INDIAN INSTITUTE OF TECHNOLOGY ROORKEE

MIN-502: Robotics & Control

Prof. S. P. Harsha
[email protected]
9917489849 (M)
2
Modeling of Robots: Coordinates Frames
• The manipulator consists of several links and joints
– Coordinated motion with constraints

• “The modeling of robot comprises of establishing a special

relationship between the manipulator and the manipulated
object”
– Vector and matrix algebra used in this regards

• The motion of each link can be described with respect to a

reference coordinate frame and hence it is convenient to have
a coordinate frame attached to the body of each link

3
Modeling of Robots: Coordinates Frames
• Description of object in space
• Position
• Orientation

4
Modeling of Robots: Coordinates Frames

5
Positions, orientations and frames
• The position of a point p relative to a coordinate system A
(Ap):

 px  ZA
A
p
 
A
p =  py 
p  YA
 z XA

6
Positions, orientations and frames
• The orientation of a body is described by a coordinate
system B attached to the body, relative to A (a known
coordinate system).

ZA ZB

XB YA

XA YB

7
Positions, orientations and frames
• A frame is a set of 4 vectors giving the position and
orientation.
• Example: frame B  
B = A
R ,A
PB Borg 
ZBZB

AAP
ZA P Borg
Borg

XX
BB
YBYB
YA

XA orientation
position
8
Coordinates Frames: Conventional 3D coordinate frame:

Space frame notation:

1 represents frame-1

In vector-matrix notation:

Direction cosines:
9
 Mapping

• “Mappings refer to changing the description of a point in

space from one frame to another frame”

• Three possibilities:
1. Second frame is rotated w. r. t. the first
2. Second frame is translated from the first
3. Second frame is rotated as well as translated w. r. t. the
first (combined)

10
 1. Mapping Between Rotated Frames
• Consider the two frames: Frame -1 (x, y, z) and
Frame – 2 (u, v, w)
 1. Mapping Between Rotated Frames
• The problem may be posed as: “The description of point
P in frame-2 is known and its description in frame-1 is to
be found”

• We need to find the projections of 2P in frame-1

 1. Mapping Between Rotated Frames
• The above relationship can be expressed in matrix form
as:
 1. Mapping Between Rotated Frames
• R is known as the rotation matrix or rotational
transformation matrix

• We can extend the discussion to:

 1. Mapping Between Rotated Frames
• Relationships:
 2. Mapping Between Translated Frames
• Consider the two coordinate frames:
 2. Mapping Between Translated Frames
• We can have:

• Also,

• Thus,
2. Mapping Between Translated Frames

• Rotation and Translation are different mathematically

as well as qualitatively
• Rotation matrix is a 3 x 3, whereas translation is
represented by 3 x 1 vector

• “A powerful representation of translation is in a 4D

space of homogeneous coordinates”
2. Mapping Between Translated Frames

• In homogeneous coordinates, a point P is space w.r.t.

frame-1 is denoted as:

• σ is a non-zero positive scale factor

 2. Mapping Between Translated Frames

• In homogeneous coordinates, vector-matrix notation

for translation can be expressed as:
3. Mapping Between Rotated and Translated Frames

• Consider the following two coordinate frames:

Intermediate reference frame 1’

3. Mapping Between Rotated and Translated Frames

• Now:

• Thus:

• Finally:
 3. Mapping Between Rotated and Translated Frames
• In homogeneous coordinates:
 Fundamental Rotation Matrices
Principal Axes Rotation:
• “Successive rotations are performed about
the principal axes of the fixed frame”
 Fundamental Rotation Matrices
Principal Axes Rotation:
• “Successive rotations are performed about
the principal axes of the fixed frame”
 Fundamental Rotation Matrices
• Principal Axes Rotation: Consider the following
rotation:

Rotation matrix 1R2 is

denoted by Rz(θ) or
R(z, θ) or Rz,θ
 Fundamental Rotation Matrices
• Principal Axes Rotation:
• In simplified form:

• Similarly:

• In homogeneous transforms:
 Fundamental Rotation Matrices
• Principal Axes Rotation:
• Rotation matrices can be multiplied together to
get sequence of finite rotations:
The order is left to right

The order is apparently important

 Fundamental Rotation Matrices
• Fixed Angle Representation:
• Fixed frame is {1} or {x y z}
• Moving frame is {2} or {u v w}
• Consider the following rotations of
moving frame about the three axes of fixed frame:
 Fundamental Rotation Matrices
• Fixed Angle Representation:
• Such a convention is known as fixed angle representation.
• The described rotations are described as XYZ-fixed angle
rotations. The overall rotation matrix:
 Fundamental Rotation Matrices
• Fixed Angle Representation:
• The three rotations about the three fixed principal axes in
fixed angle rotation produce the motions which are also
known as roll, pitch and yaw motions

• The XYZ-fixed angle transformation is equivalent to roll-pitch-

yaw (RPY) transformation
 Fundamental Rotation Matrices
• Euler Angle Representation:
• The moving frame, instead of rotating about the principal axes
of the fixed frame, can rotate about its own principal axes.
• Consider the following rotations:

• This convention for specifying the orientation is called WVU-

Euler angle representation
 Fundamental Rotation Matrices
• Euler Angle Representation:
• Equivalent rotation matrix is computed by post multiplication
of matrices of elementary rotations as follows:
 Fundamental Rotation Matrices
• Euler Angle Representation:
• Equivalent rotation matrix is computed by post multiplication
of matrices of elementary rotations as follows:
 Fundamental Rotation Matrices
• Euler Angle Representation:
• Results are same as that obtained for fixed angle
representation, but the rotations were performed in opposite
order.
• Thus,

• The above relationship is ambiguous. We modify it to the

following:
 Fundamental Rotation Matrices
• Problem:
 Fundamental Rotation Matrices
• Problem:
• In the previous problem 1T2 was obtained, which
describes the position and orientation of frame
{2} relative to frame {1}. Using this matrix,
determine the description of frame {1} relative to
frame {2}
 Fundamental Rotation Matrices
• Solution:
 Fundamental Rotation Matrices
• Problem:
 Fundamental Rotation Matrices
• Solution:
 Fundamental Rotation Matrices
Problem:
1. A vector v=3i+2j+7k is rotated by 60 about the z-axes
of the reference frame. It is then rotated by 30 about
the x-axes of the reference frame. Find the rotation
transformation.

2. A vector v=3i+2j+7k is rotated by 60 about the z-axes

of the reference frame. It is then rotated by 30 about
the x-axes of the rotated frame. Find the rotation
transformation.
3. A vector P=3i-2j+5k is first rotated by 90° about the x-
axis, then by 90° about z-axis. Finally it is translated by
-3i+2j-5k. Determine the new position of vector P.
Problem:
4. Find the new location of point G, initially at G=[ 3 0 -1
1]T, if
– It is rotated by 180° about z-axis and then translated by 3 units
along the y-axis
– It is first translated by 3 units along the y-axis and then rotated
by 180° about z-axis
• Are the two locations same? Why?

5. A moving frame is rotated about x-axis of the fixed

coordinate frame by 30°. The coordinates of a point Q in
fixed coordinate frame are given by Q=[ 2 0 3]T
• What will be the coordinates of a point Q with respect
to the moving frame?
 Kinematics in Robotic Manipulator
• To do what a manipulator does, requires control of position of
each link and joint of the manipulator to control both the
position and orientation of the tool.
• To program the tool motion and joint-link motions, a
mathematical model of the manipulator is required to refer to all
geometrical and/or time-based properties of the motion.
• Kinematic model describes the spatial position of joints and links,
and position and orientation of the end-effector
• The derivatives of kinematics deal with the mechanics of motion
without considering the forces that cause it. (Differential
Kinematics)
• The relationships between the motions and the forces and/or
torques that cause them is the dynamics problem
 Kinematics in Robotic Manipulator
• Completely familiar with the Robot Anatomy
– Links
– Joints (1-DOF)
• Joints with more than 1-DOF are less common.
– A Joint m–DOF can be modeled as m joints with 1-DOF
each connected with (m-1) links of zero length.
• Most manipulators have open serial kinematic chain
• All joints are motorized
 Mechanical Structure and Notations

• Joints may be Revolute (R) or Prismatic (P)

• Each joint has a joint axis

• By convention, z-axis of the co-ordinate frame is aligned with

the joint axis
 Mechanical Structure and Notations

• Consider the following convention for numbering links and

joints:

• Thus, an n-DOF manipulator arm consists of (n+1) links

connected by n joints
 Rotation Matrices in 3D

cosθ − sinθ 0
Rotation around the Z-Axis
R z =  sinθ cosθ 0
 0 0 1

 cosθ 0 sinθ 
R y =  0 1 0  Rotation around the Y-Axis

− sinθ 0 cosθ 

𝟏 𝟎 𝟎
𝐑𝑋 = 𝟎 𝐜𝐨𝐬𝛉 −𝐬𝐢𝐧𝛉 Rotation around the X-Axis
𝟎 𝐬𝐢𝐧𝛉 𝐜𝐨𝐬𝛉

47
 Homogeneous Matrices (H) in 3D
H is a 4x4 matrix that can describe a translation, rotation, or both in one matrix
O
1 0 0 Px 
Y
0 1 0 Py 
N H=
P
0 0 1 Pz 
X A  
Translation without rotation 0 0 0 1
Z

Y
n x ox ax 0
n oy ay 0
H= y
O
N
n z oz az 0
 
X 0 0 0 1
Rotation part:
Z Rotation without translation Could be rotation around z-axis,
A x-axis, y-axis or a combination of
the three.
48
 The Homogeneous Matrix is a concatenation of numerous
translations and rotations

W N 
Y  O
W 
W A 
 
T
P
X

Z
One more variation on finding H:

H = (Rotate so that the X-axis is aligned with T)

* ( Translate along the new t-axis by || T || (magnitude of T))
* ( Rotate so that the t-axis is aligned with P)
* ( Translate along the p-axis by || P || )
* ( Rotate so that the p-axis is aligned with the O-axis)

49
 We are interested in two kinematics topics
Forward Kinematics (angles to position)
What you are given: The length of each link
The angle of each joint

What you can find: The position of any point

(i.e. it’s (x, y, z) coordinates

Inverse Kinematics (position to angles)

What you are given: The length of each link
The position of some point on the robot

What you can find: The angles of each joint needed to

obtain that position

50
 Forward v/s Inverse

•Forward Kinematics
– Compute configuration (pose) given individual DOF
values
• Good for simulation

•Inverse Kinematics
– Compute individual DOF values that result in
specified end effector position
• Good for control

51
 What is Inverse Kinematics?
• Forward Kinematics

End Effector

Base
 What is Inverse Kinematics?
• Inverse Kinematics

End Effector

Base
 Description of Links
• From a geometric viewpoint, the link defines the
relative position and orientation of joint axes at its
two ends

link length

Two parameters ai-1 and αi-1 are known

as link parameters and are constant for
link twist a given link
Link-connection description
 Link & Joint
• Link and Joint Parameters
– Joint angle  i: the angle of rotation from the Xi-1 axis to the
Xi axis about the Zi-1 axis. It is the joint variable if joint i is
rotary.
– Joint distance d i : the distance from the origin of the (i-1)
coordinate system to the intersection of the Zi-1 axis and the
Xi axis along the Zi-1 axis. It is the joint variable if joint i is
prismatic.
– Link length ai : the distance from the intersection of the Zi-1
axis and the Xi axis to the origin of the ith coordinate system
along the Xi axis.
– Link twist angle  i: the angle of rotation from the Zi-1 axis to
the Zi axis about the Xi axis.
56
Link-connection description

variable variable
offset di angle θi
Kinematic Modeling of the manipulator
 Denavit-Hartenberg (DH) Notation
• “The definition of a manipulator with four joint-link
parameters for each link and a systematic procedure
for assigning right-handed orthonormal coordinate
frames, one to each link in an open kinematic chain,
was proposed by Denavit and Hartenberg and is
known as the Denavit - Hartenberg notation or the
DH notation”
• Base frame or inertial reference frame is frame {0}
• Tool frame is frame {n}
• Frame {i} is rigidly attached to distal end of the link
{i}
• An n-DOF manipulator will have (n+1) frames
Denavit – Hartenberg (DH) Notation
 Denavit - Hartenberg Notation
• Labeling scheme
Denavit-Hartenberg Notation
Denavit-Hartenberg Notation
Denavit-Hartenberg Notation
 Denavit-Hartenberg Notation
• Once the frames are assigned, the joint-link parameters are
easily identified.

• These parameters are used to develop the Direct Kinematic

Model.

• Whenever there is a choice in frame assignment, such an

assignment is used which results in as many zero
parameters as possible.
 Manipulator Transformation Matrix

• Consider the following frames:

 Manipulator Transformation
Matrix
• Manipulator Transformation matrix is given by:
 Kinematic Relationship between Adjacent
Links
• Consider the following links:
 Kinematic Relationship between Adjacent
Links
• The transformation from frame {i-1} to frame {i}:
 Manipulator Transformation Matrix

• Consider the following frames:

 Example-1: 2-DOF Planar Manipulator Arm
• Obtain the position and orientation of the tool point P with
respect to the base for the 2-DOF, RP manipulator as shown in
following figure:
 Example-1
• Kinematic Model:

Kinematic Model:
For θ1 = 120° and d2 = 200 mm:
Example-2: Kinematic Model of a Cylindrical Arm
• Formulate the forward kinematic model of the 3 DOF (RPP)
manipulator arm as shown in figure:
 Example-2
• Transformation Matrices:
 Example-3: Articulated Arm Kinematic Model
• Formulate the forward kinematic model of the following manipulator:
 Example-3
• Transformation Matrices:
 Example-4: RPY wrist Kinematics
• For the following RPY wrist obtain the direct Kinematic model:
 Example-4
• Transformation Matrices:
 Example-5: Kinematics of a 3-DOF Polar Arm
• For the given 3-DOF Polar Arm, obtain the orientation and position
of Tool-Point P if the joint-link variable is q=[ 90° -45° 100 mm]T:
 Example-5
• Transformation Matrices:
 Example-6: SCARA Robot Kinematics
• For the following SCARA manipulator obtain the direct Kinematic model:
 Example-6
• Transformation Matrices:
 Example-6
• Kinematic Model:
•
Example-8: Stanford Robot Kinematics
For the following manipulator obtain the direct Kinematic model:
 Example-8
• Joint-link Parameters:
 Example-8
• Transformation Matrices:
 Example-8
• Kinematic Model:
Inverse Kinematics
• Inverse Kinematic problem is :

• “The determination of all possible and feasible sets

of joint variables, which would achieve the specified
position and orientation of the manipulator’s end-
effector with respect to the base frame”

• The problem of direct and inverse kinematics are of

fundamental importance for Industrial robots

88
Inverse Kinematics
• Configuration (Position and Orientation)
• Configuration Space (Cartesian Space)
• Joint Vector
• Joint Vector Space (Joint Space)

89
Inverse Kinematics
Kinematic Descriptions

Solvability of Inverse Kinematic Model

• Inverse Kinematic is complex
• It involves transcendental functions
• Number of simultaneous equations is also generally more than
the number of unknowns, making some of the equations
mutually dependent
90
Solvability of Inverse Kinematic Model
• Examine the direct kinematic model from mathematical perspective
• For a manipulator to have all position and orientation solutions, the
number of DOF n (equal to the number of unknowns) must at least
match the number of independent constraints. That is for general
dexterous manipulation n>=6
• Case- Multiple solutions due to parallel axes of revolute joints

91
Solution Techniques
Two Approaches:
• Closed Form Approach
• Numerical Approach

• In closed form solution, joint displacement’s are determined as

explicit functions of the position and orientation of the end-
effector.
– Closed form approach is a Algebraic approach
– A composite approach is based on direct inspection, algebra and inverse
transform is used in this approach

• Numerical Approach is iterative in nature… will not discuss here

92
Closed Form Solutions

(e) In the inverse kinematic model, the right hand side of Direct Kinematic
Model is known, while the left-hand side has n unknowns (q1, q2, …, qn). The left
hand side consists of product of n link transformation matrices, that is

Also, i-1Ti is a function of only one unknown qi. Premultiplying both sides by the inverse of 0T1,
yields
93
Example-1: Articulated Arm Inverse Kinematics
• For the 3-DOF articulated arm, determine the joint displacements
for known position and orientation of the end of the arm point
Position and orientation of the
end-point of arm be given by

From direct kinematic model

94
Example-1: Solution
• Step-1: Finding the single ones

• Step-2: Inverse Transform Approach

95
Example-1: Solution
• Step-2: Inverse Transform Approach

96
Example-1: Solution
• Step-3: The last unknown

Can you think of alternate ways to find the same solution? (With same or different
techniques)

97
Example-2: Inverse Kinematics of RPY Wrist
• For the 3-DOF wrist, determine the solution for three joint
variables for a given end-effector orientation matrix:
Let the end effector orientation
matrix be given by: n, o, a

From direct kinematic model

98
Example-2: Solution
• Use the approach of pre-multiplying:

99
Example-2: Solution
• Finding the displacement variables:

• Finding other displacements:

100
Example-2: Solution
• Solution by post-multiplication:

101
Example-2: Solution
• Finding unknowns:

• On similar principles:

102
Example-3: Inverse Kinematics of SCARA Manipulator
• Solve the problem of Inverse Kinematics for the 4-DOF SCARA
Manipulator
From direct kinematic model:

103
Example-3: Inverse Kinematics of SCARA Manipulator

104
❖Sources:
❖Internet sources ….. animation
❖Introduction to Robotics (Mechanics & Control) by JJ Craig
❖Book - Robotics and control by Mittal and Nagrath
❖Book – Mechatronics – W. Bolton

105