DESIGN AND IMPLEMENTATION OF STEWART PLATFORM ROBOT FOR

ROBOTICS COURSE LABORATORY

A Thesis

presented to

the Faculty of California Polytechnic State University,

San Luis Obispo

In Partial Fulfillment

of the Requirements for the Degree

Master of Science in Mechanical Engineering

by

Trent Robert Peterson

March 2020
 © 2020
Trent Robert Peterson
ALL RIGHTS RESERVED

ii
 COMMITTEE MEMBERSHIP

TITLE: Design and Implementation of Stewart

Platform Robot for Robotics Course Lab-

oratory

AUTHOR: Trent Robert Peterson

DATE SUBMITTED: March 2020

COMMITTEE CHAIR: Saeed B. Niku, Ph.D.

Professor of Mechanical Engineering

COMMITTEE MEMBER: William R. Murray, Ph.D.

Professor of Mechanical Engineering

COMMITTEE MEMBER: John R. Ridgely, Ph.D.

Professor of Mechanical Engineering

iii
 ABSTRACT

Design and Implementation of Stewart Platform Robot for Robotics Course

Laboratory

Trent Robert Peterson

A Stewart Platform robot was designed, constructed, and programmed for use in

Cal Poly’s ME 423 Robotics: Fundamentals and Applications laboratory section. A

Stewart Platform is a parallel manipulator robot with six prismatic joints that has six

degrees of freedom, able to be defined in both position and orientation. Its purpose is

to supplement parallel robot material covered in lecture. Learning objectives include

applying and verifying the Stewart Platform inverse kinematics and investigating the

Stewart Platform’s operation, range of motion, and limitations. The Stewart Plat-

form geometry and inverse kinematics were modeled and animated using MATLAB.

The platform was then built using linear actuators, magnetic spherical bearings, and

acrylic plates. Control of the Stewart Platform is achieved using an Arduino Due and

a custom HexaMoto shield. Users interact with the system using a GUI created with

MATLAB’s App Designer.

Keywords: Stewart Platform, Parallel Robot, Linear Actuators, Robot, Inverse Kine-

matics, Arduino

iv
 ACKNOWLEDGMENTS

I would like to thank Dr. Saeed Niku for providing me the opportunity to create a

robot and give back to the university, and for the guidance and patience along the

way.

I would like to thank the members of my thesis committee, Dr. William Murray and

Dr. John Ridgely, for imparting onto me the knowledge to make this robot and thesis

possible.

I would like to thank Charlie Refvem for reviewing my board and giving me advice as

I assembled my electronics hardware, saving me lots of time and potentially wasted

effort.

I would like to thank the Cal Poly Mechanical Engineering Department for funding

this robot, and to Christine Haas for working with me through the process.

I would like to thank Progressive Automations for the (pending) donation of the linear

actuators.

I would like to thank Helical Products Company for supplying flexible couplings for

use with the robot.

Finally, I would like to thank my family and friends for all their support during my

time at Cal Poly.

v
 TABLE OF CONTENTS

Page

LIST OF TABLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . x

LIST OF FIGURES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi

CHAPTER

1 Introduction and Background . . . . . . . . . . . . . . . . . . . . . . . . . 1

1.1 Robots: A General Overview . . . . . . . . . . . . . . . . . . . . . . . 1

1.2 Parallel Robots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

1.2.1 Advantages and Disadvantages . . . . . . . . . . . . . . . . . 3

1.3 Stewart Platform . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

1.3.1 Geometry Stability . . . . . . . . . . . . . . . . . . . . . . . . 5

1.3.2 History . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

1.3.3 Kinematic Equations . . . . . . . . . . . . . . . . . . . . . . . 7

1.3.4 Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

1.3.5 Stewart Platform Specifications . . . . . . . . . . . . . . . . . 13

2 Mechanical Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

2.1 Linear Actuators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

2.1.1 Selection Guidelines . . . . . . . . . . . . . . . . . . . . . . . 18

2.1.2 Progressive Automation PA-14P-8-35 . . . . . . . . . . . . . . 18

2.2 Magnetic Spherical Joints . . . . . . . . . . . . . . . . . . . . . . . . 19

2.3 Platform Material . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

2.4 Shaft Couplers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

2.5 6-3 Configuration Geometry . . . . . . . . . . . . . . . . . . . . . . . 24

2.6 Electronics Enclosure . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

vi
3 Electronic Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

3.1 Microcontroller - Brief Overview . . . . . . . . . . . . . . . . . . . . . 30

3.1.1 Microcontroller Selection . . . . . . . . . . . . . . . . . . . . . 30

3.1.2 Arduino Due Specifications . . . . . . . . . . . . . . . . . . . 32

3.2 HexaMoto Shield . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

3.2.1 Design Requirements . . . . . . . . . . . . . . . . . . . . . . . 34

3.2.2 Robot Power’s MultiMoto Arduino Shield . . . . . . . . . . . 35

3.2.3 HexaMoto Design . . . . . . . . . . . . . . . . . . . . . . . . . 36

3.2.4 Component Details . . . . . . . . . . . . . . . . . . . . . . . . 38

3.2.4.1 STMicroelectronics L9958SBTR Motor Driver . . . . 38

3.2.4.2 I/O Hardware Components . . . . . . . . . . . . . . 40

3.3 HexaMoto Shield Assembly . . . . . . . . . . . . . . . . . . . . . . . 40

3.4 DC Power Supply . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42

3.5 Connectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

3.6 Electrical Assembly and Integration . . . . . . . . . . . . . . . . . . . 45

4 Software Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

4.1 Matlab Platform Simulation . . . . . . . . . . . . . . . . . . . . . . . 47

4.1.1 Geometrical Properties . . . . . . . . . . . . . . . . . . . . . . 47

4.1.2 Motion Profile . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

4.1.3 Calculation and Animation . . . . . . . . . . . . . . . . . . . . 50

4.2 Arduino Sketch . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52

4.2.1 Arduino Sketch Source Credit . . . . . . . . . . . . . . . . . . 53

4.2.2 Setup and Initialization . . . . . . . . . . . . . . . . . . . . . 54

4.2.3 Calibration and Homing . . . . . . . . . . . . . . . . . . . . . 54

4.3 Matlab GUI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55

vii
 4.3.1 Serial Connection . . . . . . . . . . . . . . . . . . . . . . . . . 56

4.3.2 Platform Configuration . . . . . . . . . . . . . . . . . . . . . . 56

4.3.3 Linear Actuator Position Calculation . . . . . . . . . . . . . . 57

4.3.4 Saving and Sequencing Points . . . . . . . . . . . . . . . . . . 58

4.3.5 Motion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59

5 Testing and Future work . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60

5.1 Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60

5.2 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64

5.3 Improvements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64

5.3.1 HexaMoto Shield . . . . . . . . . . . . . . . . . . . . . . . . . 64

5.3.2 Acrylic Plates . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

5.4 Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66

6 Lab Manual . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68

6.1 Lab Manual . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68

6.2 Learning Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68

BIBLIOGRAPHY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69

APPENDICES

A PA-14P Datasheet . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71

B WAC20-4mm-4mm Datasheet . . . . . . . . . . . . . . . . . . . . . . 79

C Mechanical Drawings . . . . . . . . . . . . . . . . . . . . . . . . . . . 81

D Mechanical Bill of Materials . . . . . . . . . . . . . . . . . . . . . . . 90

E MultiMoto Schematic . . . . . . . . . . . . . . . . . . . . . . . . . . . 91

F HexaMoto Schematic . . . . . . . . . . . . . . . . . . . . . . . . . . . 94

G Electrical Bill of Materials . . . . . . . . . . . . . . . . . . . . . . . . 97

H Replacement Manual . . . . . . . . . . . . . . . . . . . . . . . . . . . 98

viii
I Matlab Simulation Script . . . . . . . . . . . . . . . . . . . . . . . . . 105

J Arduino Sketch . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117

K Matlab GUI Code . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123

L Stewart Platform Lab Manual . . . . . . . . . . . . . . . . . . . . . . 134

ix
 LIST OF TABLES

Table Page

1.1 Stewart Platform Specifications . . . . . . . . . . . . . . . . . . . . 14

2.1 Magnetic Spherical Joint Product Options [1] . . . . . . . . . . . . 20

3.1 Arduino Due Technical Specifications [3] . . . . . . . . . . . . . . . 32

4.1 Inverse Kinematics Solution for 6-6 Configuration . . . . . . . . . . 52

5.1 Commanded and Actual Actuator Lengths . . . . . . . . . . . . . . 60

5.2 Stewart Platform Positional Test Measurements . . . . . . . . . . . 61

5.3 Stewart Platform Orientation Test Measurements . . . . . . . . . . 63

x
 LIST OF FIGURES

Figure Page

1.1 Closed and Open Kinematic Loop Mechanisms . . . . . . . . . . . 2

1.2 Stewart Platform Configurations . . . . . . . . . . . . . . . . . . . 4

1.3 Typical Stewart Platform Joint Configuration . . . . . . . . . . . . 5

1.4 Unstable 6-6 Configuration . . . . . . . . . . . . . . . . . . . . . . 6

1.5 Tire Testing Machine Application of Stewart Platform [13] . . . . . 7

1.6 Home Position of Stewart Platform Robot . . . . . . . . . . . . . . 8

1.7 AMiBA Radio Telescope in Hawaii [12] . . . . . . . . . . . . . . . . 13

2.1 Section View of an Electric Linear Actuator [10] . . . . . . . . . . . 17

2.2 Ball Joint Rod Ends [14] . . . . . . . . . . . . . . . . . . . . . . . . 19

2.3 Magnetic Spherical Joint Properties . . . . . . . . . . . . . . . . . . 21

2.4 Helical WAC20mm-4mm-4mm Couplings . . . . . . . . . . . . . . . 23

2.5 Range-of-Motion Limitation with Sharing Spherical Joint . . . . . . 24

2.6 Range-of-Motion Limitation with Sharing Spherical Joint . . . . . . 25

2.7 Joint Angle of Spherical Joint in 6-3 Configuration . . . . . . . . . 25

2.8 Electronic Enclosure . . . . . . . . . . . . . . . . . . . . . . . . . . 27

2.9 Arduino Due Mount . . . . . . . . . . . . . . . . . . . . . . . . . . 27

2.10 Molex Connector Retainer and Label . . . . . . . . . . . . . . . . . 28

3.1 Stewart Platform System Diagram . . . . . . . . . . . . . . . . . . 29

3.2 Arduino Due [3] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32

3.3 Progressive Automation PA-14P Pinout [19] . . . . . . . . . . . . . 35

3.4 Robot Power MultiMoto Arduino Shield [18] . . . . . . . . . . . . . 35

xi
3.5 HexaMoto Board Layout, Top Layer . . . . . . . . . . . . . . . . . 37

3.6 L9958 Block Diagram and PowerSO16 Pinout [22] . . . . . . . . . . 38

3.7 L9958 Application Circuit [22] . . . . . . . . . . . . . . . . . . . . . 39

3.8 SunLED Bi-Directional LED [9] . . . . . . . . . . . . . . . . . . . . 39

3.9 HexaMoto Shield I/O Hardware . . . . . . . . . . . . . . . . . . . . 40

3.10 HexaMoto Shield, Post Solder Paste Application and Reflow . . . . 41

3.11 HexaMoto Shield, Completed . . . . . . . . . . . . . . . . . . . . . 42

3.12 PA-14P Current Draw vs Load [19] . . . . . . . . . . . . . . . . . . 43

3.13 Power and USB Receptacle Connectors Enclosure . . . . . . . . . . 44

3.14 MicroUSB to USB-B Connector . . . . . . . . . . . . . . . . . . . . 44

3.15 Painted Electronic Enclosure with PSU and Connectors . . . . . . . 45

3.16 Enclosure Wiring . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46

4.1 Animation Screenshots . . . . . . . . . . . . . . . . . . . . . . . . . 47

4.2 Motion Profile Creation . . . . . . . . . . . . . . . . . . . . . . . . 49

4.3 Matlab Simulation Plots for Joint Values . . . . . . . . . . . . . . . 50

4.4 Matlab Simulation Plots for Orientation Values . . . . . . . . . . . 51

4.5 6-6 Configuration Model for Inverse Kinematics Solution . . . . . . 51

4.6 Arduino State Transition Diagram . . . . . . . . . . . . . . . . . . 53

4.7 Matlab GUI for Stewart Platform . . . . . . . . . . . . . . . . . . . 55

4.8 Serial Connection and Calibration . . . . . . . . . . . . . . . . . . . 56

4.9 Top and Bottom Joint Locations . . . . . . . . . . . . . . . . . . . 57

4.10 Position and Orientation Input . . . . . . . . . . . . . . . . . . . . 57

4.11 Save and Sequence Panel . . . . . . . . . . . . . . . . . . . . . . . . 58

4.12 Motion Control Panel . . . . . . . . . . . . . . . . . . . . . . . . . 59

xii
5.1 Inclinometer App Example Readings . . . . . . . . . . . . . . . . . 62

5.2 HexaMoto Shield Improvements . . . . . . . . . . . . . . . . . . . . 65

5.3 3Dconnexion 3D Mouse [2] . . . . . . . . . . . . . . . . . . . . . . . 67

xiii
 Chapter 1

INTRODUCTION AND BACKGROUND

1.1 Robots: A General Overview

A robot is a machine that can be easily commanded, typically by a computer, to

perform a set of tasks automatically. Robots are extraordinarily diverse in size,

performance, cost, and operation. They typically have multiple links and actuators

that create kinematic chains, and connect to an end effector. An end effector is a tool

that is connected to the end of a robotic mechanism. The end effector may move in

space in up to six ways called degrees of freedom (DOF). Three degrees of freedom

are translation which occur along mutually-orthogonal axes, typically referred to as

X, Y, and Z. Another three DOF may be rotation, which occur about these three

axis. These six DOF are necessary to completely define the position and orientation

of an object in space.

A joint is the connection between links that allows motion. This motion can be linear,

rotary, or spherical. Examples of linear joints include linear motors, fluid-powered

cylinders, and electric linear actuators. An electric linear actuator is articulated by

a screw mechanism, powered by a rotary motor through gearing. A linear motor has

the same operating principle as a rotary motor, but its rotor and stator are ’unfolded’

allowing for linear but finite motion. It is important to note that a joint does not

necessitate complete control. Spherical joints are difficult to control in practice. In

addition, pneumatic joints are excellent for actuating to the extended and retracted

positions of the cylinder, but controlling their actuation to intermediate positions

1
accurately is more difficult. Links are rigid members that connect the joints of a

robot.

Forward kinematics and inverse kinematics are an integral segment of robotic study

and analysis. Forward kinematics allows the position and orientation of the end

effector to be known, given all joint values. Inverse kinematics determines the joint

values given a known or desired end effector position and orientation.

The combination of joints and links comprise a kinematic loop, which can be either

closed or open. An open kinematic loop occurs when links are connected in serial

form via joints. The kinematic loop is closed when each link, which includes ground,

is connected to a minimum of two other links. A four-bar linkage is a common

mechanism that has a closed kinematic loop. The arm of an excavator is an example

of an open loop mechanism (Figure 1.1).

(a) Closed Kinematic Loop (b) Open Kinematic Loop

Figure 1.1: Closed and Open Kinematic Loop Mechanisms

1.2 Parallel Robots

J. P. Merlet, author of Parallel Robots, defines a parallel general manipulator as a

“closed-loop kinematic chain mechanism whose end effector is linked to the base by

several independent kinematic chains” [15]. The manipulator becomes an n-degree-of-

freedom robot when controlled actuation occurs through quantity n actuators. This

2
geometry gives a parallel robot its distinct characteristics, analysis, and behavior.

Parallel robots can have up to six DOF and have various constructions.

1.2.1 Advantages and Disadvantages

It is important to compare the differences between serial and parallel robots and

illustrate the areas in which each type of robot excels. An important characteristic of

a robot is its payload-to-mass ratio, which is typically higher for parallel robots. This

is due to the closed-loop construction of a parallel robot. By distributing the loading

of the robot between its actuators, this construction supports greater forces that

comes with increased mass or acceleration. In comparison, the open-loop construction

of a serial robot successively increases the load seen by each joint. The outermost

joint, the wrist, needs to support the mass and inertial forces from the end effector

only. The next joint, the elbow, must support the end effector, the wrist joint, and

the forearm link. This continues to the base of the robot. Loading of an open-

loop mechanism can cause deformation that is not detected by sensors which reduces

positional accuracy. To compensate, the serial robot must be made very stiff, and

this over-designed geometry increases robot weight. High stiffness and payload-to-

mass ratios in parallel robots is generally accompanied by a reduction in workspace

compared to similarly-sized serial robots.

Solving the forward kinematics equations for a serial robot is relatively straightfor-

ward, but their inverse kinematic solutions are much more involved. Conversely, the

inverse kinematics solution for a parallel robot is much less effort than its forward

kinematics solution, which is an ongoing area of research for some types of parallel

robots.

3
1.3 Stewart Platform

The Stewart Platform is a common type of parallel manipulator and possesses six

degrees of freedom. In their paper, Cruz, Ferreira, and Sequeira [5] state that “the

generic Stewart-Gough platform is composed of two rigid bodies connected through

a number of prismatic actuators as in a parallel arrangement of kinematic chains.

Usually six actuators are used, pairing arbitrary points in the two bodies.”

The actuators are connected to the rigid bodies by spherical or universal joints. This

would mean the Stewart Platform is a spherical-prismatic-spherical, or SPS, robot.

Although the joint positions can be arbitrary, evenly spacing them results in special

cases. When the joints are spaced 60° around the base and the moving platform, this

is called a 6-6 configuration. When a pair of actuators share the same joint spaced

120° on the moving platform, this is called a 6-3 configuration. 120° spacing on top

and bottom platforms results in a 3-3 configuration as shown in Figure 1.2.

(a) Type 6-6 (b) Type 6-3 (c) Type 3-3

Figure 1.2: Stewart Platform Configurations

4
1.3.1 Geometry Stability

In a literature survey, a true type 6-6 Stewart Platform was never found to be used

in practice. Typically, platforms were designed as a mix between the 6-6 and 3-3

configuration. Instead of pairs of actuators sharing the same joint, each actuator had

its own joint like in a 6-6 configuration; pairs of joints were separated by a small gap,

resulting in a geometry visually similar to Type 3-3. A diagram of this geometry is

shown in Figure 1.3 and example platforms are shown in Section 1.3.4.

Figure 1.3: Typical Stewart Platform Joint Configuration

Further analysis reveals that a true 6-6 platform with 60° spacing between all joints

is unstable because there is no force balance, allowing the top platform to rotate

about its axis. This is illustrated in Figure 1.4. When a small disturbance occurs, the

links cannot exert forces in opposite directions resulting in a couple about the links.

Spherical joints cannot resist this moment, leading to a collapse. This shortcoming

of a true 6-6 configuration shows the value of 6-3, 3-3, and comparable geometries.

5
 Figure 1.4: Unstable 6-6 Configuration

1.3.2 History

The early stages of robotics were dominated by serial robots, which drew heavy

inspiration from the human arm and its utility. It was the contributions by Stewart

and Gough that brought parallel robots under the scope of further research and

development, especially in industry in recent years. According to Merlet [15], Eric

Gough established a closed-loop kinematic mechanism and its principles in 1947 and

built a prototype for tire testing in 1955. In the 1960s, the rise of aeronautics created

a demand for flight simulators. D. Stewart published “A Platform with Six Degrees

of Freedom” in 1965, discussing a triangular parallel mechanism design approach

for the simulators. While Stewart had important contributions to flight simulator

development, his published design was not used in practice. On the other hand the

tire-testing mechanism designed by Gough, who was one of the reviewers of Stewart’s

paper, found heavy usage in industry. Gough’s Tire Testing Machine is shown in

Figure 1.5 and used six jacks to exert combined loads to test Dunlop tires. Regardless,

6
the platform has become known as the Stewart-Gough Platform, or more informally

as the Stewart Platform.

Figure 1.5: Tire Testing Machine Application of Stewart Platform [13]

1.3.3 Kinematic Equations

The Stewart Platform kinematics are derived from analysis of its six kinematic chains.

The following derivation is drawn from Dr. Saeed Niku’s inverse kinematic derivation

of the Stewart Platform, found in the 3rd edition of his book, Introduction to Robotics:

Analysis, Control, Applications [17].

7
 Figure 1.6: Home Position of Stewart Platform Robot

A fixed reference frame xyz is placed at the center of the platform base. The z -axis is

normal to the base and the x -axis is parallel to it, pointed towards a spherical joint.

A moving reference frame noa is similarly placed at the center of the moving platform

with the a-axis normal to the moving platform face and the n-axis pointed towards a

spherical joint. The spherical joints are assumed to be connected through the linear

actuator. Frame noa rotates by θ,φ,ψ respectively.

Four parts compose each kinematic chain C: A connects the origin of the fixed

reference frame and base spherical joint at an angle from x. B connects the origin

of the moving reference frame and moving platform spherical joint at an angle from

n. L connects the spherical joints. P connects the origins of the two frames, and is

common for all six chains. The purpose of the inverse kinematics is to determine the

length L of each linear actuator, given the position P and orientation of frame noa.

The vector equation to determine each actuator length is

8
 Li = P + Bi − Ai (1.1)

for i = 1..6.

For chain C1 , A1 and B1 lie along the x -axis and n-axis respectively and their angles

are zero. In a Type 6-6 Stewart Platform, subsequent chains will be multiples of 60°

from the x and n axes. This is not always the case, and the remaining angles can be

arbitrary values; in practice, Stewart Platforms will have some degree of symmetry.

Following the vector convention of Introduction to Robotics by Dr. Saeed Niku [16],

Ai for the general case can be written as

 
Acos(ith )
 
 
Asin(ith )
Ai =  (1.2)
 

0
 
 
 
1

where the it h term is the angle that Ai makes with the x axis.

Ai for i=1..6 will remain constant throughout the motion of the Stewart Platform for

a given geometry.

Similarly, Bi can be written as

 
Bcos(ith )
 
 
Bsin(ith )
Bi =  (1.3)
 

0
 
 
 
1

9
for the moving platform when it is in the home position as shown in Figure 1.6. The

home position is achieved after the robot is calibrated, and all the actuators are at

their minimum length.

Because Bi is attached to moving frame noa, its components will change as the plat-

form rotates. To account for this rotation, Bi must be pre-multiplied by rotation

matrices. Using the variables from the moving reference frame, the rotation matrices

are:

 
1 0 0 
 
Rot(x, θ) = 
0 cosθ −sinθ
 (1.4)
 
0 sinθ cosθ
 
 cosφ 0 sinφ
 
Rot(y, φ) = 
 0 1 0  (1.5)
 
−sinφ 0 cosφ
 
cosψ −sinψ 0
 
Rot(z, ψ) = 
sinψ cosψ 0
 (1.6)
 
0 0 1

Equations 1.4, 1.5, and 1.6 are from equations 2.20 and 2.21 in Introduction to Robotics

[16]. These 3x3 rotation matrices only account for rotation. They can be expanded

to 4x4 to include position, in which the fourth row and column are zero, except for a

value of one located in element 4,4.

In order to determine the value of Bi after rotation for each kinematic chain, it must

be pre-multiplied by the above rotation matrices, yielding:

10
    
Bx Bx
 i   i 
   
Bi y     
Bi y 
= Rot(z, ψ) Rot(y, ψ) Rot(x, θ)  (1.7)
   
  
Bi z  Bi z 
   
   
1 1
rot home

Multiplying Rot(z, ψ) and Rot(y, φ) yields

 
   cosφcosψ −sinψ cosψsinφ
 
Rot(z, ψ) Rot(y, ψ) = 
cosφsinψ cosψ sinφsinψ 
 (1.8)
 
−sinφ 0 cosφ

Multiplying the result of Equation 1.8 by Rot(x, θ) yields

  
cosφcosψ −sinψ cosψsinφ 1 0 0 
  
cosφsinψ cosψ sinφsinψ  0 cosθ −sinθ =
  
  
−sinφ 0 cosφ 0 sinθ cosθ

 
cosφcosψ cosψsinφsinθ − cosθsinψ sinψsinθ + cosψcosθsinφ
 
cosφsinψ cosψcosθ + sinφsinψsinθ cosθsinφsinψ − cosψsinθ
  (1.9)
 
−sinφ cosφsinθ cosφcosθ

After combining the rotation matrices, Equation 1.7 becomes

    
B
 i x  cosφcosψ cosψsinφsinθ−cosθsinψ sinψsinθ+cosψcosθsinφ 0 B
 i  x
    
 Bi y  cosφsinψ cosψcosθ+sinφsinψsinθ cosθsinφsinψ−cosψsinθ 0 Bi y 
= (1.10)
    
   
B z   −sinφ cosφsinθ cosφcosθ 0  Bi z 
 
 i   
    
1 0 0 0 1 1
rot home

11
[Bi ]home is known from the Stewart Platform geometry, and θ, φ, and ψ are known

upon specifying the orientation of the moving platform. Therefore, [Bi ]rot can be

solved. Expanding the vectors Equation 1.1 into component form yields

       
Lix  Pix  Bix  Aix 
       
L  = P  + B  − A  (1.11)
 iy   iy   iy   iy 
       
Liz Piz Biz Aiz

for i=1..6.

The linear actuator lengths are the desired quantity from the kinematics. The mag-

nitude of each length is determined by

q
|Li | = (Lix )2 + (Liy )2 + (Liz )2 (1.12)

for i=1..6.

These calculated lengths become the setpoints for controlling the Stewart Platform.

1.3.4 Applications

There are several applications suited for a Stewart Platform. Flight simulators make

use of the platforms 6 DOF to simulate orientation and motion. The Stewart Platform

can accurately orient to the correct roll, pitch and yaw, and their high-dynamic

capability can create linear and nonlinear forces, like a bump. Shake tables make use

of this capability to create high-acceleration, low-displacement moves. Fiber-optic

equipment can be aligned using a Stewart Platform because they must be accurately

positioned and oriented, enabling the light to refract and the optic signal to operate

12
properly. In astronomy, Stewart Platforms are used in spherical radio telescopes and

trajectory tracking, such as the one shown in Figure 1.7.

Figure 1.7: AMiBA Radio Telescope in Hawaii [12]

1.3.5 Stewart Platform Specifications

The application of this Stewart Platform is for education and experimentation to pro-

mote the ”Learn By Doing” experience in regards to parallel robots. Thus, there are

not external deterministic requirements or specifications that the Stewart Platform

must meet. Instead, an internal specification will be set. Quantitative specifications

are found in Table 1.1.

13
 Table 1.1: Stewart Platform Specifications
Metric Specification Actual Value
Mechanical Properties
Weight < 40 lb 23.7 lb
Length < 24 in 21 in
Width < 24 in 21 in
Height < 36 in 31 in
Range of Motion
θ > 30° 44°
φ > 30° 54°
ψ > 60° 90°
X 6 in 13.5 in
Y 6 in 14.5 in
Z 6 in 25.0 in
Accuracy
Angular < 2° 1°
Positional 0.25 in 0.25 in

The mechanical properties were set to allow the robot to fit comfortably on a desk

surface, and be moved by one person. The specified range of motion demonstrates

that the Stewart Platform can properly utilize each of its six degrees of freedom. The

robot is not being used for tooling or processing applications, and does not need to

be extremely accurate; it must be accurate enough to pass visual observation by the

user.

There are also some qualitative specifications set for the Stewart Platform. The

platform should be safe for students to operate and experiment with. It should be

designed to be easy to manufacture using Cal Poly Machine Shop resources. Chosen

vendor and custom parts used should be readily available or easy to replace, keeping

the Stewart Platform serviceable. It would be educationally beneficial to reconfigure

the robot. Typically, the geometry of a robot is optimized to solve a problem or

complete a task. For this lab purpose, exploring multiple configurations and their

benefits and limitations.

14
 Chapter 2

MECHANICAL DESIGN

There were several important foci of the mechanical design of the Stewart Platform.

The first was to create a platform that was easy to assemble and disassemble into var-

ious configurations. As mentioned previously in Section 1.3, the Stewart Platform has

three primary configurations, 6-6, 6-3, and 3-3. To avoid limiting the Stewart Plat-

form to just one configuration, it was designed to be re-configurable within minutes

with just a hex key.

Another design focus was ability to manufacture and assemble the Stewart Platform

with only university resources. This focus was coupled with budget considerations,

as special tooling and outsourcing manufacturing could significantly increase cost.

Therefore, all components were required to be manufactured in-house using machine

shop resources.

2.1 Linear Actuators

At its core, a linear actuator is a device that causes movement along an axis. This can

typically be achieved with hydraulic, pneumatic, or electrical power. These mediums

of actuation were considered with regards to the Stewart Platform, specifically its

performance requirements and intended lab environment. The linear actuators that

operate the Stewart Platform must have a form of positional feedback, because the

joint lengths are the controlled variables by which the Stewart Platform is actuated.

The actuators are primarily oriented vertically in the Stewart Platform, and must be

capable of both supporting and moving a payload.

15
There are three common characteristics in linear actuator terminology: stroke, force,

and speed. The stroke is the difference between the actuator length when fully ex-

tended and fully retracted. The force is the maximum load the linear actuator can

handle and this can be rated for both static and dynamic conditions. Speed is simply

how fast the actuator extends or retracts.

Hydraulic cylinders with servo-valves under position control would provide sufficient

positional accuracy for a Stewart Platform to function properly. Additionally, under

a typical hydraulic pressure of 3000 psi, a single piston with a 0.5 inch cylinder could

actuate with a force up to 589 lb. While hydraulics satisfies the force and positional

control requirements, it requires dedicated infrastructure and relatively high cost,

as hydraulic systems require pumps, filters, valves, hoses, and tubing. Therefore,

hydraulic actuation is inappropriate for this Stewart Platform.

Pneumatic systems, generally speaking, can be created with less cost and infrastruc-

ture than hydraulic systems, and provide sufficient force within the defined scope

of this Stewart Platform. However, the positional control of pneumatic cylinders is

difficult due to the compressibility of air and its nonlinear behavior. Even indus-

trial, purpose-built pneumatic positioning controllers [24] after optimized tuning, can

achieve as as little as ±1 mm (or ±0.039 inches) positional accuracy. Products that

can achieve these specifications are beyond the budget for the project, and pneumatic

actuation is also inappropriate for the Stewart Platform.

Electric linear actuators are another common means of linear actuation. Firgelli

Automations [10] gives an overall introduction to electric linear actuators. They

convert the rotational motion of a motor into linear motion through a gear train,

which connects to a lead screw and nut as shown in Figure 2.1.

16
 Figure 2.1: Section View of an Electric Linear Actuator [10]

Most electric linear actuators have factory-installed limit switches that cut the power

to the motor once the actuator has reached its maximum or minimum travel distance.

Typically, a 12 VDC motor is used to drive linear actuators, but DC motors at other

voltages and AC motors are available as well. Some electric linear actuators also have

position feedback, usually provided from a potentiometer.

Progressive Automations, Thomson, and Firgelli Automations offer linear actuators

in their product line that range from miniature with light payload ratings, around 5

lb, to heavy duty models able to carry several hundred pounds.

The electric linear actuator is lower cost, relatively easy to integrate, can exert moder-

ate forces, and is very suited to position control. These factors make electric actuators

the best choice for this Stewart Platform.

17
2.1.1 Selection Guidelines

As described above, there are several specifications available for choosing electric

linear actuators. The scope of the project helped dictate these specifications. The

physical envelope of the robot was established to have an approximate 1.5 ft. by 1.5

ft footprint, with a vertical range between 1.5 - 2 feet. Using these guidelines, a stroke

between 6 inches to 10 inches was desired. It is important to note that when fully

extended, the overall length of the linear actuator will be at least twice the stroke. The

intended usage of the Stewart Platform is for experimentation with kinematics and

motion, and will not be carrying heavy loads or aligning precise equipment. Therefore,

high-speed, low-load actuators are the best fit for this design. Non-hardware design

decisions include low cost and ease of use and implementation, as excessive cost or

time to implement would be harmful to the project completion and cause delays

if spare parts are needed or installed. After looking through several vendors, the

PA-14P-8-35 linear actuator offered by Progressive Automations best fits the design

parameters.

2.1.2 Progressive Automation PA-14P-8-35

The PA-14P-8-35 is an electric linear actuator that has an 8-inch stroke and exerts

a dynamic force of 35 lb. At no load, the actuator moves at 2 in/sec, drawing 1.0

A. At full load, the actuator moves at 1.38 in/sec and draws its maximum 5.0 A.

It has a potentiometer that provides the positional feedback required to control the

Stewart Platform. As of this writing, these actuators cost $138.99 each. While

there are cheaper linear actuators, Progressive Automations provides a sponsorship

opportunity which would retroactively reduce the price of the actuators if approved,

making them budget-friendly. Finally, Progressive Automations has sample Arduino

18
code provided for controlling multiple actuators and their timing, monitoring their

current, and more. Further information can be found in the PA-14P Datasheet [19]

in Appendix A.

2.2 Magnetic Spherical Joints

The Stewart Platform has 6, 9, or 12 passive spherical joints, depending on the

configuration of the platform. The contribution of spherical joints to the degrees of

freedom of the platform was discussed in Section 1.3. A common hardware choice for a

spherical joint is a ball joint, shown in Figure 2.2. The swivel angle is a characteristic

of a ball joint that specifies the amount of rotation possible from the axis normal to

the axis of the rod. This is an important characteristic because the range of the ball

swivel affects the range of the platform operation. On McMaster Carr’s website, most

ball joint rod ends have a maximum ball swivel between 20°-30°[14].

Figure 2.2: Ball Joint Rod Ends [14]

Another choice that satisfies the spherical joint requirement of the Stewart Platform is

a magnetic spherical joint. These spherical joints are comprised of a steel ball with a

protruding threaded rod, coupled with a Neodymium Iron Boron (NdFeB) permanent

magnet housed in cylindrical base as shown in Figure 2.3a. The magnetic joints come

in a range of sizes and holding force ratings. Table 2.1, adapted from AliExpress [1],

19
shows a range of common values for these magnets. The terminology in the table

corresponds to the labels in Figure 2.3b, and all dimensions are in millimeters.

Table 2.1: Magnetic Spherical Joint Product Options [1]

Type D L A B THD Holding Force Cost

KD310 10 20 10 12 M3 18 N $3.65
KD312 10 20 12 12 M3 18 N $3.95
KD413 13 20 13 12 M4 40 N $4.84
KD418 13 20 18 12 M4 40 N $5.11
KD625 20 25 25 16 M5 150 N $6.55
KD725 25 35 25 16 M5 200 N $8.85

Cost, size, and holding force were all considered when selecting the magnetic spher-

ical joints. In general, all three of these characteristics increase together. A lower

cost spherical joint is preferred, especially since 12 joints are needed. This must be

balanced by the need for sufficient holding force. If the holding force is too low, then

the magnetic joint may separate during normal operation which must be avoided.

The holding force cannot be too high either; if a binding condition occurs, the joint

needs to be able to fail and separate. This prevents the linear actuator from stalling

and overheating while relieving the acrylic plate loadings. The maximum force the

linear actuators can exert is 35 lb, equivalent to 156 N. The KD625 seems to perfectly

meet this requirement but there is very little cushion between the holding force and

actuator force. On the other end, the KD310 is the lowest cost joint, but is dwarfed

by the linear actuators, and can hold little more than the weight of the linear actu-

ators. The KD418 was chosen as a moderate cost, size, and holding force solution,

and it has functioned as intended in operation of the Stewart Platform.

There are additional advantages in using the magnetic joints with this Stewart plat-

form. The magnetic ball joints have a ball swivel angle of 242°, as shown in Figure

2.3c, around an order of magnitude greater than the ball joints previously mentioned.

20
 (a) Spherical Components (b) Dimension Key (c) Angle Limit

Figure 2.3: Magnetic Spherical Joint Properties

This angle is well beyond the angles of the linear actuators relative to the base and

removes the joints as a limiting factor of range of motion. Friction was a concern ini-

tially because of a lack of lubrication between the steel and magnet surfaces. Friction

was very apparent when the steel ball was rotated against the magnet by hand, and

smooth rotation was difficult to achieve. However, the rotation was much smoother

once the linear actuator was attached to the joint. This is because the linear actu-

ator provides several inches of leverage, whereas twisting the sphere or pushing the

threaded rod only gives about an inch of leverage. When installed on the Stewart

Platform, the actuators easily overcome friction and rotate smoothly.

2.3 Platform Material

Low weight, low cost, ease of manufacturing, and pleasing aesthetics were all con-

sidered when selecting the material for the top and base platform plates. Wood,

acrylic, and aluminum were top material candidates. Weight is not critical for this

application, but lower plate weight reduces the overall Stewart Platform weight. Ad-

ditionally, the weight of the top plate takes away from the payload capability of the

21
platform. Wood and acrylic have lower densities than aluminum, but aluminum is

stiffer, allowing for plate thickness and weight reduction; no material has a significant

advantage for weight. Wood is around 2.5 times cheaper than acrylic and several

times cheaper than aluminum for the same raw material volume. Reducing the plate

thickness for aluminum still results in a much higher relative cost. All materials are

able to be manufactured at Cal Poly; wood and acrylic can be cut using a laser cut-

ter, and the aluminum can be cut with a water jet. While wood can be aesthetic,

acrylic and aluminum are best suited to the Stewart Platform’s more industrial and

mechanical appearance.

Acrylic was chosen because it is the best overall fit due to its moderate weight, cost,

and easy to laser-cut. The base and top plates are both 1/4 inch thick acrylic sheets

and weigh 2.67 lb and 1.66 lb respectively. For reference, if made out of aluminum

6061-T6 of equivalent size, their weights would be 6.02 and 3.75 lb respectively. Laser

cutting of the acrylic was straightforward and resulted in aesthetic Stewart Platform

plates. Additionally, acrylic is transparent which is beneficial because the electronics

under the base platform are visible, exposing students to the components that control

and drive the system.

2.4 Shaft Couplers

In a Stewart Platform, all six linear actuators must actuate together in a coordinated

manner. If they do not, then the position and orientation of the robot will be incor-

rect. At worst, a large deviation in one or more actuators can cause the robot to move

outside of its envelope or achieve an unstable position, causing the magnetic joints to

separate. It is evident that the Stewart Platform, and parallel robots in general, are

very sensitive to actuator accuracy. This necessitates a mechanical solution called

22
compliance. Bram Vanderborght gives a helpful introduction to compliance, sum-

marizing that a compliant member will allow deviations from its equilibrium when a

force acts on it, whereas a stiff member will not, within the device limits [23].

The magnetic joints, linear actuators, and acrylic are considered stiff components.

However, the acrylic is the least stiff of the three so if the robot encounters a binding

position or an unintended loading condition, the acrylic plates will take the load,

risking deformation and breakage. This should be avoided by placing a compliant

member in series with the linear actuator. The compliant member must be in a

compact and professional package, and removable if the Stewart Platform can be run

without the compliant member.

Helical shaft couplings were chosen to be the compliant member because they sat-

isfied the above requirements. Shown in Figure 2.4, they are primarily intended for

torque transmission but offer the flexibility in axial and bending motions to provide

the necessary compliance. The shaft couplings chosen are the WAC20-4mm-4mm

Aluminum Alloy Couplings and the data sheet is located in Appendix B.

Figure 2.4: Helical WAC20mm-4mm-4mm Couplings

23
2.5 6-3 Configuration Geometry

In a 6-3 configuration, the six base joints are evenly spaced at 60° angles and the three

top joints are evenly spaced at 120° angles. Two actuators have to share a joint in this

configuration. When two magnets are attached to a single steel ball to accomplish

this, there is interference shown in Figure 2.5, where the Stewart Platform is in a

normal position and orientation.

Figure 2.5: Range-of-Motion Limitation with Sharing Spherical Joint

The only way for a spherical joint to be shared by two magnetic joints is for their

cylindrical axis to be greater than 92.9° apart as shown in Figure 2.6. This is only pos-

sible outside of the range-of-motion of the Stewart Platform. Therefore, the magnetic

portions of the joints cannot share the same steel ball.

24
 Figure 2.6: Range-of-Motion Limitation with Sharing Spherical Joint

To fix this, the spherical joints had to be separated. Separation means the robot will

not truly be in a 6-3 configuration, only close to it. The geometry of the spheres drove

the angle separation of 3.47° between each joint from the 120° centerline is shown in

Figure 2.7.

Figure 2.7: Joint Angle of Spherical Joint in 6-3 Configuration

25
This solution creates a new problem; the joint angles for the 6-3 configuration are no

longer at 120°because of this angle of separation. This joint angle must be included in

the modeling and kinematic equations of the Stewart Platform. This implementation

is discussed in Section 4.1.1.

2.6 Electronics Enclosure

There are a few key electronic components required for the operation of the Stewart

Platform. These components include a power supply, microcontroller, motor driver,

and connection ports, plugs, and switches. These components are further discussed

and detailed in Chapter 3. The components must be housed in an electronics en-

closure for safety and reliability, protecting users from contact with live power and

internal components from careless or accidental handling. Optimally, the enclosure

should be relatively lightweight and portable, provide sufficient airflow, and be easy

to manufacture at a low material price. Furthermore, the enclosure is an appropriate

means to connect the PA-14P motor and potentiometer through the included 6-pin

connector.

Sheet metal was chosen to be the electronics enclosure material because of its easy

manufacturability, high strength, and low cost. Alternatives considered were plastic

3D printed enclosures or acrylic. The volume was too large and geometry too simple to

justify all the 3D printed material. Acrylic could have been cut on the a laser cutter,

but each surface would have to be fastened together. The sheet metal enclosure was

designed, cut with a water jet, and bent using Cal Poly resources. The enclosure is

shown in Figure 2.8

The front of the enclosure has three cutouts for connecting USB and AC power along

with a power switch. The rear has six cutouts for the female 6-pin Molex connectors

26
 (a) Enclosure Front (b) Enclosure Rear

Figure 2.8: Electronic Enclosure

that attach to the linear actuator connectors. On the walls of the enclosure are vents

that allow air circulation. The enclosure will be attached underneath the base of the

Stewart Platform, making the Stewart Platform and its electronics a complete unit.

The power supply resides in the bottom of the electronics enclosure. Suspended above

it is the Arduino due and motor driver, to be discussed in Chapter 3, which helps

keep the enclosure compact. This uninstalled bracket is shown in Figure 2.9. The

drawings for all manufactured mechanical parts are located in Appendix C.

Figure 2.9: Arduino Due Mount

27
When the 6-pin female Molex connectors were inserted into the enclosure walls, the

wall thickness was insufficient to retain the connectors. When the linear actuator was

plugged in, the force would push the connector back into the enclosure, which would be

impossible to bring back into place when the enclosure is installed under the Stewart

Platform base. Retainers were printed to increase the wall thickness, allowing the

connectors to stay in place when the linear actuators are plugged in. These retainers

are shown in Figure 2.10. An additional benefit from these retainers is a permanent

labeling solution, ensuring that the actuator number cannot be switched or removed.

Figure 2.10: Molex Connector Retainer and Label

The bill of materials for all mechanical components of the Stewart Platform can

be found in Appendix D. The total weight of the Stewart Platform is 23.7 pounds,

which includes mechanical and electrical hardware. The electronics are discussed in

Chapter 3.

28
 Chapter 3

ELECTRONIC DESIGN

The electronic design of the Stewart Platform must meet the operational needs of the

robot. The robot needs power to drive the actuators through its motions, which need

to be accurately calculated and controlled. A controller can apply a control loop and

send actuation signals based on calculated inputs and current positions of the linear

actuators.

Figure 3.1: Stewart Platform System Diagram

A high level overview of the system is presented in Figure 3.1. At the top of the

diagram are the PA-14P actuators discussed in Chapter 2. This chapter discusses

the design and realization of the middle portion, the electronics. A controller, motor

29
drivers, and a power supply comprise the core electronic components. The bottom of

the diagram incorporates the code inside the controller and the user interface with

the Stewart Platform, discussed in Chapter 4.

3.1 Microcontroller - Brief Overview

A microcontroller uses a microprocessor, along with peripherals including read-only

memory (ROM), random-access memory (RAM), and input and output capabilities

(I/O) to make a computer on a chip [4]. They are typically used as part of an

electronic or mechanical system with dedicated functionality. This is known as an

embedded system, which are often subject to real-time constraints. There are count-

less microcontrollers available to control the Stewart Platform.

3.1.1 Microcontroller Selection

An Arduino board was chosen as the microcontroller platform to control the robot for

several reasons. Arduinos are generally implemented in an easy-to-use prototyping

environment comprised for makers, hobbyists, students, and more. The setup, usage,

and intended audience of an Arduino platform make it a suitable choice for the Stewart

robot. The board’s design and software are open-source, providing a basis to make

modifications as necessary. For board designs and hardware, modifications can be

made to existing designs to implement additional features or tailor performance to

a specific application. Hardware modifications were not necessary for the Stewart

Platform, but having this option was helpful in case the necessity did arise further into

the design. An additional benefit to using an Arduino is the simplicity of replacement.

If the Arduino board becomes faulty, it can be replaced within a few days by ordering

30
a ready-made board, installing it, and downloading the project to the new board.

This method is a near-turnkey solution to replacing the microcontroller.

Availability of usable software and resources was another driving factor in deciding

to use the Arduino platform. This served to reduce development time, errors, and

shift the development focus from setup to application and refinement. Progressive

Automations has several examples of Arduino code that apply directly to their linear

actuators. Examples of code applicable to this project include “Controlling Multiple

Actuators with the Multimoto Arduino Shield” and “Controlling the Timing of a

Single Linear Actuators Motion” [20]. The provided code helped with the setup

and initialization required for working with the PA-14P actuators. Further details

regarding the software used for operating the robot are discussed in Section 4.2.

Arduino offers several board choices, with varying levels of features and performance.

Boards include the Uno, Nano, Mega, and Due. The selected board, at the minimum,

has to have available a minimum number of necessary inputs and outputs. The PA-

14P actuators each have a potentiometer, so there must be 6 analog inputs to read the

position of the actuators. The motor driver, covered in further detail in Section 3.2,

outputs voltage at a level and polarity based on a PWM pin and a direction pin. The

motor drivers also share an Enable and Disable pin. This required 14 digital outputs,

6 of which must be capable of PWM output. The Arduino Mega 2560 and Due fulfill

both requirements. They have some similarities in regards to I/O capabilities, but

there are differences in their features. However, the Due offers a clock 5.25x faster

than the Mega 2560, twice the flash memory size, and 12 times the SRAM at the same

price point. The Due is selected as the microcontroller for controlling the Stewart

Platform, and is shown in Figure 3.2.

31
 (a) Isometric View (b) Top View

Figure 3.2: Arduino Due [3]

3.1.2 Arduino Due Specifications

The following section describes the selected Arduino Due in further detail and its

application to the Stewart platform. Table 3.1 summarizes some of the technical

specifications of the Arduino Due.

Table 3.1: Arduino Due Technical Specifications [3]

Feature Value
Microcontroller AT91SAM3X8E
Operating Voltage 3.3V
Digital I/O Pins 54
Analog Input Pins 12
Analog Output Pins 2
Flash Memory 512
SRAM 96 KB
Clock Speed 84 MHz

A distinct characteristic of the Arduino Due is the presence of two USB ports. Ac-

cording to Arduino’s web page on the Due, one port is a Programming port and the

other is a Native port. The Programming port is connected to an ATmega16U2,

which connects to the SAM3X though hardware UART and also provides a virtual

32
COM port to a connected computer. The Native USB connects to the SAM3X, en-

abling serial communication over USB. The Native port can also emulate various USB

devices, or act as a USB host. Both ports can be used for programming the Due,

but there are some notable differences. When the Programming port is connected

at 1200 baud and opened or closed, a SAM3X “hard-erase” procedure is performed,

activating the Erase and Reset pins before UART communication. In comparison,

the Native port performs a “soft-erase” procedure when opened and closed at 1200

baud. Flash memory is erased and the bootloader restarts the board. Programming

via the Programming port is preferred because it is more reliable compared to the

soft-erase procedure, which may not perform when the microcontroller has crashed.

However, the Programming port is limited to a baud rate of 115200, which can be an

issue if high-speed serial communication is required. The Native port is capable of

much higher serial communication speeds because of the USB connection, and setting

a baud rate in the Arduino sketch is ignored. This functionality and performance dif-

ference was notable in programming and operating the robot. Uploading code using

the Native port was several times faster than uploading via the Programming port.

However, the Programming port was used when uploaded sketches crashed because of

its reliability. For normal operations, functionality, and data transmission, the Native

port was used [3]. All further information on the Arduino Due, including overview,

technical specifications, and documentation, are available on the cited Arduino Due

web page.

3.2 HexaMoto Shield

The Arduino Due is capable of controlling motors through its PWM outputs but

lacks the hardware to power them so an additional motor driver board is necessary

33
to complete the electronic hardware design. Two approaches include purchasing an

aftermarket motor driver board or designing and assembling a custom board.

3.2.1 Design Requirements

The motor driver shield will need to have six channels to power each linear actuator.

Each motor driver must provide a form of speed control. The robot operation would

not be acceptable if the actuators could only operate at maximum speed. Addition-

ally, it would be ideal to incorporate the potentiometer signals on the Arduino shield.

This would provide an inclusive solution in which all inputs and outputs (except for

12V motor power) interface with the Arduino Due through the shield.

A potentiometer outputs a voltage as its analog signal. An 8-bit analog-to-digital

converter (ADC) can output 256 discreet values. For a 4-inch stroke, each discreet

change in the ADC reading of the potentiometer corresponds to a length of 0.0156

inches. For a 12-inch stroke, this value is 0.0469 inches. Similarly, a 10-bit ADC that

outputs 1024 values means each increment is a change in length of 0.004 inches for

a 4-inch stroke and 0.012 inches for a 12-inch stroke. It is important to note that

ADCs have sources of measurement error, and they are not completely accurate to

the above values.

In regards to electronics integration, Progressive Automation’s DC linear actuator

with a potentiometer will have five wires. Two wires for motor power will provide

+12VDC and 0VDC (GND) and three for the potentiometer will provide +5VDC,

signal and GND. The shield should accommodate these five lines for each of the six

actuators. This is illustrated in the PA-14P datasheet shown in Figure 3.3.

34
 Figure 3.3: Progressive Automation PA-14P Pinout [19]

3.2.2 Robot Power’s MultiMoto Arduino Shield

Progressive Automations has various products suggested for controlling and powering

their linear actuators. One such product is a MultiMoto Arduino Shield, designed by

Robot Power as shown in Figure 3.4.

(a) Isometric View (b) Top View

Figure 3.4: Robot Power MultiMoto Arduino Shield [18]

35
According to Robot Power’s website, the MultiMoto features four independently con-

trolled and fully-reversible channels due to the use of the H-bridge design. Each

channel can handle 6.5A continuous on a 6-36V input voltage. With current and

over-temp limiting safety features, the general design and scope of the board is ap-

propriate for the environment and application for the Stewart Platform. However,

this board is two channels short of controlling the 6 linear actuators required on the

Stewart Platform. It will however provide to be a useful reference point in designing a

purpose-built driver. The schematics for the Robot Power Multimoto Arduino Shield

are included in Appendix E.

3.2.3 HexaMoto Design

The HexaMoto is designed to meet all the requirements set in Section 3.2.1. It uses

six L9958SBTR motor drivers, pin headers to interface with the Arduino Due, and

13 screw terminals for I/O. The specifics of these components are detailed in the

following subsection. The schematic for the HexaMoto may be found in Appendix F.

Investigating different PCB vendors, JLCPCB offered two-layer boards under 100

mm by 100 mm at $2 for quantity five. This drove the overall dimension of the

HexaMoto. By its own virtue, the shield must have pin headers that match with the

Arduino due, so these were placed about the center plane. The motor drivers and

their screw terminals were arranged around the bottom half of the board, and the

potentiometer screw terminals were placed around the top half perimeter.

Figure 3.5 shows the board layout with the ground planes deleted from the top and

bottom layer to illustrate the traces and polygons. Signal traces were routed with 16

mil traces, and power traces were routed with 24 mil to 32 mil traces. Traces on the

top of the board were primarily routed horizontally and traces on the bottom were

36
primarily routed vertically. This trace design helps keep traces organized and reduce

overlapping and space issues, which are common occurrences when routing the last

few traces of a board.

Figure 3.5: HexaMoto Board Layout, Top Layer

The board needs to support up to 5A per motor channel, which requires very large

traces to reduce resistance. Polygons are used instead of traditional traces to carry

this high current. The screw terminal at the center of the board takes the 12V input

from the power supply and distributes it to the motor drivers through the polygons,

which are between 50-100+ mils wide.

37
3.2.4 Component Details

3.2.4.1 STMicroelectronics L9958SBTR Motor Driver

The primary component of the HexaMoto Shield is the motor driver. The MultiMoto

uses the L9958 motor driver [21] and it contains the functionality required for driving

a Progressive Automation linear actuator. It is designed to drive brushed DC motors,

and can output up to 8.6A at 4V to 28V. Protection features include undervoltage

and overvoltage protection, and H-bridge over temperature and short circuits [22].

The block diagram and pinout of the selected PowerSO16 power package are shown

in Figure 3.6.

Figure 3.6: L9958 Block Diagram and PowerSO16 Pinout [22]

The L9958 motor driver is controlled by PWM, DIR, EN, and DI inputs. The EN

(Enable) and DI (disable) pins determine whether the bridge is set to Tri-state or

On-state. The DIR (Direction) and PWM (Pulse Width Modulation) pins control

the direction and speed of the connected motor. Serial peripheral interface (SPI) is

available for this motor driver, but its usage was not required and was not connected

on the motor driver. The application circuit in the L9958 manual is shown in Figure

3.7.

38
 Figure 3.7: L9958 Application Circuit [22]

All associated resistors and capacitors were included on the HexaMoto Shield. The

10nF capacitors on the outputs are decoupling capacitors, to reduce high frequency

noise to the outputs. The 100µF and 1µF capacitors connected to Vs are for decou-

pling to increase the robustness of the output short protection.

An LED was used to visually indicate the direction of each L9958 motor driver’s

output. The extension of the linear actuators is indicated by a green output. Retrac-

tion of the linear actuators are indicated by a red output. This was accomplished

by connecting the LED between OUTA and OUTB of the L9958 motor driver. This

LED is made by SunLED, and shown in Figure 3.8.

Figure 3.8: SunLED Bi-Directional LED [9]

39
3.2.4.2 I/O Hardware Components

Screw terminals were used for connecting the linear actuator wiring. 2-position screw

terminals were used for the input voltage from the DC power supply unit, discussed

in Section 3.4, and for the six outputs to the linear actuator motors. 3-position

screw terminals, as shown in Figure 3.9b, connect the linear actuators’ potentiometer,

comprised of the sensor, 3.3V, and GND. Pin headers are used to easily connect the

HexaMoto Shield to the Arduino Due. Standard 0.10 inch pin headers are shown

in Figure 3.9a. Note that the Arduino Due operates on +3.3VDC logic instead of

+5VDC, hence the difference in the voltage used for the potentiometer.

(a) Pin Header, 0.10 in Pitch (b) Screw Terminal, 3

[7] Position[8]

Figure 3.9: HexaMoto Shield I/O Hardware

3.3 HexaMoto Shield Assembly

The HexaMoto Shield assembly required soldering both through-hole and surface-

mount parts. Through-hole parts are typically easier to solder because they are larger,

easier to handle, and can be soldered on the opposite side of the board. Surface mount

parts can be more difficult, especially in smaller package sizes. For this application,

surface-mount components were soldered using solder paste, a stainless steel stencil,

40
and an oven. The stencil is a sheet of metal that is laser cut to match the solder pads

on the PCB. The solder paste is applied to the stencil and distributed, resulting in a

small and even deposit of solder paste on each pad. The components are placed on

the PCB and then the PCB is put into an oven, allowing the solder to reflow. Figure

3.10 shows the HexaMoto Shield after reflow and cooling. Please note that solder

paste can be irritating to the eyes, skin, and respiratory system. Make sure to handle

solder paste with care, avoiding contact with eyes, skin, and vapor inhalation. Use

isopropyl alcohol for solder paste cleanup.

Figure 3.10: HexaMoto Shield, Post Solder Paste Application and Reflow

Next, the capacitors, pin headers, and screw terminals were hand-soldered to the

HexaMoto Shield. These are through-hole parts that were not compatible with the

stencil. After soldering, the HexaMoto Shield was fully assembled, as shown in Figure

3.11. All traces were checked for shorts and continuity before any further integration.

41
 Figure 3.11: HexaMoto Shield, Completed

After the assembly of the HexaMoto Shield was complete, it was fit-tested on the

Arduino Due. The pin headers aligned and the shield fit on the Due successfully.

During the fit test, it became apparent that the DI pin header was connected to the

Arduino TX pin. The TX pin is the transmit pin for serial communication and while

it theoretically can be used as a general purpose digital I/O pin, it is not realistic in

practice. Therefore, the trace to the TX pin was broken and a jumper wire connected

it to an available pin header.

3.4 DC Power Supply

To function properly, each linear actuator must be properly powered. According to

Progressive Automations, each linear actuator draws a minimum of 1A during a no-

load condition, and can draw up to 5A of current at a maximum-load condition as

illustrated in Figure 3.12.

42
 Figure 3.12: PA-14P Current Draw vs Load [19]

A power supply unit was selected to support the maximum current draw of the

system. If all linear actuators encountered a simultaneous maximum-load condition,

then 30A would be drawn. A 12VDC, 360W power supply was chosen to supply the

necessary power. With a maximum output of 30A, the unit can sufficiently power the

robot. Additionally, the PSU has a built-in fan that always operates when powered.

This provides intake towards the rear of the unit, and exhausts heated air through

the front of the unit. Elevated temperatures were not noticed during extended cycle

times. The electrical bill of materials may be found in Appendix G.

3.5 Connectors

The power supply unit is connected to AC power through a fused power entry recep-

tacle. The receptacle has a Digikey part number 723W-X2/02 and is shown in Figure

3.13a. The live AC wire is connected in series through a switch before reaching the

power supply, allowing the Stewart Platform power to be toggled without having to

unplug. Additionally, the power receptacle has a fuse, increasing safety to users and

preventing damage to the robot if a short occurs.

43
The PC host connects to the Arduino with a USB-B cable that plugs into the USB-

B receptacle in the electronics enclosure. The receptacle is Digikey part number

MUSB-D511-00 and is shown in Figure 3.13b.

(a) Power Receptacle (b) USB-B Receptacle

Figure 3.13: Power and USB Receptacle Connectors Enclosure

The Arduino Due requires a micro-USB connection. A micro-USB cable was spliced

into the USB-B receptacle, as shown in Figure 3.14. This provides the intermediate

connection between the host PC and the Arduino Due.

Figure 3.14: MicroUSB to USB-B Connector

44
3.6 Electrical Assembly and Integration

The following details the progression of the electrical assembly and wiring of the

electronic enclosure. First, the PSU, switch, and power receptacle were inserted and

wired into the enclosure as shown in Figure 3.15.

Figure 3.15: Painted Electronic Enclosure with PSU and Connectors

Then, the Arduino Due and Hexamoto Shield were installed on the bracket. Once the

female Molex connectors were inserted into the enclosure, everything was ready for

wiring. The initial “rat’s nest” is shown in Figure 3.16a. Zip ties, wire routers and

wraps, and heat shrink were used to organize the electronics enclosure as shown in

Figure 3.16b. The acrylic base allows the internals of the electronics to be visible but

protected. Typically, the electronics is enclosed and hidden away. For professional

consumer products this is preferable but for an educational lab, students may enjoy

seeing the electronics “under the hood” of the Stewart Platform. Organized wiring

can make it easier to compare the system to the schematic for troubleshooting and

replacement purposes.

45
 (a) Beginning (b) Completion

Figure 3.16: Enclosure Wiring

Microcontrollers and PCBs can last for many years, especially if steps to properly

handle and care are taken. However, replacement of such components is a possibility.

A spare HexaMoto was assembled and tested and is ready for replacement. An

Arduino Due is an off-the-shelf component that is readily available online. Appendix

H details replacing these components of the Stewart Platform Electronics.

46
 Chapter 4

SOFTWARE DESIGN

4.1 Matlab Platform Simulation

During the design stage, a kinematic simulation of the Stewart Platform was created

using Matlab. This simulation uses the geometrical properties of the platform, and

given a predetermined motion profile, calculates the joint lengths required to perform

the move. Joint velocities are found by numerical differentiation of the joint lengths.

Platform position, orientation, joint lengths and their respective velocities are plotted

over time, with an accompanying 3D animation to visually verify the motion profile.

Figure 4.1 shows the Stewart Platform oscillating by 15° about ψ, 2 inches above and

below a Pz value of 20 inches, and increase linearly in θ from -15° to +15°.

Figure 4.1: Animation Screenshots

4.1.1 Geometrical Properties

The simulation takes in the geometrical properties that pertain to the kinematic

chains of the Stewart Platform to accurately calculate joint lengths. The geometry

comprises two of the known vectors, defined in Section 1.3.3. The first is the base

vector, which needs the distance and angle from the base center to each of the joint

47
connections. Similarly, the top vector is comprised of the distance and angle of the

joint connections on the top plate. The top vectors must be rotated according to the

orientation of the plate.

The simulation can be set for a true 6-6 configuration or 6-3 configuration. In a true

6-6 configuration, each joint on the base and top plate are spaced at 60°, which is

set automatically in the simulation. In reality, this configuration is unstable, per the

discussion in Section 1.3.1. Joints spaced other than 60° degrees apart will still be

considered 6-6 configurations but the angles are set manually in the simulation setup.

In the 6-3 configuration, there is an angle offset of the joints from the 120° centerline,

as discussed in Section 2.5. This modifies Equation 1.3 to become

   
B ∗cos((i−1)∗60−angle) B ∗cos((i−1)∗60+angle)
   
   
B ∗sin((i−1)∗60−angle) B ∗sin((i−1)∗60+angle)
Bi =   , Bi+1 =  (4.1)
   

0 0
   
   
   
1 1

for i = 1, 3, and 5 where angle is 3.47° for this Stewart Platform.

This joint angle difference is accounted for in the Matlab simulation as part of the

6-3 configuration initialization. Ultimately, the angles could be input manually into

the script with the angle offset already accounted for. However, the option of 6-

3 configuration makes for an improved user experience because the angle offset is

already accounted for. The Matlab simulation script can be found in Appendix I.

48
4.1.2 Motion Profile

The motion profile in the simulation is predetermined and set independently for each

degree of freedom. The duration of the motion is set in seconds. The number of

steps within that duration can be set depending on how coarse or fine the simulation

needs to be. Each of the six degrees of freedom can be set to a constant value, a

continuous function, or a piecewise-defined function to achieve the desired motion.

The motion profile used to create the animation in Figure 4.1 is included in Figure

4.2 as an example.

Figure 4.2: Motion Profile Creation

It is important to note that the motion profile should be continuous to avoid large

spikes in joint lengths and velocities that are not possible to achieve on the physical

Stewart Platform. For example, because the simulation uses discreet time steps, a

piecewise-defined function may orient the moving platform at θ = 20° at time t = 5

seconds and θ = −10° at t = 5.01 seconds. This would result in a calculated average

velocity of 3000°/sec about θ across this time interval, which is beyond the capabilities

of the Stewart Platform.

49
4.1.3 Calculation and Animation

The calculations for the joint lengths are completed in the Matlab simulation accord-

ing to equations 1.11 and 1.12 in Section 1.3.3. In order to achieve the animation, the

motion profile above (Figure 4.2) is linearly divided into hundreds of discreet steps

over a specified time span. The inverse kinematics are calculated for each step giving

a resolution equal to the motion profile time divided by the total number of steps.

For example, the following plots show the length and velocities of linear actuators

(Figure 4.3) for the given motion profile. This motion profile was specified to take 10

seconds and was calculated with 400 steps, for a resolution of 25 ms/step.

(a) Joint Lengths (b) Joint Velocitiess

Figure 4.3: Matlab Simulation Plots for Joint Values

The script also outputs plots for the platform position and orientation and the velocity

of each axis with respect to time. A plot of the angles of rotation about each axis

is shown in Figure 4.4. The plot matches the values of θ, φ, and ψ specified in the

motion profile.

50
 Figure 4.4: Matlab Simulation Plots for Orientation Values

Table 4.1 is provided as an example of these calculations for a Stewart Platform in

a 6-6 configuration with the following values: a = 9 in, b = 7 in, θ = 0°, φ = −30°,

ψ = 45°, Px = 0, Py = 0, Pz = 22 in. A 3D representation of these values is shown in

Figure 4.5. These calculations are done for every step of the animation.

Figure 4.5: 6-6 Configuration Model for Inverse Kinematics Solution

51
 Table 4.1: Inverse Kinematics Solution for 6-6 Configuration
Loop 1 2 3 4 5 6
ax 9.00 4.50 -4.50 -9.00 -4.50 4.50
ay 0 7.79 7.79 0 -7.79 -7.79
az 0 0 0 0 0 0
bx 7.00 3.50 -3.50 -7.00 -3.50 3.50
by 0 6.06 6.06 0 -6.06 -6.06
bz 0 0 0 0 0 0
bx rotated 4.29 -2.14 -6.43 -4.29 2.14 6.43
by rotated 4.29 6.43 2.14 -4.29 -6.43 -2.14
bz rotated 3.50 1.75 -1.75 -3.50 -1.75 1.75
dx -4.71 -6.64 -1.93 4.71 6.64 1.93
dy 4.29 -1.36 -5.65 -4.29 1.36 5.65
dz 25.50 23.75 20.25 18.50 20.25 23.75
d 26.28 24.70 21.11 19.57 21.35 24.49

4.2 Arduino Sketch

The Arduino code must fulfill 3 primary requirements: Manage I/O, communicate

over a serial connection, and run a control loop for the actuators. A state transition

diagram, shown in Figure 4.6 was created to lay out the design of the code to achieve

these design requirements.

First, the robot needs to establish serial communication with the host PC. Once this

connection is made, then the robot will perform an initial calibration and homing

procedure. Once successfully calibrated and homed, the code will wait for a com-

manded position for each actuator. When received, the Arduino Due will run the

control loop in the sketch and move the linear actuators.

For terminology reference, a sketch is the name for a program written for an Arduino.

The sketch is comprised of two special functions, setup() and loop(). Upon the start

52
 Figure 4.6: Arduino State Transition Diagram

of the sketch, setup() runs once and after completion, loop() is repeatedly called.

Custom functions can be declared in the sketch and called within loop().

4.2.1 Arduino Sketch Source Credit

When researching Progressive Automations and their linear actuators, it was discov-

ered that they sponsored a similar Stewart Platform project. The sponsored team

is from the University of British Columbia and worked on the project as part of an

Engineering Physics course. The team documented the project, including their code

for the Stewart Platform. Finding this code, which was released as open-source, pre-

sented two options; one option is to utilize the code, given proper credit, and adapt

it to the project; the alternative is to develop the code from scratch.

53
The code satisfied the design requirement set above. Rather than developing brand-

new code that would just result in near-identical function, the code was taken, with

proper credit, verified, and adapted to this particular Stewart Platform, and may be

found an Appendix J.

4.2.2 Setup and Initialization

The setup waits for the serial connection between the host PC and the NativeUSB

port before starting. Once connected, the direction, PWM, and potentiometer pins

are declared, and all outputs are set low. The control loop error is cleared to prevent

premature or unexpected actuation. This setup is followed by a calibration and

homing routine.

4.2.3 Calibration and Homing

After initialization, the robot first performs a homing and calibration procedure.

The procedure records the maximum and minimum potentiometer readings by fully

extending then retracting the platform. The Arduino Due ADC is 10 bits, so the range

of potentiometer values in software is 0-1023. When fully retracted, the potentiometer

readings on the actuators are between 80 and 90 counts; the readings are between 950

and 960 counts when fully extended. The calibration routine scales these raw readings

to the full 10-bit range of 0-1023. This eliminates potentiometer variations between

each actuator, enabling them to have an identical control loop. The calibration routine

also detects if the motor power is off and will return an error. When the calibration

routine is successful and the robot is fully retracted, it waits at its home position,

ready to receive position commands.

54
4.3 Matlab GUI

A graphical user interface (GUI) was created using Matlab App Designer to allow

users to easily connect to the robot and operate it. The user interface is shown in

Figure 4.7, and the source code may be found in Appendix K. This section will cover

the design choices and software flow of the app and its intended usage.

Figure 4.7: Matlab GUI for Stewart Platform

55
4.3.1 Serial Connection

Matlab first connects to the Arduino using serial communication over the COM port

specified by the GUI. Determining the correct COM port is detailed in the Lab Man-

ual. Once the correct COM port is entered, and the ’Calibrate’ button is clicked, the

Stewart Platform begins its calibration procedure. When the calibration completes,

the ’Calibrate’ button changes to ’Calibrated’. Only when the serial connection is

successful and calibration is complete can the joint lengths be calculated. This portion

of the GUI is shown in Figure 4.8.

Figure 4.8: Serial Connection and Calibration

4.3.2 Platform Configuration

The next step is to set up the geometrical configuration of the Stewart Platform.

The first option to select is whether the Stewart Platform is currently in a generic

configuration or a 6-3 configuration. If the 6-3 selection is configured, the location of

the top joints is automatically set to 0°, 120°, and 240°. This cannot be set by the

user because the top acrylic plate only allows for the steel balls to be placed together

at these locations. The generic configuration allows the user to set the locations of

56
the joints on the top plate. Both options allow the user to input the joint positions

on the base plate. This portion of the GUI is shown in Figure 4.9.

Figure 4.9: Top and Bottom Joint Locations

4.3.3 Linear Actuator Position Calculation

The actuator lengths can now be calculated with the correct joint locations. The

position and orientation of the platform is set by the user as desired. There are no

software limits on the position and orientation values. There are dozens of possible

configurations of the Stewart Platform, and each one has a different range of motion.

Students will investigate the range of motion for two of these configurations. This

portion of the GUI is shown in Figure 4.10.

Figure 4.10: Position and Orientation Input

57
Once the position and orientation of the Stewart Platform are given, the joint lengths

can be calculated. It is important to note that if the joint positions are not set

correctly, the length calculation will not be correct. At best, the robot will actuate

to an incorrect position. Worse cases include the robot actuating to an unstable

condition and collapsing. The results of the length calculation will be presented in

two forms. The first form is in units of inches. If the length of any actuator is

calculated to exceed its stroke to attain a certain position, an error will result. The

robot will not be allowed to move to this location or be saved.

4.3.4 Saving and Sequencing Points

The user has the option to save the current calculated position under the ’Save and

Sequence Setpoints’ panel shown in Figure 4.11. It can be saved in any of the six

options. If a saved setpoint is no longer desired, it can be cleared, and saved setpoints

can be overwritten. An option is provided to move to a setpoint after it has been

saved.

Figure 4.11: Save and Sequence Panel

58
4.3.5 Motion

This GUI panel, shown in Figure 4.12 contains the settings that alter the robot motion

and the buttons that command robot motion. By checking the ’Enable Path Planning’

button, the controller will divide the move into smaller moves specified in the ’Points’

text box. This path is normalized in each degree of freedom via interpolation.

Figure 4.12: Motion Control Panel

The ’Set Platform Speed’ slider changes the Stewart Platform’s speed by scaling the

PWM signal to the motor drivers. When at 100%, no scaling occurs and up to the

maximum voltage (12VDC) can be sent to the actuators, as calculated by the control

loop. If set to 50%, all PWM signals calculated by the control loop are scaled by

50% before being output. The ’Move to Current Setpoint’ button moves the Stewart

Platform to the position and orientation specified in the ’Position and Orientation’

panel by sending the actuator lengths from the ’Calculated Actuator Lengths’ panel

to the controller. The ’Move Through Sequence’ button will command the Stewart

Platform through up to six consecutive moves, from P1 to P6. The sequence will skip

any empty inputs, which may occur if a saved setpoint was cleared after being saved.

In other words, the only setpoints that will be in the sequence will have a check next

to them.

59
 Chapter 5

TESTING AND FUTURE WORK

5.1 Testing

Testing was performed on the linear actuators to ensure that they would achieve

the length commanded. Table 5.1 contains the data taken for each actuator as they

were commanded from 0 to 8 inches in 1 inch increments. The measurements were

taken with a tape measure with a resolution of 1/16” (0.063”), and measured values

interpolated to the nearest 1/32” (0.031”). The data shows that the actuators were

accurate to within 1/16 of an inch across the entire stroke.

Table 5.1: Commanded and Actual Actuator Lengths

Input Actuator 1 Actuator 2 Actuator 3 Actuator 4 Actuator 5 Actuator 6
0 0.00 0.00 0.00 0.00 0.00 0.00
1 1.00 1.00 1.00 1.00 1.00 1.00
2 2.00 2.00 2.00 2.00 2.00 2.00
3 3.00 3.00 3.00 3.00 3.00 3.00
4 4.00 4.06 4.03 4.03 4.03 4.03
5 5.00 5.06 5.06 5.03 5.03 5.03
6 6.03 6.06 6.06 6.03 6.03 6.03
7 7.06 7.06 7.06 7.03 7.03 7.03
8 8.06 8.06 8.06 8.03 8.03 8.03

The simulation script and its animation confirmed that the inverse kinematics equa-

tions are correct and working properly. Tests were performed on the Stewart Platform

to ensure that the inverse kinematics work on the actual system, and that values spec-

ified in the GUI were achieved within tolerance.

60
The position of the Stewart Platform was measured first. The height (Z) was simply

measured from the top surface of the base platform to the top surface of the moving

platform. This measurement was taken with a tape measure with markings every

1/16 inches. The moving platform position in the x-axis and y-axis was measured by

attaching a weighted pendulum to the top platform, which would make contact with

a ruled grid placed on the base platform below. The grid had markings every tenth

of an inch. Table 5.2 contains the nominal values input to the GUI and the actual

position achieved by the Stewart Platform.

Table 5.2: Stewart Platform Positional Test Measurements

X [in] Y [in] Z [in]
Input Measured Input Measured Input Measured
-4 -4.25 -4 -4.20 18 17.94
-3 -3.20 -3 -3.10 19 18.97
-2 -2.10 -2 -2.05 20 20.00
-1 -1.05 -1 -1.00 21 21.00
0 0.00 0 0.00 22 22.00
1 1.05 1 1.00 23 23.00
2 2.10 2 2.00 24 24.03
3 3.15 3 3.10
4 4.20 4 4.20

The platform was most accurate at the center of its positional range of motion. The

height of the platform was very accurate and had a maximum error of 1/16” at the

minimum height and 1/32” at the maximum height. The error along the x-axis and

y-axis increased fairly linearly, by approximately 0.05 inches per inch.

The angles of the Stewart Platform were also measured using an inclinometer app on a

smartphone. A 30-60-90 degree drafting triangle and a 45-90 degree drafting triangle

were used to verify the app readings at 30°, 45°, and 60°. The moving platform was

leveled with respect to gravity and then rotated by θ about the x-axis and by φ about

61
the y-axis; each axis was measured separately. The measurement resolution of the

inclinometer was 1°. Examples of these measurements are shown in Figure 5.1.

Figure 5.1: Inclinometer App Example Readings

A protractor with a resolution of 1° was used to measure ψ. The protractor was

zeroed to the moving platform, and then held in place independently as the platform

was rotated about the z-axis. Table 5.3 documents the input and measured values

for θ, φ, and ψ.

62
 Table 5.3: Stewart Platform Orientation Test Measurements
θ [degrees] φ [degrees] ψ [degrees]
Input Measured Input Measured Input Measured
-30 -30 -30 -31 -60 -61
-25 -24 -25 -26 -50 -51
-20 -19 -20 -20 -40 -41
-15 -14 -15 -15 -30 -30
-10 -9 -10 -10 -20 -20
5 -4 -5 -5 -10 -10
0 0 0 0 0 0
5 4 5 4 10 10
10 10 10 9 20 19
15 15 15 14
20 20 20 19
25 25 25 24
30 30 30 29

The orientation error of the platform was constant, and did not significantly increase

across the range of motion, unlike the positional error. The platform was generally 1°

less in magnitude in the −θ domain and +φ domain. The platform also slightly over-

rotated at the edge of the −φ and −ψ domains. The primary causes to these sources

of error can be attributed to geometry inaccuracies and residual control-loop error.

The nominal design values for joint-to-joint lengths, which include the spherical joints

and actuator geometry, and platform bolt-circle diameters were used in the equations.

These nominal values were checked on the actual platform for accuracy, but there was

variation up to 1/16 of an inch from kinematic chain to kinematic chain. The control-

loop error originates from the controls design. When the actuators are within a few

potentiometer counts, the effective voltage sent to the motors is very low, resulting

in no movement but an audible whine. If the actuator is within this threshold, the

PWM output from the Arduino is set to zero, eliminating the whine. This solution

introduces an error in the actuator lengths, up to 1/16” as demonstrated in Table

5.1, which will result in the small error in platform position and orientation.

63
5.2 Conclusion

A Stewart Platform was successfully designed, assembled, and implemented as part

of a lab experiment. It covers many facets of robots at an introductory level, in-

cluding inverse kinematics, simulation, assembly and disassembly, and motion imple-

mentation. Using the platform does not require a steep learning curve and performs

satisfactorily for student experimentation. The mechanical design allows for easy re-

configuration and uses the magnetic joints as the mode of failure, protecting the rest

of the system components. While the Stewart Platform is an overall success, it is

open to further improvements and future work.

5.3 Improvements

5.3.1 HexaMoto Shield

Overall, the Hexamoto Shield design proved successful but it can be improved. As

discussed in 3.3, the disable pin was connected to the TX pin, which is used in serial

communication. This was fixed on the current PCB revision by cutting the trace and

rerouting to another pin. Revision 2 of the Hexamoto Shield schematics F fixes this

design flaw. It is labeled ’A’ in Figure 5.2.

64
 Figure 5.2: HexaMoto Shield Improvements

The placement of screw terminals could also be changed to improve ease of assembly.

The labels 1-6 correspond to the linear actuator. The potentiometer and power

terminals are separated across the board. Pairing these terminals by mirroring across

the horizontal would make more sense, as only actuators 3 and 4 would be separated

across the board. During board design, pairing the power and potentiometer terminals

was attempted, but it was ultimately decided to keep the motor drivers together.

5.3.2 Acrylic Plates

Using acrylic plates for the Stewart Platform was appropriate when considering cost

and available manufacturing resources. In retrospect, it may have been justifiable

to use aluminum for the base and platform instead. The aluminum, while more

expensive, would be stronger and stiffer than the acrylic. Thicknesses between 1/8”

to 1/4” are recommended. Threaded inserts should be used in the top plate because

the steel spheres will be assembled and disassembled frequently when the Stewart

65
Platform is reconfigured per the lab manual. Aluminum threads can wear down

under these conditions.

The current angle spacing on the base and moving platforms is 20°. A spacing of 30°

may be more appropriate; 20° spacing offered a few too many configuration options,

and 30° spacing is more intuitive when using sine, cosine, and the x and y-axes.

Additionally, the angles increased in the clockwise direction from a top view. This

should be changed to counter-clockwise in order to follow the right-hand rule for

Cartesian coordinates. Currently, CW results in +Z in the downwards direction.

To achieve this, the material would have to be sourced and purchased. If using only

Cal Poly resources is desired then the aluminum should be cut with a water jet. Once

cut, the threads in the top platform would need to be tapped, ideally for threaded

inserts.

5.4 Future Work

A method of jogging the robot would be a helpful addition to the Stewart Platform.

This could be accomplished using buttons mapped to increasing or decreasing the

value of a degree of freedom. A joystick, a phone, or a 3D mouse (Figure 5.3) would

be particularly suited to controlling a Stewart Platform.

This device is able to sense in all 6 degrees of freedom, which matches the Stewart

Platform. Any user input to the 3D mouse would be matched by the moving platform.

The mouse can also sense magnitude. A light push or twist of the mouse could jog

the robot slowly, and larger twists etc. would result in faster jogging.

Incorporation of this device can be accomplished through either the Arduino Due or

the host PC. The 3D mouse can be plugged into the Native USB port on the Arduino

66
 Figure 5.3: 3Dconnexion 3D Mouse [2]

Due, which can act as a USB host for USB peripherals. Software would have to be

developed in order to parse the data from the mouse. The host PC could utilize

Matlab’s Space Mouse functionality in Simulink as another means of communicating

with the mouse. Either device would have to receive and interpret the 3D mouse

data and use it in a velocity-control loop. This device would be especially useful to

decrease teaching time of a Stewart Platform. For example, if a laser was attached

to the moving platform and had to be positioned and aligned with a receiver, the

3D mouse could quickly jog the robot to achieve this. Currently, the alternative is

to guess a numerical position and orientation in the GUI, move the platform, and

increment the numerical input until satisfied.

67
 Chapter 6

LAB MANUAL

6.1 Lab Manual

A lab manual was developed in conjunction with the Stewart Platform. The experi-

ment investigates inverse kinematics, reconfiguration, range-of-motion, and use of the

Stewart Platform as a flight simulator. The simulation script from Section 4.1 will be

used by students to verify their inverse kinematics. Next, the students will reassemble

the robot into two configurations, and determine the range-of-motion of each. Finally,

students will plan and run a movement sequence to simulate basic flight motion. The

full lab manual can be found in Appendix L.

6.2 Learning Objectives

The usage of the Stewart Platform in the lab will complement the parallel robots

material covered in lecture. The simulation provides students with the opportunity

to apply their understanding of inverse kinematics with low-risk. The accompanying

animation gives a glimpse into a Stewart Platform’s motion, and can show how catas-

trophe can result if the inverse kinematics are incorrect. Re-configuring the robot and

testing its range-of-motion highlights the various geometries for an SPS robot and the

resulting impact on its capabilities and performance. In addition, this allows students

to become familiar with the robot and the GUI. The flight simulation helps answer

the often asked question, “What is the point in learning this?” by using the Stewart

Platform in this common commercial application.

68
 BIBLIOGRAPHY

[1] AliExpress. KD418 Universal ball and socket joint, 2019. Accessed 8 October

2019.

[2] AmericanXplorer13. Space navigator. Web, March 2007. Accessed 5 February

2020.

[3] Arduino. ARDUINO DUE, 2019. Accessed 22 June 2019.

[4] K. J. Ayala. The 8051 microcontroller. Cengage Learning, 2004.

[5] P. Cruz, R. Ferreira, and J. Silva Sequeira. Kinematic modeling of

stewart-gough platforms. ICINCO, 2005.

[6] B. Dasgupta and T. Mruthyunjaya. The stewart platform manipulator: a

review. Mechanism and Machine Theory, 35(1):15–40, December 1998.

[7] Digi-Key. Molex 0022284060. Web, August 2019. Accessed 18 August 2019.

[8] Digi-Key. Molex 0398800303. Web, August 2019. Accessed 18 August 2019.

[9] Digi-Key. SunLED XZMDKVG55W-4. Web, August 2019. Accessed 18

August 2019.

[10] Firgelli Automations Team. Linear Actuators 101 - Everything you need to

know about Linear Actuators. Web, November 2018. Accessed 13 June

2019.

[11] IBS Magnet. Magnetic ball joints, 2019. Accessed 13 June 2019.

[12] K. Lancaster. AMiBA Radio Telescope. Web, June 2006. Accessed 10 January

2020.

69
[13] Matthewbarry. Eric gough’s tire testing machine. Web, March 2016. Accessed

10 January 2020.

[14] McMaster-Carr. Ball Joint Rod Ends Datasheet, 2019. Accessed 13 June 2019.

[15] J. Merlet. Parallel Robots. Springer Netherlands, P.O. Box 17, 3300 AA

Dordrecht, The Netherlands, 2006.

[16] S. B. Niku. Introduction to Robotics: Analysis, Control, Applications. John

Wiley and Sons, 2nd edition, 2011.

[17] S. B. Niku. Introduction to Robotics: Analysis, Control, Applications. John

Wiley and Sons, 3rd edition, 2020.

[18] Progressive Automations. Multimoto arduino shield. Web, August 2019.

Accessed 18 August 2019.

[19] Progressive Automations. PA-14P Datasheet, 2019. Accessed 13 June 2019.

[20] Progressive Automations. Resources. Web, August 2019. Accessed 19 August

2019.

[21] Robot Power. Multimoto product information. Web, August 2019. Accessed 18

August 2019.

[22] STMicroelectronics. L9958, December 2013. Accessed 20 August 2019.

[23] B. Vanderborght. Compliant robots. Web, January 2012. Accessed 10 October

2019.

[24] S. Vincent and J. Bridges. Positioning with air electropneumatic-positioning

systems bring extreme accuracy to high-speed automation. Web, April

2015. Accessed 13 June 2019.

70
 APPENDICES

Appendix A

PA-14P DATASHEET

PA-14P
Data Sheet

Contents

Specifications 2
Dimensions 3
Speed/Current vs Load 4
Connectors & Feedback 5
Mounting Brackets 6
Internal Components 7
Internal Descriptions 8

71
Specifications

Load (LBS) No Load Current (A) Full Load Current (A) Speed (inch/sec)
Dynamic Static 12VDC 24VDC 36VDC 48VDC 12VDC 24VDC 36VDC 48VDC No Load Full Load
35 75 1.0 0.5 0.3 0.3 5.0 2.5 1.7 1.3 2.00 1.38
50 100 1.0 0.5 0.3 0.3 5.0 2.5 1.7 1.3 1.14 0.83
75 150 1.0 0.5 0.3 0.3 5.0 2.5 1.7 1.3 0.95 0.70
110 220 1.0 0.5 0.3 0.3 5.0 2.5 1.7 1.3 0.79 0.59
150 300 1.0 0.5 0.3 0.3 5.0 2.5 1.7 1.3 0.37 0.28

Stroke 1" to 40"

Limit Switch Internal - Non-Adjustable
Limit Switch Feedback Customizable
Screw Type ACME Screw
Motor Type Brushed or Brushless DC Motor
Connector Type See Page 5
Wire Length 40" (customizable)
Housing Material 6062 Aluminum Alloy
Rod Material Aluminum Alloy/Stainless Steel (customizable)
Gear Material Polyformaldehyde (35 lbs only)/Powder Metallurgy Steel Alloy
Color (Shaft) Silver
Color (Motor End) Silver
Noise <45dB
Duty Cycle 25% (5 minutes on, 15 minutes off)
Operational Temperature -25ºC to 65ºC (-13ºF to 149ºF)
Protection Class IP54 (IP65 customizable)
Feedback Options Potentiometer (see page 5)
Certifications CE/RoHS
Mounting Brackets See Page 6
Mounting Ends Customizable

72
 Dimensions
(Dimensions in inches)

8.64

2.58
0.11 0.28

1.17
0.78

3.49
5.16

1.57

0.91
B
0.76 A 1.57
0.35 0.12 1.35
Ø0.78
Ø0.25
Ø0.25
0.81

1.40
0.70
0.59 R0.59
0.10 0.35
0.37 0.70 2.96
Ø0.39 0.85
0.75
1.50

R0.79
0.54
Ø1.50

0.04 0.12 0.65

Hole to Hole
Stroke 1 2 3 4 6 8 9 10 12 14 16 18 20 22 24 30 40
PA-14P A 6.51 7.51 8.51 9.51 11.51 13.51 14.51 15.51 17.51 19.51 21.51 23.51 25.51 27.51 29.51 35.51 45.51
B 7.51 9.51 11.51 13.51 17.51 21.51 23.51 25.51 29.51 33.51 37.51 41.51 45.51 49.51 53.51 66.61 85.51

For Stroke Length

A = Stroke Length + 5.51"
B = Stroke Length x 2 + 5.51"

73
Speed vs Load

Current vs Load

74
Connectors & Feedback

2-Pin Connector (Standard)

Motor
1 2
M- M+
1
Component Part Name Part Number Mating Part Number
2 Molex Mini Fit Jr 39-01-3029/
Housing 39-01-2025
2-Pin Receptacle 39-01-2026

Potentiometer Specifications
*For Stroke Length up to 40"

Signal

+5VDC gnd

Resistance* Number of Turns Tolerance

0-10kΩ 10 +/- 5%
*Actual resistance value may vary within the 0-10kΩ range based on stroke length

75
Mounting Brackets
(Dimensions in inches)

BRK-14
R0.32 Ø0.23 0.12

0.32

1.74 1.74
1.43
0.99
0.76 1.15 0.68
1.26 1.34

0.14
2.30 1.04
Ø0.10
0.18

0.57
Ø0.07 0.52
Ø0.32
0.36
Ø0.38
1.04

0.49 R0.18
Ø0.23

2.30

BRK-03
M4 x 16 Bolt
1.38

M4 Hex Nut

76
Internal Components

77
Internal Descriptions

Item Description Qty Item Description Qty

1 Actuator base 1 29 Limit Switches Spacer 1
2 Shaft Bearing 1 30 Limit Switches Base 1
3 Shaft Bearing Lock 1 31 Limit Switches Wiring 1
4 Shaft Base Spacer 1 32 Diode 2
5 Shaft Base Spacer Lock 1 33 Shaft Encloser Bottom Washer 1
6 Shaft Gear Wheel Holder 1 34 Shaft with Mounting Hole 1
7 Shaft Gear Wheel 1 35 Shaft Encloser 1
8 Base Cover with Mounting Support 1 36 Shaft Enclosure Top Cap 1
9 Electric Motor Gear Wheel 1 37 Power Cable 1
10 Small Intermediate Gear Wheel 1 38 Motor Enclosure Screw 2
11 Medium Intermediate Gear Wheel 1 39 Motor Screw Spring Washer 2
12 Teflon Washer 1 40 Motor Screw Washer 2
13 Electric Motor Base Washer 1 41 Shaft Eclosure Top Cap Screw 1
14 Electric Motor Base 1 42 Shaft Enclosure Base Screw 3
15 Brush Holder PCB 1 43 Brush Holder PCB Screw 2
16 Electric Motor Brush 2 44 Motor Base Screw 3
17 Electric Motor Brush Spring 2 45 Base Cover Washer 3
18 Electric Motor Rotor 1 46 Base Cover Screw 3
19 Motor Enclosure Bottom Washer 1 47 Base Extension 1
20 Electric Motor Encloser with Stator 1 48 Base Extension gasket 2
21 Motor Enclosure Top Washer 1 49 Base Extension Screw 2
22 Electric Motor Cap with Rotor Bearing 1 50 Potentiometer 1
23 Shaft Spacer 1 51 Potentiometer Bracket 1
24 Treaded Shaft Drive / Lead Screw 1 52 Potentiometer Bracket Screw 2
25 Shaft Base with Limit Switches Arm 1 53 Potentiometer Washer 3
26 Shaft Drive End Support 1 54 Potentiometer Nut 1
27 Shaft Drive End Support Screw 1 55 Potentiometer Gear 1
28 Limit Switch 2 56 Shaft Gear 1

78
 Appendix B

WAC20-4MM-4MM DATASHEET

Search:

 Navigation

PRODUCT INFORMATION

Part Description: WAC20-

4mm-4mm
Attachment Method: Integral
Clamp

Material Information
Finish: Clear
Anodize

Material: 7075-T6 VIEW CAD MODEL

Aluminum Alloy

REQUEST
QUOTE

Dimensional Information Torque

Outside Diameter: 20mm Momentary Dynamic: 1.3 Nm

Length: 28mm Non-Reversing: 0.65Nm

Reversing: 0.33Nm
Standard Bore Diameters Rated Speed: 10000 RPM
Major Bore Diameter: 4mm

Minor Bore Diameter: 4mm Misalignment

Angular Misalignment: 5 deg
Attachment Data Parallel Misalignment: .25mm
A1 Type: Cap Screw
Axial Motion: 0.25mm
A1 Screw Material: Alloy Steel

79
A1 Screw Size: M3x.5
Additional Speci cations
A1 Seating Torque (numeric): 2.0Nm
Torsional Rate (deg/Nm): 2.7
A1 Hex Wrench Size: 2.5mm degree/Nm
A1 Center Line (Metric): 3.8mm Maximum Operating
A2 Type: Cap Screw Temperature: 93C

A2 Screw Material: Alloy Steel

1. Dynamic torque ratings are momentary
A2 Screw Size: M3x.5
values. For non-reversing applications,
A2 Seating Torque (numeric): 2.0Nm divide by 2. Divide by 4 for reversing
applications. Should the torque ratings
A2 Hex Wrench: 2.5mm
be marginal for your application, contact
A2 Center Line (Metric): 3.8mm us for analysis.
2. Manufacturing dimensional tolerances
unless otherwise speci ed are:
Inch: Metric:
fraction x + .5 mm
+1/64
x.xx +.010 x.x + .25
mm

3. Inertia is based on smallest standard

bore diameter.
4. Keyways available in MC Series and W
Series with 40 & 50mm OD

PRINT

Products
Engineering
About Us
Request A Sample
Sales
News
Applications

Flexible Couplings
Beam Couplings

80
 Appendix C

MECHANICAL DRAWINGS

4.00

NOTES
UNLESS OTHERWISE SPECIFIED:
.75 1. ALL DIMENSIONS IN INCHES
2. TOLERANCES:
X.XX= 0.020
X.XXX= 0.010
ANGLES= 1
8-32 UNC 0.50
BOTH ENDS

Cal Poly Mechanical Engineering Section:01 Mat'l: Aluminum Title: LEG, BASE Drwn. By:TRENT PETERSON
ME 599 SPRING 2019 Dwg. #: 03 Nxt Asb:SGP-001 Date:6/28/2019 Scale: 1:1 Chkd. By: ME STAFF

SOLIDWORKS Educational Product. For Instructional Use Only.

81
 3X 6.94°
22X M4x0.7 THRU ALL
3X 3.47° .010 A B

13.00 B.C.

3
A

14.00

18X 20° NOTES

B UNLESS OTHERWISE SPECIFIED:
1. ALL DIMENSIONS IN INCHES
.25 2. TOLERANCES:
X.XX= 0.020
X.XXX= 0.010
ANGLES=+1 /-0
3. CREATE NUMBERING
3
THROUGH ENGRAVING
OR ETCHING PROCESS

Cal Poly Mechanical Engineering Section:01 Mat'l: Acrylic, 0.25 THK Title: TOP PLATE Drwn. By:TRENT PETERSON
ME 599 SPRING 2019 Dwg. #: 02 Nxt Asb:SGP-001 Date:6/28/2019 Scale: 1:4 Chkd. By: ME STAFF

SOLIDWORKS Educational Product. For Instructional Use Only.

82
 10.66 (TYP)
5.33 (TYP) 3
A

17.75

SCALE 1:8
17.00 B.C.

NOTES
UNLESS OTHERWISE SPECIFIED:
1. ALL DIMENSIONS IN INCHES
1.88 (TYP) 2. TOLERANCES:
3.76 (TYP) X.XX= 0.020
X.XXX= 0.010
ANGLES= 1
3. CREATE NUMBERING
3
THROUGH ENGRAVING
OR ETCHING PROCESS

10.00 (TYP)
6X 4-40 THRU
.010 A B

4X #14 THRU 18X 20° 18X 1/8 THRU

.020 A B .010 A B
5.00 (TYP)

10.00 (TYP) B

.25
Cal Poly Mechanical Engineering Section:01 Mat'l: Acrylic, 0.25 THK Title: BASE PLATE Drwn. By:TRENT PETERSON
ME 599 SPRING 2019 Dwg. #: 01 Nxt Asb:SGP-001 Date:6/28/2019 Scale: 1:4 Chkd. By: ME STAFF

SOLIDWORKS Educational Product. For Instructional Use Only.

83
 B
4.00

.50
2X .125 THRU
2.00 1.00
BOTH TABS A
.010 A B C

6X .125 THRU
.010 A B C

DOWN 90° R .13

DOWN 90° R .13

5.26
3.60 (TYP)

3.00

1.902

6.08

1.663 (TYP) .379 (TYP) NOTES

C UNLESS OTHERWISE SPECIFIED:
2.413 (TYP) 1. ALL DIMENSIONS IN INCHES
3.62 (TYP) 2. TOLERANCES:
X.XX= 0.020
X.XXX= 0.010
ANGLES= 1
3. BEND RADIUS TO NOT
EXCEED R=0.15

Cal Poly Mechanical Engineering Section:01 Mat'l: AL, #14-#18 GA Title: MOUNT, ARDUINO DUE Drwn. By:TRENT PETERSON
ME 599 SPRING 2019 Dwg. #: 04 Nxt Asb:ELCE-001 Date:6/28/2019 Scale: 1:2 Chkd. By: ME STAFF

SOLIDWORKS Educational Product. For Instructional Use Only.

84
 3

4 12X

5 6X

2
ITEM NO. PART NUMBER Default/
QTY.
1 ARDUINO DUE 1
2 MOUNT, ARDUINO 1
3 HexaMoto Shield 1
4 92196A106 (McMaster) 12
5 91780A427 (McMaster) 6
Cal Poly Mechanical Engineering Section:01 Mat'l: N/A Title: Arduino, Subassy Drwn. By:TRENT PETERSON
ME 599 SPRING 2019 Dwg. #:01 Nxt Asb: ELEC,ENCL Date:6/29/2019 Scale: 1:2 Chkd. By: ME STAFF

SOLIDWORKS Educational Product. For Instructional Use Only.

85
 C .756
7.00 .88
.559 .394
A .25 (TYP) 2.00 2.93
6X .38 1.39
.95 .220
3X 2.14
4.00 2X 2.83 3.30
.40
A .157 .063
3X .20
1.00 .50 3X .20 (TYP) 6X DETAIL A
3X .40 SCALE 2 : 1
NOTES
1.25 7.17 UNLESS OTHERWISE SPECIFIED:
9.50
11.25 1. ALL DIMENSIONS IN INCHES
2. TOLERANCES:
X.XX= 0.020
VIEW A X.XXX= 0.010
ANGLES= 1
3. ALL SLOT WIDTHS = 0.10
4. VIEW A DIMENSIONS APPLY TO OTHER TAB
5. ALL DRILLED HOLES ø1/8
3X 2.00
4X 1.00
B 2X .63 5.32
2X 2.83 .800

3X 1.76 .91
1.97
1.88 (TYP) 4X R.08
5.01 3.72
3X 1.48 1.07

.58
.144 (TYP) 4X R.05
2X 2.14
2X .20

Cal Poly Mechanical Engineering Section:01 Mat'l: AL, #14-#18 GA Title: ENCLOSURE, ELECTRONICS Drwn. By:TRENT PETERSON
ME 599 SPRING 2019 Dwg. #: 04 Nxt Asb:ELCE-001 Date:6/28/2019 Scale: 1:4 Chkd. By: ME STAFF

SOLIDWORKS Educational Product. For Instructional Use Only.

86
 UP 90° R .13

4.06 (TYP)
DOWN 90° R .13
DOWN 90° R .13

DOWN 90° R .13

18.94

DOWN 90° R .13

4.72 (TYP) NOTES

UNLESS OTHERWISE SPECIFIED:
1. ALL DIMENSIONS IN INCHES
UP 90° R .13 2. TOLERANCES:
X.XX= 0.020
X.XXX= 0.010
13.14 ANGLES= 1
3. BEND RADIUS TO NOT
FLAT PATTERN (REFERENCE) EXCEED R=0.15

Cal Poly Mechanical Engineering Section:01 Mat'l: AL, #14-#18 GA Title: MOUNT, ARDUINO DUE Drwn. By:TRENT PETERSON
ME 599 SPRING 2019 Dwg. #: 04 Nxt Asb:SGP-001 Date:6/28/2019 Scale: 1:2 Chkd. By: ME STAFF

SOLIDWORKS Educational Product. For Instructional Use Only.

87
 3

7 6X

1
ITEM NO. PART NUMBER QTY.
1 COVER, ELECTRONICS 1

2 12V 30A REGULATED 1

PSU
3 ArduinoSubAssy 1
9 12X 4 MUSB-D511-00-ND 1

5 723WX202 1

4 5 6 6 CR102J3RS215QF7 1
8 4X 7 76650-0076 6
8 91780A427 4
9 92196A106 12
Cal Poly Mechanical Engineering Section:01 Mat'l: N/A Title: ENCLOSURE, ELECTRONICS Drwn. By:TRENT PETERSON
ME 599 SPRING 2019 Dwg. #:01 Nxt Asb: SGP-001 Date:6/30/2019 Scale: 1:2 Chkd. By: ME STAFF

SOLIDWORKS Educational Product. For Instructional Use Only.

88
 4

5 6X

6 6X

1 6X

ITEM NO. PART NUMBER QTY.

1 PA-14P 6
2 KD418 12
3 Base Plate 1
4 Top Plate 1
12X 2 5 Helical_WAC20-4mm- 6
4mm
6X 7 6 94595A236 6

6X 8 7 91274A118 1
8 93475A230 1
9 Enclosure 1
10 4X
10 LEG, AL, 4 IN LG 4
3 9 11 4X
11 2515T12 4
Cal Poly Mechanical Engineering Section:01 Mat'l: N/A Title: STEWART GOUGH PLATFORM Drwn. By:TRENT PETERSON
ME 599 SPRING 2019 Dwg. #:01 Nxt Asb: Date:6/30/2019 Scale: 1:2 Chkd. By: ME STAFF

SOLIDWORKS Educational Product. For Instructional Use Only.

89
 Appendix D

MECHANICAL BILL OF MATERIALS

Bill of Materials Stewart-Gough Platform

Component - Stewart Platform Price Quantity Shipping/Taxes Total Cost Vendor Notes
KD418 Magnetic Ball Joint $ 4.75 12 $ 12.83 $ 69.83 https://www.aliexpress.com/item/KD418-Universal-ball-and-socket-joint-Steel-ball-Brass-rod-end-with-thread-hole
Long lead
PA-14P-8-35 Feedback Actuator $ 138.99 6 $ 12.98 $ 846.92 https://www.progressiveautomations.com/linear-actuator-with-potentiometer
50% Off with sponsorship
1/4 x 24 x 24 Acrylic Sheet $ 20.75 1 $ 20.75 https://freckleface.com/shopsite_sc/store/html/acrylicsheetonequarterinchthick.html
Base Plate
1/4 x 16 x 24 Acrylic Sheet $ 13.83 1 $ 13.90 $ 27.73 https://freckleface.com/shopsite_sc/store/html/acrylicsheetonequarterinchthick.html
Top Plate
WAC20-4mm-4mm $ - 6 $ - $ - http://heli-cal.com/product/wac20-4mm-4mm/
Sponsored
94595A236 Threaded Rod M4 x 0.7 mm, 12 mm Long $ 3.48 2 $ - $ 6.96 https://www.mcmaster.com/94595a236
Pack of 10, need 12
93475A230 M4 SST Washer $ 1.86 1 $ - $ 1.86 https://www.mcmaster.com/93475a230
Pack of 100
90591A141 M4 Nut $ 1.49 1 $ - $ 1.49 https://www.mcmaster.com/90591a141
Pack of 100
91274A118 M4x12mm SHCS $ 4.53 1 $ - $ 4.53 https://www.mcmaster.com/91274a118
Pack of 50
McMaster Taxes and Shipping $ - 0 $ 7.93 $ 7.93

$ 988.00 Total - Main Platform

$ 571.03 Total After Rebate

Component - Electronic Enclosure Hardware Price Quantity Shipping/Taxes Total Cost Vendor Notes
92196A106 4-40 x 0.25 SST SHCS $ 3.92 1 $ - $ 3.92 https://www.mcmaster.com/92196A106
Pack of 100
91780A427 4-40 x 13/32 Standoff $ 0.38 10 $ - $ 3.80 https://www.mcmaster.com/91780A427
Each
7565K65 Cable Holder $ 5.11 1 $ - $ 5.11 https://www.mcmaster.com/7565K65
Pack of 25
71535K51 Power Cord, 6 ft. $ 5.44 1 $ - $ 5.44 https://www.mcmaster.com/71535K51
Each
92141A005 4-40 Washer $ 1.40 1 $ - $ 1.40 https://www.mcmaster.com/92141A005
Pack of 100
8974K11 6061 0.75" Dia x 12" Lg $ 4.24 2 $ - $ 8.48 https://www.mcmaster.com/8974k11
1 ft length, Robot Feet (Qty 4, 5" Length each)
9284K313 1/8 ID Polyester Sleeving $ 2.08 1 $ - $ 2.08 https://www.mcmaster.com/9284K313
10 ft
7130K52 Zip Tie, 4" $ 2.99 1 $ - $ 2.99 https://www.mcmaster.com/7130K52
Pack of 100
McMaster Taxes and Shipping $ - 0 $ 9.74 $ 9.74

$ 42.96 Total - Electronics Enclosure (Hardware)

GRAND TOTAL (BEFORE SPONSORSHIP) $ 1,223.61

GRAND TOTAL (AFTER PROG. AUTO. 50% RETROACTIVE REBATE) $ 782.89

90
 Appendix E

MULTIMOTO SCHEMATIC

91
92
93
 Appendix F

HEXAMOTO SCHEMATIC

1 2 3 4 5 6

VIN
V+/2.3A
536 B260-13-F

1
2
3
X1-1

SMBJ30
C8 R14 D2 C7
A VIN A

D1
10uF 330uF

POT_1
GND
X1-2

10nF
C33
Power PWM X2-1
1 1
RESET PWM_1/2.1B +3.3V X2-2
+3.3V
2 PWM_2/2.3C
2
3 3 GND
VCC/2.1A PWM_3/2.4B GND/2.1C X2-3
GND/2.1C
4 PWM_4/3.1B
4
5 5

POT_2
GND/2.1C PWM_5/3.3C
6 6

GND

POT_4
VIN PWM_6/3.4A

GND
10nF
C34
X3-1 VCC/2.1A
J1 J3

10nF
C36
X5-1
+3.3V X3-2

R_LED
B B
SpiceOrder 1

+3.3V X5-2
Potentiometer Digital1 GND/2.1C X3-3
GND/2.1C X5-3 SpiceOrder 2

1 1

POT_3
POT_1 DIR_1/2.1B
2 2
GND

POT_5
POT_2 DIR_2/2.3C
3 3

GND
10nF
C35

POT_3 DIR_3/2.4B X4-1

4 4

10nF
C37
POT_4 DIR_4/3.1B X6-1

R13
536
5 5
POT_5 DIR_5/3.3C +3.3V X4-2
6 6
POT_6 DIR_6/3.4B +3.3V X6-2
7 7
ENABLE/2.1A GND/2.1C X4-3
8
2 8 GND/2.1C X6-3 ENABLE/2.1A
2 J2 J4

POT_6
GND

10nF
C38
C X7-1 C
+3.3V X7-2
GND/2.1C X7-3

D D

ME599
Notes: This schematic was designed by referencing the MultiMoto board by Robot Power (Copyright 2014). Project: Cal Poly SLO Stewart Robot
Its design was released under the Creative Commons Non-Commercial, Attribution, Share-Alike License.
This design modifies the MultiMoto board by expanding it to 6 channels, and adding 6 inputs for potentiometers.
This design is also released under the Creative Commons Non-Commercial, Attribution, Share-Alike License.
Title: HexaMoto Shield
Cal Poly
E Drawn By: TRENT PETERSON E
Mechanical Engineering
Date: 1/22/2020 23:39 San Luis Obispo, CA

Rev: 2 Sheet: 1/3

HexaMoto_R2 .sch
Document Number: TRP-001

94
 1 2 3 4 5 6
VCC/1.6B C16 C17
VCC/1.6B C4 C5
V+/1.6A 100nF 10uF V+/1.6A
100nF 10uF
C2 C3 C14 C15
GND
C1 C13
A IC1 GND 100uF 1uF IC3 100uF 1uF A
16 100nF 16 100nF
VDD VDD
13 13
VS GND VS GND
4 4
CP CP
9 6 9 6
EN OUT1 EN OUT1
7 10 C39 7 10 C43
ENABLE/1.6C DI OUT2 ENABLE/1.6C DI OUT2
11 M1-1 11 M3-1
PWM_1/1.2A PWM PWM_3/1.2B PWM
BI_LED1 2.1k

BI_LED3 2.1k
5 10nF 5 10nF
R2

R6
DIR_1/1.2B DIR DIR_3/1.2B DIR
GND

GND
M1-2 M3-2
3 3
CS SpiceOrder 1
MKDSN1,5/2-5,08 CS SpiceOrder 1
MKDSN1,5/2-5,08
15 10nF 15 10nF
SCK SCK
14 14
SI SI
2
SO C40 2
SO C44
B B
SpiceOrder 2 SpiceOrder 2

12 12
REXT REXT
8 8
PGND PGND
GND

GND
1 1
GND GND
17 17
EXP EXP
VCC/1.6B C11 C12
10k

10k
L9958SBTR L9958SBTR
R1

R5
V+/1.6A
100nF 10uF
C9 C10
C6
IC2 GND 100uF 1uF
16 100nF
VDD
13
C VS
CP
4
GND C
9 6
EN OUT1
7 10 C41
ENABLE/1.6C DI OUT2
11 M2-1
PWM_2/1.2A PWM
BI_LED2 2.1k

5 10nF
R4

DIR_2/1.2B DIR
GND

M2-2
3
CS SpiceOrder 1
MKDSN1,5/2-5,08
15 10nF
SCK
14
SI
2
SO C42 SpiceOrder 2

12
REXT
8
D PGND D
GND

1
GND
17
EXP
10k

L9958SBTR
R3

ME599
Notes: Project: Stewart Robot

Title: TRP_MOTOR_DRIVER
Cal Poly
E Drawn By: TRENT PETERSON E
Mechanical Engineering
Date: 1/22/2020 23:39 San Luis Obispo, CA

Rev: 1 Sheet: 2/3

HexaMoto_R2 .sch
Document Number: TRP-001

95
 1 2 3 4 5 6
VCC/2.4A C31 C32
VCC/2.4A C21 C22 V+/2.6A
V+/2.6A 100nF 10uF
100nF 10uF C29 C30
C19 C20 C28
IC6 GND 100uF 1uF
C18
A IC4 GND 100uF 1uF 16 100nF A
100nF VDD
16 13
VDD VS GND
13 4
VS GND CP
4
CP
9 6
EN OUT1
9 6 7 10 C49
EN OUT1 ENABLE/2.4A DI OUT2
7 10 C45 11
ENABLE/2.4A DI OUT2 PWM_6/1.2B PWM M6-1

BI_LED6 2.1k
R12
11 5 10nF
PWM_4/1.2B PWM M4-1 DIR_6/1.2C DIR
BI_LED4 2.1k

10nF

GND
5
R8

DIR_4/1.2B DIR M6-2

GND

M4-2 3
CS SpiceOrder 1
MKDSN1,5/2-5,08
3 15 10nF
CS SpiceOrder 1
MKDSN1,5/2-5,08 SCK
15 10nF 14
SCK SI
14
SI
2
SO C50 SpiceOrder 2

2
SO C46
B B
SpiceOrder 2

12
REXT
12
REXT
8
PGND

GND
8 1
PGND GND
GND

1 17
GND EXP
17
EXP

R11
10k
VCC/2.4A L9958SBTR
C26 C27
10k

L9958SBTR
R7

V+/2.6A
100nF 10uF
C24 C25
C23
IC5 GND 100uF 1uF
16 100nF
VDD
13
C VS
CP
4
GND C
9 6
EN OUT1
7 10 C47
ENABLE/2.4A DI OUT2
11 M5-1
PWM_5/1.2B PWM
BI_LED5 2.1k
R10

5 10nF
DIR_5/1.2C DIR
GND

M5-2
3
CS SpiceOrder 1
MKDSN1,5/2-5,08
15 10nF
SCK
14
SI
2
SO C48 SpiceOrder 2

12
REXT
8
D PGND D
GND

1
GND
17
EXP
10k

L9958SBTR
R9

ME599
Notes: Project: Stewart Robot

Title: TRP_MOTOR_DRIVER
Cal Poly
E Drawn By: TRENT PETERSON E
Mechanical Engineering
Date: 1/22/2020 23:39 San Luis Obispo, CA

Rev: 1 Sheet: 3/3

HexaMoto_R2 .sch
Document Number: TRP-001

96
 Appendix G

ELECTRICAL BILL OF MATERIALS

Digikey Components
Qty Ordered Cost (ea) Total Digikey PN MFGR PN Qty Value Device Package Parts Description
BI_LED1, BI_LED2, BI_LED3, BI_LED4,
10 $ 0.40 $ 3.95 1497-1314-1-ND XZMDKVG55W-4
7 LED_E LED_1206 BI_LED5, BI_LED6, R_LED LED
1 $ 0.16 $ 0.16 WM20124-ND 423751856 1 MA03-1 MA03-1 VIN PIN HEADER
C1, C4, C6, C11, C13, C16, C18, C21,
20 $ 0.13 $ 2.56 478-1395-1-ND 08055C104KAT2A
12 100nF C-USC0805K C0805K C23, C26, C28, C31 CAPACITOR, American symbol
10 $ 0.15 $ 1.53 399-6546-ND ESH107M025AE3AA 6 100uF C-USC0805K C0805K C2, C9, C14, C19, C24, C29 CAPACITOR, American symbol
10 $ 0.14 $ 1.40 RNCF0805DTE10K0CT-ND RNCF0805DTE10K0 7 10k R-US_M0805 M0805 R1, R3, R5, R7, R9, R11, R13 RESISTOR, American symbol
C33, C34, C35, C36, C37, C38, C39, C40,
25 $ 0.09 $ 2.20 478-10823-1-ND 08051C103K4T4A C41, C42, C43, C44, C45, C46, C47, C48,
18 10nF (10k pF) C-USC0805K C0805K C49, C50 CAPACITOR, American symbol
14 $ 0.12 $ 1.67 1276-1096-1-ND CL21A106KOQNNNE 7 10uF C-USC0805K C0805K C5, C8, C12, C17, C22, C27, C32 CAPACITOR, American symbol
10 $ 0.06 $ 0.60 1276-1026-1-ND CL21B105KOFNNNE 6 1uF C-USC0805K C0805K C3, C10, C15, C20, C25, C30 CAPACITOR, American symbol
10 $ 0.08 $ 0.82 P2.10KCCT-ND ERJ-6ENF2101V 6 2.1k R-US_M0805 M0805 R2, R4, R6, R8, R10, R12 RESISTOR, American symbol
3 $ 0.73 $ 2.19 565-1693-ND EKZE350ELL331MJ16S 1 330uF CPOL-USTT5D10 TT5D10 C7 POLARIZED CAPACITOR, American symbol
2 $ 0.38 $ 0.76 P19275CT-ND ERJ-6DQF2R0V 1 2 R-US_M0805 M0805 R14
2 $ 0.36 $ 0.72 P20563CT-ND ERJ-PB6B5360V 1 536 R-US_M0805 M0805 R13 RESISTOR, American symbol
2 $ 0.42 $ 0.84 B260-FDICT-ND B260-13-F 1 B260-13-F ZENER-DIODESMB SMB D2 Z-Diode
3 $ 0.35 $ 1.05 WM50014-08-ND 22284080 2 Digital1, Pot MA08-1 MA08-1 J4, J2 PIN HEADER
7 $ 5.46 $ 38.22 497-11667-1-ND L9958SBTR 6 L9958SBTR L9958SBTR SOP80P1410X360-17N IC1, IC2, IC3, IC4, IC5, IC6 L9958 Series H-Bridge - PowerSO16
10 $ 0.59 $ 5.93 WM4393-ND 398800302 7 MKDSN1,5/2-5,08 MKDSN1,5/2-5,08 MKDSN1,5/2-5,08 X1, M1, M2, M3, M4, M5, M6 MKDSN 1,5/ 2-5,08 Printklemme
10 $ 0.75 $ 7.48 WM6289-ND 398800303 6 MKDSN1,5/3-5,08 MKDSN1,5/3-5,08 MKDSN1,5/3-5,08 X2, X3, X4, X5, X6, X7 MKDSN 1,5/ 3-5,08 Printklemme
3 $ 0.26 $ 0.78 WM50014-06-ND 22284060 2 PWM, Power MA06-1 MA06-1 J1, J3 PIN HEADER
2 $ 0.54 $ 1.08 SMBJ30ABCT-ND SMBJ30A 1 SMBJ30 ZENER-DIODESMB SMB D1 Z-Diode
6 $ 1.26 $ 7.56 WM8369-ND ‎76650-0076 KIT CONN MINI-FIT JR 6 CIRCUITS
The right hand side above this note is the minimum BOM required to make one Motor Driver PCB. The quantity
1 $ 2.53 $ 2.53 Q853-ND 3025010-03 USB A MALE TO MICRO B MALE 3'
ordered (top left) includes a reasonable spare parts, a best design practice for PCB creation. Spare part total (~$10)
1 $ 2.33 $ 2.33 Q361-ND 3021001-03 CBL USB A-B CON 3' 28/28 AWG
1 $ 1.52 $ 1.52 Q204-ND 723W-X2/02 PWR ENT MOD RCPT IEC320-C14 PNL
1 $ 11.28 $ 11.28 MUSB-D511-00-ND MUSB-D511-00 <----Additional Digikey Parts (Enclosure Electronic Hardware) ----> CONN RCPT USB2.0 TYPEB 4POS PCB
5 $ 0.23 $ 1.15 507-1254-ND 5ST 6.3-R FUSE GLASS 6.3A 250VAC 5X20MM

Merchandies Total $ 100.31

Shipping $ 4.99
Taxes $ 7.77
Digikey Total $ 113.07

Components - Electronics
Price Quantity Shipping Total Cost Vendor Notes
Arduino Due $ 47.50 1 $ - $ 47.50 https://www.progressiveautomations.com/products/lc-062
Dept Order, 50% Off with sponsorship

12V DC 30A Regulated

Power Supply $ 19.50 1 $ - $ 19.50 https://www.ebay.com/itm/231849834132
Motor Driver - PCB $ 0.40 5 $ - $ 2.00
Motor Driver - Solder
Stencil $ 6.00 1 $ 4.58 $ 10.58 https://jlcpcb.com/

Total - Electronics (Non-

Digikey) $ 79.58
Total After Rebate $ 55.83

97
 Appendix H

REPLACEMENT MANUAL

STEWART-GOUGH PLATFORM

REPLACEMENT GUIDE- ELECTRONICS

INTRODUCTION:
This is an instruction manual that will detail how to replace the HexaMoto Arduino Shield and
the Arduino Due.

SOFTWARE STEPS (Arduino Due):

1. Download the Arduino IDE from https://www.arduino.cc/en/main/software by using the
“Windows Installer, for Windows XP and up” link on the page. Run the executable and
complete the installation.

2. Connect to the Arduino Due with a PC with the Arduino IDE installed.

3. The program will launch, and the loading screen should appear.

4. Once the program is launched, hover over the Include Library selection under the Sketch
tab. Select the Add .ZIP library… option that appears and select the provided
StewartPlatformLibrary.zip file.

5. Ensure that the Arduino Due boards are loaded in the IDE. If they are, they will be
located under the Board: … option under the Tools tab at the bottom of the list. If they
are not present, they can be added with the Boards Manager.

98
Install the Arduino SAM Boards (32-bits ARM Cortex-M3) which will allow the Arduino
Due board to be selected once installed.

99
6. Load the Stewart Platform sketch into the Arduino IDE. The program should look similar
to the following.

7. Ensure the device selected is an Arduino Due and that the COM Port is correct. The
COM Port may be found in Device Manager in Windows. The Arduino Due has two
USB ports: Native and Programming. For this manual’s purpose, the Native port has a
much faster serial connection allowing for quick uploads, but the Programming port is
more robust when uploading sketches. Either should work, as long as the Arduino IDE
selection of Native or Programming port matches the connection made with the Arduino
Due.

100
8. Once the sketch is loaded and the Arduino IDE is properly configured to upload to an
Arduino Due, use the upload tool to put the sketch on the Arduino Due. The IDE should
show the sketch compiling, then reading/writing to the Due, followed by a completion
message.

9. The Arduino Due may be simply unplugged from the PC.

101
HARDWARE STEPS:

1. Unplug the USB and power cables from the ports in the front of the electronics platform.
Unscrew the six (6) fasteners that attach the electronics enclosure to the bottom platform and
lower it onto the working surface.

2. Unplug the six (6) actuator connectors from the electronics enclosure.

102
3. Remove the HexaMoto Arduino shield from the Arduino Due. The shield is connected by
pins. Great care should be taken to not bend the pins upon removal. This is best
accomplished by grabbing the sides of the shield, and slightly rocking it back and forth and
left and right, as depicted by the arrows, to slowly lift the shield. To install a new HexaMoto,
align the shield’s pins with the Due’s female pin headers and insert, with care not to bend the
pins. Complete these steps in reverse order to conclude the Hexamoto replacement. If
replacing the Arduino Due, continue to Step 4.

103
4. Remove the screws that attach the Arduino Due to the sheet metal mount. Note, due to hole
alignment, not all 6 screws may be present. Unplug the USB cable from the Native USB
Port, pointed to by the arrow.

5. Replace the old Arduino Due with a new Arduino Due that has the Stewart Platform code
loaded. Ensure that the USB cable is plugged back into the Native USB Port. Follow all steps
in reverse order to reassemble the electronics enclosure.

104
 Appendix I

MATLAB SIMULATION SCRIPT

Table of Contents
Stewart Gough Platform Equations of Motion .......................................................................... 1
Clear Previous .................................................................................................................... 1
Set Constants ...................................................................................................................... 1
Motion Profile Creation ....................................................................................................... 2
Calculate EOMs .................................................................................................................. 4
Derivative Calculation .......................................................................................................... 5
Plots ................................................................................................................................. 5
Animation ........................................................................................................................ 10

Stewart Gough Platform Equations of Motion

This script will take in, in variable form, the dimensions of the bottom and top of the robot, a and b
respectively, the values of the theta, phi, and psi, respective to the x, y, and z axis respectively, and the
position vector P, and output the link lengths of each linear actuator. The outputs are then simulated for
a predetermined motion profile and various metrics are

%By Trent Peterson

Clear Previous
close; clc; clear;

Set Constants
base_diameter = 18; %[inches], acrylic
platform bolt circle
top_diameter = 14; %[inches], acrylic
platform bolt circle
a = (base_diameter)/2; %[inches], distance
from center of base to actuator
b = (top_diameter)/2; %[inches], distance
from center of top to actuator

%Set base and top angles

base_angle = [ 0 60 120 180 240 300 ]; %[degrees], Spacing of
actuator pairs, base
base_angle = sort(base_angle); %sorted for
calculation integrity

top_angle = [ 0 40 120 160 240 280 ]; %[degrees], Spacing of

actuator pairs, top
top_angle = sort(top_angle); %sorted for
calculation integrity

% Set configuration
% 6-3 Configuration when the base joints are spaced 60° and the top
joints

105
 % pairs are spaced ~120° (Physical joint angle accounted for)
% 6-6 Configuration when all joints are spaced 60° (unstable
configuration)
configuration_6_3 = false; % Run preset 6-3
configuration simulation
configuration_6_6 = false; % Run preset 6-6
configuration simulation
animation = false; % Display Animation
plots = true; % Display Plots
%Set physical parameters
stroke_length = 8; %[inches]
joint_angle = 3.47; %[degrees], for 6_3
configuration
min_joint_length = 16.7; %[inches], from CAD
max_joint_length = 24.7; %[inches], from CAD
min_joint_velocity = -2; %[inches/sec], from
Progressive Automations
max_joint_velocity = 2; %[inches/sec], from
Progressive Automations
% Set up motion variables
n= 1:400; % Array dimension
[theta,phi,psi,Px,Py,Pz]=deal(zeros(1,length(n))); % Create position
array
time_duration= 10; % Seconds
time = 0:(time_duration/(max(n)-1)):time_duration; % Create time
array, position
time_dot = time; % Create time
array, velocity
time_dot(end)=[]; % Delete array
end, discrete differentiation
motion_profile = 2; % Choose motion
profile

% Preallocation, create joint locations and distance arrays

a_i = zeros(4,6,length(theta));
b_i = zeros(4,6,length(theta));
b_rot = zeros(4,6,length(theta));
L_vector = zeros(4,6,length(theta));
L_length = zeros(length(theta),6);
L_length_dot = L_length;
L_length_dot(length(n),:)=[];

Motion Profile Creation

Create function based motion per degree of freedom (independently) based on the time duration over n
steps. All motions are allowed. Resulting plots will determine if move profile is valid or not

if motion_profile == 1
for m = 1:length(n)
theta(m) = -30 + 30 * m/max(n); %[degrees], relative
to x axis
phi(m) = 0; %[degrees], relative
to y axis

2
106
 psi(m) = -30 + 60 * m/max(n); %[degrees], relative
to z axis
Px(m) = 0; %[inches], x position
Py(m) = 0; %[inches], y position
Pz(m) = 20 + 2*cosd(m/max(n)*360); %[inches], z position
end
end

if motion_profile == 2
for m = 1:length(n)
theta(m) = 10; %[degrees], relative
to x axis
phi(m) = 10*cosd(m/max(n)*360); %[degrees], relative
to y axis
psi(m) = 0; %[degrees], relative
to z axis
Px(m) = 3 * m/max(n); %[inches], x position
Py(m) = 0; %[inches], y position
Pz(m) = 22 - 3*m/max(n) ; %[inches], z position
end
end

if motion_profile == 3
for m = 1:length(n)
theta(m) = 0; %[degrees], relative
to x axis
phi(m) = 10; %[degrees], relative
to y axis
psi(m) = -15 + 25 * m/max(n); %[degrees], relative
to z axis
Px(m) = 3; %[inches], x position
Py(m) = -2 + 3 * m/max(n); %[inches], y position
Pz(m) = 20; %[inches], z position
end
end

if motion_profile == 4
for m = 1:length(n)
theta(m) = -15 + 30 * m/max(n); %[degrees], relative
to x axis
phi(m) = 10 - 5 * m/max(n); %[degrees], relative
to y axis
psi(m) = 15*sind(m/max(n)*360); %[degrees], relative
to z axis
Px(m) = 1.5; %[inches], x position
Py(m) = -3 + 4*m/max(n); %[inches], y position
Pz(m) = 19.8 + 2*cosd(m/max(n)*360);%[inches], z position
end
end

P = [Px;Py;Pz;zeros(1,length(theta))]; %combine into array

3
107
Calculate EOMs
if configuration_6_6
for j = 1:length(theta)
for i = 1:6
%calculate a vector for each base location
a_i(1,i,j)=a*cosd((i-1)*60);
a_i(2,i,j)=a*sind((i-1)*60);
a_i(3,i,j)=0;
a_i(4,i,j)=1;
%calculate b vector for each top location
b_i(1,i,j)=b*cosd((i-1)*60);
b_i(2,i,j)=b*sind((i-1)*60);
b_i(3,i,j)=0;
b_i(4,i,j)=1;
%convert b from global to rotated

b_rot(:,i,j)=stewartrot(theta(j),phi(j),psi(j),b_i(:,i,j));
L_vector(:,i,j) = P(:,j) + b_rot(:,i,j) - a_i(:,i,j);
L_length(j,i) = sqrt(L_vector(1,i,j)^2 + L_vector(2,i,j)^2
+ L_vector(3,i,j)^2);
end
end

elseif configuration_6_3
for j = 1:length(theta)
%calculate b vector for each top location
for i = [1 3 5]
b_i(1,i,j)=b*cosd((i-1)*60-joint_angle);
b_i(2,i,j)=b*sind((i-1)*60-joint_angle);
b_i(3,i,j)=0;
b_i(4,i,j)=1;
%------------------------------------------
b_i(1,i+1,j)=b*cosd((i-1)*60+joint_angle);
b_i(2,i+1,j)=b*sind((i-1)*60+joint_angle);
b_i(3,i+1,j)=0;
b_i(4,i+1,j)=1;
end

for i = 1:6
%calculate a vector for each base location
a_i(1,i,j)=a*cosd((i-1)*60);
a_i(2,i,j)=a*sind((i-1)*60);
a_i(3,i,j)=0;
a_i(4,i,j)=1;

%convert b from global to rotated

b_rot(:,i,j)=stewartrot(theta(j),phi(j),psi(j),b_i(:,i,j));
L_vector(:,i,j) = P(:,j) + b_rot(:,i,j) - a_i(:,i,j);
L_length(j,i) = sqrt(L_vector(1,i,j)^2 + L_vector(2,i,j)^2
+ L_vector(3,i,j)^2);
end

4
108
 end

else %for all non-true 6-6 and 6-3 cases

for j = 1:length(theta)
for i = 1:6
a_i(1,i,j)=a*cosd(base_angle(i));
a_i(2,i,j)=a*sind(base_angle(i));
a_i(3,i,j)=0;
a_i(4,i,j)=1;
%------------------------------------------
b_i(1,i,j)=b*cosd(top_angle(i));
b_i(2,i,j)=b*sind(top_angle(i));
b_i(3,i,j)=0;
b_i(4,i,j)=1;

b_rot(:,i,j)=stewartrot(theta(j),phi(j),psi(j),b_i(:,i,j));
L_vector(:,i,j) = P(:,j) + b_rot(:,i,j) - a_i(:,i,j);
L_length(j,i) = sqrt(L_vector(1,i,j)^2 + L_vector(2,i,j)^2
+ L_vector(3,i,j)^2);
end
end
end

Derivative Calculation
[theta_dot,phi_dot,psi_dot,Px_dot,Py_dot,Pz_dot,P_dot]=deal(zeros(1,length(n)-1)); %pr

for i=2:length(n)

Px_dot(i-1) = (Px(i) - Px(i-1))/(time(i)-time(i-1));

Py_dot(i-1) = (Py(i) - Py(i-1))/(time(i)-time(i-1));
Pz_dot(i-1) = (Pz(i) - Pz(i-1))/(time(i)-time(i-1));
P_dot(i-1) = sqrt((Px_dot(i-1))^2 + (Py_dot(i-1))^2 +
(Pz_dot(i-1))^2);
theta_dot(i-1) = (theta(i) - theta(i-1))/(time(i)-time(i-1));
phi_dot(i-1) =(phi(i) - phi(i-1))/(time(i)-time(i-1));
psi_dot(i-1) =(psi(i) - psi(i-1))/(time(i)-time(i-1));

for j = 1:6
L_length_dot(i-1,j) = (L_length(i,j) - L_length(i-1,j))/
(time(i)-time(i-1));
end
end

Plots
if plots
figure(1)
hold on
for i=1:6
plot(time,L_length(:,i))

5
109
 end

xlabel('Time (seconds)')
ylabel('Length (inches)')
plot(time, min_joint_length*ones(1,length(time)),'--k', time,
max_joint_length*ones(1,length(time)),'--k','LineWidth',2)
legend('Actuator 1', 'Actuator 2', 'Actuator 3', 'Actuator
4', 'Actuator 5', 'Actuator 6', 'Min Length', 'Max
Length', 'Location', 'north')
title('Joint Lengths, from Sphere to Sphere')

hold off

figure(2)
hold on
for i=1:6
plot(time_dot,L_length_dot(:,i))
end
plot(time, min_joint_velocity*ones(1,length(time)),'--k', time,
max_joint_velocity*ones(1,length(time)),'--k', 'LineWidth',2)
xlabel('Time (seconds)')
ylabel('Velocity (inches/sec)')
ylim([-2.5 2.5])
% plot(time, min_joint_length*ones(1,length(time_dot)), time,
max_joint_length*ones(1,length(time)))
legend('Actuator 1', 'Actuator 2', 'Actuator 3', 'Actuator
4', 'Actuator 5', 'Actuator 6', 'Min Length', 'Max
Length', 'Location', 'northwest')
title('Joint Velocities')

hold off

figure(3)
plot(time, Px, time, Py, time, Pz)
xlabel('Time (seconds)')
ylabel('Coordinate Value (inches)')
legend('X Centroid', 'Y Centroid', 'Z
Centroid', 'Location', 'east')
title('Position of Centroid P')

figure(4)
plot(time, theta, time, phi, time, psi)
xlabel('Time (seconds)')
ylabel('Angle of Rotation (degree)')
legend('Theta (about X)', 'Phi (about Y)', 'Psi (about Z)')
title('Angle of Rotation about Respective Axis')

figure(5)
plot(time_dot, Px_dot, time_dot, Py_dot, time_dot, Pz_dot,
time_dot, P_dot)
xlabel('Time (seconds)')
ylabel('Coordinate Value Velocity (inches/sec)')
legend('X Centroid', 'Y Centroid', 'Z Centroid', 'Centroid
Velocity', 'Location', 'southeast')

6
110
 title('Velocity of Centroid P')

figure(6)
plot(time_dot, theta_dot, time_dot, phi_dot, time_dot, psi_dot)
xlabel('Time (seconds)')
ylabel('Angular Velocity (degree/sec)')
legend('Theta Dot (about X)', 'Phi Dot (about Y)', 'Psi Dot (about
Z)', 'Location', 'north')
title('Angular Velocity about Respective Axis')
end

7
111
8
112
9
113
Animation
if animation
figure(100)

alpha_base = 0:10:360; %creates angles for circular base plot

x_base = base_diameter*0.5*cosd(alpha_base); %x values for
circular base
y_base = base_diameter*0.5*sind(alpha_base); %y values for
circular base
z_base = zeros(length(x_base)); %z values for circular base

if configuration_6_3 == 0 && configuration_6_6 == 0

alpha = base_angle;
else
alpha = 0:60:360;
end

x = a*cosd(alpha); %x values for actuators base side

y = a*sind(alpha); %y values for actuators base side
z = zeros(length(x));
z1 = ones(length(x));

for j = 1:length(psi)
hold off

10
114
 plot3(x_base, y_base, z_base, 'b-', 'LineWidth', 3); %plot
circular base every time
hold on
for i = 1:6
% for each instance in time, plot the xyz values of the
end of the
% actuators based on the length
x_d(i) = x(i) + L_vector(1,i,j);
y_d(i) = y(i) + L_vector(2,i,j);
z_d(i) = z(i) + L_vector(3,i,j);
plot3([x(i) x_d(i)], [y(i) y_d(i)], [z(i)
z_d(i)], 'LineWidth', 3)
end

% periodic boundary condition to complete top plate drawing

x_d(i+1)=x_d(1);
y_d(i+1)=y_d(1);
z_d(i+1)=z_d(1);
plot3(x_d, y_d, z_d,'LineWidth', 3)

% plot settings
axis square;
axis([-(1.5*a) (1.5*a) -(1.5*a) 1.5*a 0 (max(Pz)+b/2)]);
grid on;
set(gcf,'WindowState','fullscreen')
pause(1/100)
if j < length(psi)
clf;
end
end
end

Published with MATLAB® R2018b

11
115
Published with MATLAB® R2018b

1 116
 Appendix J

ARDUINO SKETCH

117
118
119
120
121
122
 Appendix K

MATLAB GUI CODE

classdef StewartPlatformApp_exported < matlab.apps.AppBase

% Properties that correspond to app components

properties (Access = public)
%These properties (GUI object creation) were omitted for
brevity
end

properties (Access = private)

ser;
Actuator_Current_Pos = [0;0;0;0;0;0];
serial_state;
P1=zeros(1,6);
P2=zeros(1,6);
P3=zeros(1,6);
P4=zeros(1,6);
P5=zeros(1,6);
P6=zeros(1,6);
MoveComplete;

end

methods (Access = private)

function MoveToPoint(app,P)

speed = round(app.SetPlatformSpeedSlider.Value);

if app.PathPlanningCheckBox.Value == 1
j=1;
printOnce = true;
move_points = app.PointsEditField.Value;
pos_array = zeros(length(P), move_points);
for i = 1:length(P)
pos_array(i,:) =
round(linspace(app.Actuator_Current_Pos(i),P(i),move_points));
end

if app.serial_state == 1
while j <= move_points
if (printOnce == true)
fprintf(app.ser,
[strjoin({num2str(pos_array(1,j)), num2str(pos_array(2,j)),
num2str(pos_array(3,j)), num2str(pos_array(4,j)),
num2str(pos_array(5,j)), num2str(pos_array(6,j)),
num2str(speed)}), '\n']);
printOnce = false;
end
MoveCheck = fread(app.ser,1);

123
 if MoveCheck == 'D'
j = j +1;
printOnce = true; %intermediate move is done,
send the next position
end
end
app.ErrorMessage.Value = 'Move Complete';
end
end
if app.PathPlanningCheckBox.Value == 0 && app.serial_state ==
1
fprintf(app.ser, [strjoin({num2str(P(1)), num2str(P(2)),
num2str(P(3)), num2str(P(4)), num2str(P(5)), num2str(P(6)),
num2str(speed)}), '\n']);
MoveCheck = fread(app.ser,1);
if MoveCheck == 'D'
app.ErrorMessage.Value = 'Move Complete';
end
end

app.ErrorMessage.Visible = 'on';
for k = 1:length(P)
app.Actuator_Current_Pos(k) = P(1,k);
end
end
end

methods (Access = private)

% Code that executes after component creation

function startupFcn(app)
app.ErrorMessage.Visible = 'on';
app.SerialSwitch.Enable = 'off';
app.TopJointLocations63Panel.Visible = 'off';
app.TopJointLocations66Panel.Visible = 'on';
% app.t=timer();
% app.t.Period = 1;
% app.t.ExecutionMode = 'fixedRate';
% app.t.TimerFcn = @

end

% Selection changed function: StewartPlatformGUIButtonGroup

function StewartPlatformGUIButtonGroupSelectionChanged(app,
event)

if(app.GenericConfigurationButton.Value)
app.TopJointLocations63Panel.Visible = 'off';
app.TopJointLocations66Panel.Visible = 'on';
app.BaseJointLocationsPanel.Position =
[15,255,197,171];
end
if(app.ConfigurationButton_2.Value)

2
124
 app.TopJointLocations63Panel.Visible = 'on';
app.TopJointLocations66Panel.Visible = 'off';
app.BaseJointLocationsPanel.Position =
[15,324,197,171];
end

end

% Button pushed function: CalibrationButton

function CalibrationButtonPushed(app, event)

app.CalcJointButton.Enable = 'off';
app.ErrorMessage.Value = 'Calibrating...';

Serial configuration
ComPortNumber = num2str(app.COMPortEditField.Value);
PORT = strcat('COM', ComPortNumber);
BAUD_RATE = 115200;
DATA_BITS = 8;
PARITY = 'none';
STOP_BITS = 1;

% Establish a new serial connection (clear previous

connections)
delete(instrfindall);
app.ser = serial( PORT, ...
'BaudRate', BAUD_RATE, ...
'DataBits', DATA_BITS, ...
'Parity', PARITY, ...
'StopBits', STOP_BITS ...
);
fopen(app.ser);
app.SerialSwitch.Value = 'On';
app.serial_state = 1;
pause(14);
CalibrationState = fread(app.ser,1);
if CalibrationState == 1
app.CalibrationButton.Text = 'Calibrated';
app.CalibrationButton.Enable = 'off';
app.CalcJointButton.Enable = 'on';
app.ErrorMessage.Value = '';
app.SerialSwitch.Enable = 'off';
app.MovePlatformReset.Enable = 'on';
elseif CalibrationState == 0
app.ErrorMessage.Value = 'Calibration Failed';
else
app.ErrorMessage.Value = 'Serial Compare Failed';
end

end

% Button pushed function: CalcJointButton

3
125
 function CalcJointButtonPushed(app, event)
app.ErrorMessage.Value = '';
app.ErrorMessage.BackgroundColor = 'w';
% SET UP VARIABLES

%------------------------------------------------------------
a_length = 17/2; %Length [inches] of a vector on base
b_length = 13/2; %Length [inches] of a vector on top
ball_joint_height = 1.032; %Height from magnetic base to
sphere center [inches]
ball_diameter = 18/25.4; %[inches]
platform_thiccness = 0.25; %[inches]
joint_joint_offset = 15.61; %[inches]
%need to verify both dimensions
angleOffset_6_3 =3.5; %[degrees]
a_i = zeros(4,6);
b_i = zeros(4,6);
b_rot = zeros(4,6);
d_vector = zeros(4,6);
d_length = zeros(1,6);

%Take in

Px = app.PxInput.Value;
Py = app.PyInput.Value;
Pz = app.PzInput.Value;

%Remove offset from base top to ball center

Pz = Pz - ball_joint_height - ball_diameter/2 -
platform_thiccness;

P = [Px;Py;Pz;1];

% SET UP A and B Matrices

%------------------------------------------------------------

a_i(1,1) = a_length*cosd(app.Base_Act_1.Value); a_i(2,1) =

a_length*sind(app.Base_Act_1.Value);
a_i(1,2) = a_length*cosd(app.Base_Act_2.Value); a_i(2,2) =
a_length*sind(app.Base_Act_2.Value);
a_i(1,3) = a_length*cosd(app.Base_Act_3.Value); a_i(2,3) =
a_length*sind(app.Base_Act_3.Value);
a_i(1,4) = a_length*cosd(app.Base_Act_4.Value); a_i(2,4) =
a_length*sind(app.Base_Act_4.Value);
a_i(1,5) = a_length*cosd(app.Base_Act_5.Value); a_i(2,5) =
a_length*sind(app.Base_Act_5.Value);
a_i(1,6) = a_length*cosd(app.Base_Act_6.Value); a_i(2,6) =
a_length*sind(app.Base_Act_6.Value);
a_i(3,:) = 0;
a_i(4,:) = 1;

if(app.GenericConfigurationButton.Value)

4
126
 b_i(1,1) = b_length*cosd(app.Top_Act_1.Value);
b_i(2,1) = b_length*sind(app.Top_Act_1.Value);
b_i(1,2) = b_length*cosd(app.Top_Act_2.Value);
b_i(2,2) = b_length*sind(app.Top_Act_2.Value);
b_i(1,3) = b_length*cosd(app.Top_Act_3.Value);
b_i(2,3) = b_length*sind(app.Top_Act_3.Value);
b_i(1,4) = b_length*cosd(app.Top_Act_4.Value);
b_i(2,4) = b_length*sind(app.Top_Act_4.Value);
b_i(1,5) = b_length*cosd(app.Top_Act_5.Value);
b_i(2,5) = b_length*sind(app.Top_Act_5.Value);
b_i(1,6) = b_length*cosd(app.Top_Act_6.Value);
b_i(2,6) = b_length*sind(app.Top_Act_6.Value);

end

if(app.ConfigurationButton_2.Value)

b_i(1,1) = b_length*cosd(app.Top_Act_1_2.Value -
angleOffset_6_3); b_i(2,1) = b_length*sind(app.Top_Act_1_2.Value -
angleOffset_6_3);
b_i(1,2) = b_length*cosd(app.Top_Act_1_2.Value +
angleOffset_6_3); b_i(2,2) = b_length*sind(app.Top_Act_1_2.Value +
angleOffset_6_3);
b_i(1,3) = b_length*cosd(app.Top_Act_3_4.Value -
angleOffset_6_3); b_i(2,3) = b_length*sind(app.Top_Act_3_4.Value -
angleOffset_6_3);
b_i(1,4) = b_length*cosd(app.Top_Act_3_4.Value +
angleOffset_6_3); b_i(2,4) = b_length*sind(app.Top_Act_3_4.Value +
angleOffset_6_3);
b_i(1,5) = b_length*cosd(app.Top_Act_5_6.Value -
angleOffset_6_3); b_i(2,5) = b_length*sind(app.Top_Act_5_6.Value -
angleOffset_6_3);
b_i(1,6) = b_length*cosd(app.Top_Act_5_6.Value +
angleOffset_6_3); b_i(2,6) = b_length*sind(app.Top_Act_5_6.Value +
angleOffset_6_3);

end

b_i(3,:) = 0;
b_i(4,:) = 1;

% Rotate the B matrix and calculate the lengths of

actuators
for i = 1:6
b_rot(:,i)=stewartrot(-app.ThetaInput.Value, -
app.PhiInput.Value, -app.PsiInput.Value, b_i(:,i));
d_vector(:,i) = P(:,1) + b_rot(:,i) - a_i(:,i);
d_length(1,i) = sqrt(d_vector(1,i)^2 + d_vector(2,i)^2
+ d_vector(3,i)^2);
end

%Remove physical constants so length is on scale from 0

inches to 8 inches
d_length = d_length - joint_joint_offset;

5
127
 app.Actuator_1_Length_Set.Value = d_length(1,1);
app.Actuator_2_Length_Set.Value = d_length(1,2);
app.Actuator_3_Length_Set.Value = d_length(1,3);
app.Actuator_4_Length_Set.Value = d_length(1,4);
app.Actuator_5_Length_Set.Value = d_length(1,5);
app.Actuator_6_Length_Set.Value = d_length(1,6);

if max(d_length) > 8 || min(d_length) < 0

app.ErrorMessage.Visible = 'on';
app.MovePlatform.Enable = 'off';
app.ErrorMessage.Value = 'CALCULATED LENGTH(S) OUT OF
RANGE';
app.ErrorMessage.BackgroundColor = 'r';
app.ErrorMessage.FontWeight = 'bold';
else
app.MovePlatform.Enable = 'on';
app.SaveP1Button.Enable = 'on';
app.SaveP2Button.Enable = 'on';
app.SaveP3Button.Enable = 'on';
app.SaveP4Button.Enable = 'on';
app.SaveP5Button.Enable = 'on';
app.SaveP6Button.Enable = 'on';
% app.ErrorMessage.Visible = 'off';
end

output_length = round(interp1([-8 16],[-1024

2048],d_length));
app.Actuator_1_Pot_Set.Value = output_length(1,1);
app.Actuator_2_Pot_Set.Value = output_length(1,2);
app.Actuator_3_Pot_Set.Value = output_length(1,3);
app.Actuator_4_Pot_Set.Value = output_length(1,4);
app.Actuator_5_Pot_Set.Value = output_length(1,5);
app.Actuator_6_Pot_Set.Value = output_length(1,6);

end

% Button pushed function: MovePlatform

function MovePlatformButtonPushed(app, event)
output_length(1) = app.Actuator_1_Pot_Set.Value;
output_length(2) = app.Actuator_2_Pot_Set.Value;
output_length(3) = app.Actuator_3_Pot_Set.Value;
output_length(4) = app.Actuator_4_Pot_Set.Value;
output_length(5) = app.Actuator_5_Pot_Set.Value;
output_length(6) = app.Actuator_6_Pot_Set.Value;

MoveToPoint(app, output_length)
end

% Value changed function: SerialSwitch

function SerialSwitchValueChanged(app, event)
%value = app.SerialSwitch.Value;
switch app.SerialSwitch.Value

6
128
 case 'On'
fopen(app.ser);
app.serial_state = 1;
case 'Off'
app.serial_state = 0;
fclose(app.ser);
end

end

% Button pushed function: Logo_3

function Logo_3ButtonPushed(app, event)
web('https://www.calpoly.edu/');
end

% Button pushed function: Logo_2

function Logo_2ButtonPushed(app, event)
web('https://me.calpoly.edu/');
end

% Button pushed function: Logo_4

function Logo_4ButtonPushed(app, event)
web('http://heli-cal.com/');
end

% Button pushed function: Logo

function LogoButtonPushed(app, event)
web('https://www.progressiveautomations.com/');
end

% Button pushed function: SaveP1Button

function SaveP1ButtonPushed(app, event)
app.P1(1) = app.Actuator_1_Pot_Set.Value;
app.P1(2) = app.Actuator_2_Pot_Set.Value;
app.P1(3) = app.Actuator_3_Pot_Set.Value;
app.P1(4) = app.Actuator_4_Pot_Set.Value;
app.P1(5) = app.Actuator_5_Pot_Set.Value;
app.P1(6) = app.Actuator_6_Pot_Set.Value;

app.P1CheckBox.Value = 1;
app.MovePlatformPtoP.Enable = 'on';
app.ClearP1Button.Enable = 'on';
app.MovetoP1Button.Enable = 'on';
end

% Button pushed function: SaveP2Button

function SaveP2ButtonPushed(app, event)
app.P2(1) = app.Actuator_1_Pot_Set.Value;
app.P2(2) = app.Actuator_2_Pot_Set.Value;
app.P2(3) = app.Actuator_3_Pot_Set.Value;
app.P2(4) = app.Actuator_4_Pot_Set.Value;
app.P2(5) = app.Actuator_5_Pot_Set.Value;
app.P2(6) = app.Actuator_6_Pot_Set.Value;

7
129
 app.P2CheckBox.Value = 1;
app.MovePlatformPtoP.Enable = 'on';
app.ClearP2Button.Enable = 'on';
app.MovetoP2Button.Enable = 'on';
end

% Button pushed function: SaveP3Button

function SaveP3ButtonPushed(app, event)
app.P3(1) = app.Actuator_1_Pot_Set.Value;
app.P3(2) = app.Actuator_2_Pot_Set.Value;
app.P3(3) = app.Actuator_3_Pot_Set.Value;
app.P3(4) = app.Actuator_4_Pot_Set.Value;
app.P3(5) = app.Actuator_5_Pot_Set.Value;
app.P3(6) = app.Actuator_6_Pot_Set.Value;

app.P3CheckBox.Value = 1;
app.MovePlatformPtoP.Enable = 'on';
app.ClearP3Button.Enable = 'on';
app.MovetoP3Button.Enable = 'on';
end

% Button pushed function: SaveP4Button

function SaveP4ButtonPushed(app, event)
app.P4(1) = app.Actuator_1_Pot_Set.Value;
app.P4(2) = app.Actuator_2_Pot_Set.Value;
app.P4(3) = app.Actuator_3_Pot_Set.Value;
app.P4(4) = app.Actuator_4_Pot_Set.Value;
app.P4(5) = app.Actuator_5_Pot_Set.Value;
app.P4(6) = app.Actuator_6_Pot_Set.Value;

app.P4CheckBox.Value = 1;
app.MovePlatformPtoP.Enable = 'on';
app.ClearP4Button.Enable = 'on';
app.MovetoP4Button.Enable = 'on';
end

% Button pushed function: SaveP5Button

function SaveP5ButtonPushed(app, event)
app.P5(1) = app.Actuator_1_Pot_Set.Value;
app.P5(2) = app.Actuator_2_Pot_Set.Value;
app.P5(3) = app.Actuator_3_Pot_Set.Value;
app.P5(4) = app.Actuator_4_Pot_Set.Value;
app.P5(5) = app.Actuator_5_Pot_Set.Value;
app.P5(6) = app.Actuator_6_Pot_Set.Value;

app.P5CheckBox.Value = 1;
app.MovePlatformPtoP.Enable = 'on';
app.ClearP5Button.Enable = 'on';
app.MovetoP5Button.Enable = 'on';
end

% Button pushed function: SaveP6Button

function SaveP6ButtonPushed(app, event)

8
130
 app.P6(1) = app.Actuator_1_Pot_Set.Value;
app.P6(2) = app.Actuator_2_Pot_Set.Value;
app.P6(3) = app.Actuator_3_Pot_Set.Value;
app.P6(4) = app.Actuator_4_Pot_Set.Value;
app.P6(5) = app.Actuator_5_Pot_Set.Value;
app.P6(6) = app.Actuator_6_Pot_Set.Value;

app.P6CheckBox.Value = 1;
app.MovePlatformPtoP.Enable = 'on';
app.ClearP6Button.Enable = 'on';
app.MovetoP6Button.Enable = 'on';
end

% Button pushed function: ClearP1Button

function ClearP1ButtonPushed(app, event)
app.P1 = zeros(1,6);
app.P1CheckBox.Value = 0;
app.ClearP1Button.Enable = 'off';
app.MovetoP1Button.Enable = 'off';
end

% Button pushed function: ClearP2Button

function ClearP2ButtonPushed(app, event)
app.P2 = zeros(1,6);
app.P2CheckBox.Value = 0;
app.ClearP2Button.Enable = 'off';
app.MovetoP2Button.Enable = 'off';
end

% Button pushed function: ClearP3Button

function ClearP3ButtonPushed(app, event)
app.P3 = zeros(1,6);
app.P3CheckBox.Value = 0;
app.ClearP3Button.Enable = 'off';
app.MovetoP3Button.Enable = 'off';
end

% Button pushed function: ClearP4Button

function ClearP4ButtonPushed(app, event)
app.P4 = zeros(1,6);
app.P4CheckBox.Value = 0;
app.ClearP4Button.Enable = 'off';
app.MovetoP4Button.Enable = 'off';
end

% Button pushed function: ClearP5Button

function ClearP5ButtonPushed(app, event)
app.P5 = zeros(1,6);
app.P5CheckBox.Value = 0;
app.ClearP5Button.Enable = 'off';
app.MovetoP5Button.Enable = 'off';
end

% Button pushed function: ClearP6Button

9
131
 function ClearP6ButtonPushed(app, event)
app.P6 = zeros(1,6);
app.P6CheckBox.Value = 0;
app.ClearP6Button.Enable = 'off';
app.MovetoP6Button.Enable = 'off';
end

% Button pushed function: MovePlatformPtoP

function MovePlatformPtoPButtonPushed(app, event)
output_length(1,:) = app.P1;
output_length(2,:) = app.P2;
output_length(3,:) = app.P3;
output_length(4,:) = app.P4;
output_length(5,:) = app.P5;
output_length(6,:) = app.P6;

% This checks for any rows of 0 and removes, leaving an

array of non-zero positions
% Limitation of inability to go to reset position as part
of
% the sequence
output_length=output_length(any(output_length,2),:);

for w = 1:size(output_length,1)
MoveToPoint(app, output_length(w,:))
end
end

% Button pushed function: MovePlatformReset

function MoveReset(app, event)
Home_Pos = [10,10,10,10,10,10];
if app.serial_state == 1
fprintf(app.ser, [strjoin({num2str(Home_Pos(1)),
num2str(Home_Pos(2)), num2str(Home_Pos(3)), num2str(Home_Pos(4)),
num2str(Home_Pos(5)), num2str(Home_Pos(6)), num2str(60)}), '\n']);
end

app.ErrorMessage.Visible = 'on';
app.ErrorMessage.Value = [strjoin({num2str(Home_Pos(1)),
num2str(Home_Pos(2)), num2str(Home_Pos(3)), num2str(Home_Pos(4)),
num2str(Home_Pos(5)), num2str(Home_Pos(6)), num2str(60)}), '\n'];

for k = 1:length(Home_Pos)
app.Actuator_Current_Pos(k) = Home_Pos(k);
end
end

% Button pushed function: MovetoP1Button

function MovetoP1ButtonPushed(app, event)
MoveToPoint(app, app.P1)
end

% Button pushed function: MovetoP2Button

function MovetoP2ButtonPushed(app, event)

10
132
 MoveToPoint(app, app.P2)
end

% Button pushed function: MovetoP3Button

function MovetoP3ButtonPushed(app, event)
MoveToPoint(app, app.P3)
end

% Button pushed function: MovetoP4Button

function MovetoP4ButtonPushed(app, event)
MoveToPoint(app, app.P4)
end

% Button pushed function: MovetoP5Button

function MovetoP5ButtonPushed(app, event)
MoveToPoint(app, app.P5)
end

% Button pushed function: MovetoP6Button

function MovetoP6ButtonPushed(app, event)
MoveToPoint(app, app.P6)
end

% Callback function
function CalibrationButton_2Pushed(app, event)
MoveDone = fread(app.ser,1);
if MoveDone == 'D'
app.ErrorMessage.BackgroundColor = 'r';
end
end
end

% App initialization and construction

methods (Access = private)

%These properties (GUI object creation) were omitted

for brevity

end

methods (Access = public)

% Construct app
function app = StewartPlatformApp_exported

% Create and configure components

createComponents(app)

% Register the app with App Designer

registerApp(app, app.UIFigure)

% Execute the startup function

runStartupFcn(app, @startupFcn)

11
133
 Appendix L

STEWART PLATFORM LAB MANUAL

EXPERIMENT ONE

STEWART-GOUGH PLATFORM

INTRODUCTION:
In this experiment, you will be working with a type of parallel robot called a Stewart Platform.
The platform was initially developed from the late 1940s to mid-1960s. Eric Gough established
the principles of a closed-loop kinematic mechanism in 1947 and built a prototype for tire testing
in 1955. In 1965 D. Stewart published “A Platform with Six Degrees of Freedom" as flight
simulators rose in popularity. Despite significant contributions by both men the platform has
eventually become known as the Stewart Platform. Figure 1 shows the Stewart Platform that
will be used in this experiment.

Figure 1: Stewart Platform

The following sections introduce the primary components used in the 6 DOF SPS Stewart
Platform.

134
Progressive Automation PA-14P Linear Actuators

This Stewart Platform is comprised of 6 linear actuators made by Progressive Automations.

These actuators have a stroke of 8 inches and are powered by 12VDC motors, enabling extension
and retraction at up to 2 in/second. They feature internal potentiometers for position feedback
and limit switches that prevent extending or retracting beyond the actuator design. These
comprise the prismatic portion of the Stewart Platform’s kinematic chains.

KD418 Magnetic Ball Joint

Magnetic ball joints are used at either end of the actuators to fulfill the S portion of the SPS
mechanism. The joints are comprised of two pieces: a steel sphere with a threaded rod, and a
cylindrical magnet that interfaces with the sphere. Magnetic ball joints are available
commercially at various sizes and strengths. The KD418 has a holding force of 5kgs (11 lbf) and
will separate under greater loading. This holding force allows the joint to operate as intended
under normal conditions, and separate if a crash or binding occurs. They are designed to be the
failure point, preventing excess loading on the actuators and acrylic plates.

Figure 2: KD418 Separation Technique

There will be a portion of the experiment that requires separating the magnets and spheres. It is
recommended to follow Figure 2 for easiest removal. Do not separate along the cylindrical axis;
hold the cylinder with a finger or two and use the base of your thumb to push the sphere laterally.
The magnets were tested for a crushing or pinching risk by placing a finger between the sphere
and magnet, and no magnetic clamping force was experienced.

135
Arduino Due

The controller behind the Stewart Platform is an Arduino Due, powered by an AT91SAM3X8E
microcontroller that operates on 3.3VDC at a clock speed of 84 Mhz. The Due has 54 digital I/O
pins, of which 12 can be configured to provide PWM output. It also has 12 analog input pins.

The Stewart Platform uses 13 digital output pins: 6 direction pins to specify extension or
retraction, 6 PWM outputs, and an enable pin. 6 analog input pins will read the potentiometer’s
signal voltage and correlate it to position in software. The input pins are set up for 10-bit analog-
to-digital (ADC) conversion meaning the 8” stroke will have a range of 1024 position values,
equal to 0.0078”/count. The Due has a control loop programmed that will act on inputs from the
host PC.

HexaMoto Shield

The Arduino Due is unable to output any significant amount of current to power the linear
actuators, so an external power source and motor drivers are needed. The HexaMoto Shield is a
custom PCB that attaches to the top of the Due and has 6 motor drivers. They are connected to a
12VDC power supply and use the Due’s PWM and direction signals to power the actuators in a
controlled manner. The HexaMoto also features screw terminals to easily connect the
potentiometers to the Due’s analog inputs. These main components are labeled in Figure 3.

Figure 3: HexaMoto Components

136
Inverse Kinematics:
The Stewart Platform has 6 degrees of freedom which allows control over both position and
orientation of the moving platform. Forward kinematics for the Stewart Platform involves
determining position and orientation of the platform given actuator link lengths which requires
advanced numerical analysis methods beyond the scope of this course.

The inverse kinematics return the link lengths given position and orientation and are much more
straightforward for the Stewart Platform. Figure 4 sets up the derivation of the inverse kinematic
relations.

Figure 4: Vector Definition for Inverse Kinematic Derivation

The desired vector for each actuator i is

(Equation 1)

where

(Equation 2)

and

137
 (Equation 3)

However, 𝑩 ̅ changes as the top platform rotates about any of the three axis and must be
accounted for with rotation matrixes. Note that theta, phi, and psi correspond to the x-, y-, and z-
axis respectively.

(Equation 4)

̅ matrix into the vector equation yields

Substituting the rotated 𝑩

(Equation 5)

138
Simulation and Verification of Equations - Foreword

The first part of the experiment, you will use Matlab to first run a simulation of the Stewart
Platform. A majority of the simulation will be provided, with some intentional omissions. To get
the simulation working, you must:

1. Write a function called stewartrot that uses the rotation matrices in Equation 4 to
̅ 𝒉𝒐𝒎𝒆 to 𝑩
transform 𝑩 ̅ 𝒓𝒐𝒕 . The inputs to stewartrot must be theta, phi, psi, and b_home.
The function will apply the rotation matrixes to b_home and output a 4x1 matrix b_rot.

It is strongly advised to create and test this function before the lab. Save it to a flash
drive to load it into the Matlab folder. This will help complete the simulation
portion quickly.

2. In lab you will complete the inverse kinematics for the generic case. The inverse
kinematics for the 6-6 and 6-3 Stewart Platform configurations are already included in
the simulation. You may refer to the setup and usage of these two versions of the inverse
kinematics to create the general case kinematics. Further detail will be given in the
experiment.

EXPERIMENT
Starting the Stewart Platform

• Ensure the power cable is plugged into an outlet and to the electronics enclosure. Connect
the USB cable between the Stewart Platform and the host PC. Verify the Arduino Due is
connected by observing a green LED on the board. Turn on the power switch and confirm
the power supply is on by observing a green LED on the unit.

• Open the StewartPlatformStarterApp shortcut on the desktop. The following GUI should
appear:

139
• In order to communicate with the Arduino Due, a serial connection must be established.
From the Start Menu in Windows, open Device Manager and locate the COM Port used
by the Arduino Due. Insert this value into the GUI. In the provided example, the value to
enter would be ‘3’. Exit this window once the value is obtained.

• Press the Calibrate button. This will open the serial connection and the robot will extend
the actuators to their maximum length, pause, and retract to their minimum length. The
Arduino is reading the potentiometer readings and verifying proper functionality. A
calibration status message will be returned. The Stewart Platform is now ready to move.

• Input the following values into the GUI

Move 1 Move 2 Move 3 Move 4 Move 5

Actuator 1 2.0 2.0 2.0 2.0 6.8

Actuator 2 2.5 3.0 3.0 3.0 7.9

Actuator 3 2.0 2.3 2.3 2.3 6.8

Actuator 4 2.5 5.2 5.2 5.2 7.9

Actuator 5 2.0 2.8 3.8 1.8 6.8

Actuator 6 2.5 2.8 2.8 2.8 7.9

• Observe the resulting position and motion of the Stewart Platform. Note that you are
commanding actuator lengths, and it would take complicated forward kinematics to
determine the platform position and orientation from these lengths. This lab will focus on
the use of inverse kinematics to calculate the link lengths from a specified the platform’s
position and orientation.

• Note that only actuator 5 changes in Move 3 and Move 4. Note the effect that one
actuator’s movement has on the platform position and orientation.

• Before closing the GUI, move the platform to reset position.

140
Simulation and Verification of Equations

• Open Matlab and open Stewart_Platform.m. Transfer your written and tested function
stewartrot.m to the same folder as Stewart_Platform.m.

• In the %%Set Constants section of the simulation, base_diameter and top_diameter are
missing. Using a tape measure, measure the bolt circle diameters of the base and moving
platform, rounded to the nearest tenth of an inch. Update these variable values.

• Under %Set base and top angles create base_angle and top_angle which are 1x6
matrices containing the angular positions of the spherical joints, in ascending order.
o Use base angles of 0°, 80°, 120°, 200°, 240°, and 300°.
o Use top angles of 0°, 60°, 120°, 180°, 240°, and 300°.

• In %% Calculate EOMs locate the inverse kinematics for the 6-6 and 6-3
configurations and observe their usage. Using these as a reference:
o Create a_i and b_i in a similar manner, using base_angle and top_angle as the
inputs to sine and cosine.
o Using your tested function stewartrot.m, rotate b_i. Store the output of this
function in b_rot.
o Using Equation 5, calculate the array L_vector containing the x, y, and z lengths
of the actuators using P, a_i, and b_rot.
o Calculate the magnitude of L_vector and store it into L_length and end both
nested for loops and the outer if statement.

• Run the simulation with animation true and plots false. The platform should increase in Z
and in theta.

• Set configuration_6_3 to true. The geometry of the 6-3 configuration is already set.
Verify the configuration change but the motion remains the same.

• Under %% Motion Profile Creation, there are several motion profiles that specify the
values of each degree of freedom throughout the simulation. Select a motion profile to
simulate. An example of a motion profile is shown below.

141
 • With animation true and plots false, run the script. Observe the animation and verify the
motion matches the motion profile.

• With animation false and plots true, run the simulation again to generate the plots. Save
them for inclusion in the report. In the report, identify if this motion profile would be
possible on the actual Stewart Platform and justify your determination.

• Using a configuration of your choice, make a final motion profile that includes motion in
all degrees of freedom. Ensure that this motion profile stays within the Stewart Platform
link length limits and actuator speed limits. You may use linear motion, as seen in the
theta and psi equations, but for more interesting plots and animations, attempt a
complicated motion profile that uses sin, cos, or other continuous functions. When
complete, show the instructor your animation, and save the plots for inclusion in your
report.

Instability Investigation

• Configure the Stewart Platform to be a true 6-6 configuration, where all joints are evenly
spaced at 60° apart. With at least two people, disassemble the Stewart Platform:
o First, turn off power and unplug all connections.
o Remove one linear actuator by separating the magnet at the end of the actuator
from the steel sphere on the moving platform. The top plate is now unconstrained
and must be held. Finish removing the actuator by separating the sphere at the
base of the actuator and the magnet attached to the base platform and set it on the
desk.
o Finish removing remaining 5 actuators and set the moving platform down.

• Move the magnets on the base platform to 0°, 60°, 120°, 180°, 240°, and 300°. Move the
steel spheres on the moving platform to the same angles.

• Reconnect the linear actuators to both platforms. Connect actuator 1 to the joint at the
smallest angle on both the moving and base platforms. In this case, actuator 1 would
connect to both joints at 0°, and actuator 6 would connect to both joints at 300°. Follow
this pattern at all times when reconfiguring the robot. Keep hold of the top platform
even when this process is complete.

142
 • Is the robot stable? Try to position the robot into a stable position using your hands. You
can let go slightly to check stability, but do not let the platform fall. If you find a stable
position, what happens when it is twisted slightly? Observe carefully and in your report,
provide an explanation regarding the stability of the Stewart Platform in a true 6-6
configuration.

Range of Motion: 6-6 Configuration

Reconfigure

• Change the Stewart Platform configuration to a modified 6-6 configuration. Disassemble

the Stewart Platform as before and move the magnets on the base platform to 0°, 80°,
120°, 200°, 240°, and 320°. Move the steel sphere joints on the moving platform to 0°,
40°, 120°, 160°, 240°, and 280°.

Startup

• Ensure the power cable is plugged into an outlet and to the electronics enclosure. Connect
the USB cable between the Stewart Platform and the host PC. Verify the Arduino Due is
connected by observing a green LED on the board. Turn on the power switch and confirm
the power supply is on by observing a green LED on the unit.

• Open StewartPlatformApp.mlapp from the desktop shortcut. The following GUI should
appear.

Figure 5: Stewart Platform GUI

143
• Set the configuration to Generic Configuration. This is a modified 6-6 that allows all 12
joint locations to be input. Input the values in ascending order according to the Stewart
Platform setup.

• Reconnect all cables. Enter the COM port of the Arduino Due and calibrate the robot
using the same procedure with the basic GUI.

Testing

• Verify that the top and base joint locations are correct. Find the maximum value the
Stewart Platform can achieve about each axis, one at a time. Achieve this by using the
“Position and Orientation” panel and “Motion” panel. First, input desired position and
orientation values and select “Calculate Actuator Lengths”. The output of the calculation
appears on “Calculated Actuator Lengths” for your review.

• Under the “Motion” panel, use “Move to Current Setpoint” to move the Stewart Platform
to the position currently calculated. Throughout the experiment, use the ‘Path Planning’
and ‘Speed’ features and observe the motion differences. If you desire to move the
Stewart Platform back to its reset position, use the “Move to Reset Position” button at
any time.

• You may alter other axis values to maximize the desired axis value. For example, Figure
6 shows that actuators 2, 4, and 6 have effectively reached their 8-in, 1024 potentiometer
count limit with all other “Position and Orientation” inputs equal to zero. However,
changing the value of Ψ from 0° to -20° allows Pz to increase by 0.4 inches.

Figure 6: Tip for Maximizing Range of Motion

• It’s recommended that you find the maximum z-coordinate, theta rotation, and phi
rotation first. These can be found without the robot actuating into an unstable position.

144
 • Finding the limits of Px, Py, and Ψ are more difficult. You cannot rely on the actuator
lengths calculation to determine the maximum value, as doing so may cause the robot to
collapse. Find the value by incrementing it, moving the robot and checking stability, and
incrementing again. You may want to have someone keep their hands close to the moving
plate in case the robot has been extended too far. If it does overextend, a magnetic joint
or more will separate, and the platform will fall. Simply reattach the joints and move the
platform location to a more centered value. Hold the moving platform initially to support
it as it moves to a more stable position.
• In this configuration the Stewart Platform can achieve a much larger Ψ value in the
negative direction than in the positive direction. Verify this is the case and report the
largest magnitude of Ψ.

Value to Maximize Px (in) Py (in) Pz (in) Θ (°) Φ (°) Ψ (°)

Px
Py
Pz
Θ (about n-axis)
Φ (about o-axis)
Ψ (about a-axis)

Range of Motion: 6-3 Configuration

Reconfigure

• Change the configuration of the robot to a 6-3 configuration. Disassemble the Stewart
Platform as before and move the magnets on the base platform to 0°, 60°, 120°, 180°,
240°, and 300°.
• On the moving platform attach two spherical joints at 0°, 120°, and 240° using the pair of
threaded holes adjacent to these angles. Connect actuators 1 and 2 to the 0° joints,
actuators 3 and 4 to the 120° joints, and actuators 5 and 6 to the 240° joints. This is also
specified in the GUI.

Note that if two magnetic cylinders were to be attached to the same spherical joint, they
cylinders would be greater than 90° apart, as shown in Figure 7. This cannot be achieved
with the given Stewart Platform geometry, so pairs of spherical joints are used to best
represent a true 6-3 configuration. The resulting angular offset from the nominal 0°, 120°,
and 240° should be neglected in the inverse kinematic equations. In Stewart_Platform.m, the
6-3 inverse kinematics utilize joint_angle to account for the offset. This joint angle is utilized
by the GUI as well.

145
 Figure 7: Interference Resulting from True 6-3 Configuration

• Perform the same range of motion tests for the six degrees of freedom and record them.

Value to Maximize Px (in) Py (in) Pz (in) Θ (°) Φ (°) Ψ (°)

Px
Py
Pz
Θ (about x-axis)
Φ (about y-axis)
Ψ (about z-axis)

• In the report, compare the range-of-motion differences between the two configurations
and give a brief summary describing how the Stewart Platform geometry affects its range
of motion.

146
Flight Simulator

This portion of the experiment investigates one of the original uses of a Stewart Platform: a flight
simulator. This portion will build upon the experiment thus far by incorporating motion
sequencing.

After a position and orientation has been input and calculated, it can be saved using the buttons
found in the “Save and Sequence Setpoints” panel. If a saved setpoint is incorrect or needs to be
changed, it can be saved again and overwritten. If no longer desired, the saved setpoint can be
cleared.

When the “Move Through Sequence” button is selected, the platform will move sequentially
through all the saved setpoints. It will ignore any unsaved or cleared setpoints.

Using the axis specified in Figure 8, simulate:

Figure 8: Axis Specification for Aircraft Flight Simulator

147
 • A takeoff, in which the plane’s pitch and acceleration causes the passengers to feel a
combined total of 0.5 Gs of acceleration in the -Y direction. Calculate the Stewart
Platform pitch necessary to simulate this acceleration by using gravity and save your
result in P1.
• While still accelerating and rising, a turn to the left, in which the aircraft yaws 10°
and rolls -8°. Save this calculated setpoint to P2
• The aircraft leveling out, and yawing -45°. Save this to P3
• The aircraft banking to the right with a roll of 15° (maintaining yaw). Save this to P4
• A landing approach, in which the aircraft has a negative pitch and no roll or yaw.
The passengers experience 0.75Gs in the -Z direction (which includes acceleration
due to gravity). Determine the pitch necessary to simulate this and save it to P5.
• A landed aircraft on a level tarmac. Save this to P6.

While only changing pitch, roll, and yaw is necessary to complete this simulation, you may wish
to add translation along the X, Y and Z axes to mimic the aircraft’s relative position. For
example, a takeoff could include an increase along Z by a couple inches, or a banking/turning
left would result in a decrease along X. You have all 6 degrees of freedom at your disposal, be
creative!

• Run the sequence of setpoints to run the flight simulator, and have your instructor verify
your motion.

Cleanup Procedure

• When completed with the experiment, move the Stewart Platform to the reset position for
the next lab group.
• Turn off the switch on the electronics enclosure to power off the motors and unplug the
USB cable from the computer to disconnect serial and power off the Arduino.
• Close the GUI and delete all generated files from this lab session.

148