Design of a Stepping Hexapod Bio-Inspired Robot

Răzvan Ștefan Dologa Honoriu Vălean Alexandru Ianosi

Automation Department Automation Department Department of Mechatronics and
Technical University of Cluj-Napoca Technical University of Cluj-Napoca Machine Dynamics
Cluj-Napoca, Romania Cluj-Napoca, Romania Technical University of Cluj-Napoca
2024 IEEE International Conference on Automation, Quality and Testing, Robotics (AQTR) | 979-8-3503-6193-3/24/$31.00 ©2024 IEEE | DOI: 10.1109/AQTR61889.2024.10554226

[email protected] [email protected] Cluj-Napoca, Romania

[email protected]

Abstract: This paper introduces a hexapod robotic platform II. RELATED WORK
designed for versatile locomotion and modularity. A. Type of Robots
Characterized by its distinctive six-legged configuration, the
hexapod achieves an optimal balance between stability and Stepping robots can be classified based on both the
speed. The primary objectivewas to establish a comprehensive, number of legs they possess and the organism whose
cost-effective, and highly modular platform serving as a movements they seek to replicate [1]. Legged robots have
foundational base for future development. To provide context, demonstrated a distinct advantage particularly in situations
the paper begins by offering an overview of the current state of involving soft surfaces (sandy terrain). Those leave a way
the industry, highlighting exemplarylegged robots from various significantly smaller footprint when compared to wheeled or
categories. The development phases of the platform are tracked robots, resulting in more energy-efficient movement.
systematically presented across three chapters, eachfocusing on This efficiency is attributed to the reduced soil deformation in
the significant contributions within its respective domain. the smaller contact surface [2]. Another notable advantage is
Beginning with the motivation behind the selection of specific the ability of the hexapod like robots to change their
components and the foundational principles that underpin the orientation by rotation at fix point and to adjust their height in
platform design continuing with the electrical diagram and the order to compensate for variations in terrain. Unlike other type
associated PCB for the electrical part of the platform and ending of robots, hexapods can keep on moving despite one or more
with kinematics principles and the control structure used to legs becoming unusable, as long as the robot manage to
remotely operate the platform. Subsequently the paper delves maintain its center of weight inside the polygon described by
into the kinematic principles underlying the hexapod’s
the position of the remaining legs. In specific scenarios, these
movement, marking the core concepts that govern its
robots have been programmed to use their extra legs or front
locomotion. Finally, the inverse kinematics equations are
introduced, laying the groundwork for future advancements
legs to lift and relocate different objects. A series of
and refinements in the platform’s capabilities. drawbacks for hexapod robots can be observed in their
relatively low movement speed, the necessity of a high
Keywords: Hexapod; CAD; PCB; Kinematics; modular number of actuators (depending on the number of degrees of
platform; biomimetic; application. freedom) while maintaining a complex structure with an
intricated control algorithm. Current research in the industry
I. INTRODUCTION of walking robots is directed towards refining control
In the contemporary landscape robots are progressively algorithms to enhance the precision and efficiency of robotic
becoming more ingrained in the fabric of our everyday reality movements. Another significant focus is spent on the
while our reliance on them continues to grow. While wheeled trajectory determination, seeking more effective methods for
robots may seem to dominate the robotics scene, they are platforms to navigate their surroundings. Simultaneously,
subjects of constraints of both the terrain they navigate and there is an emphasis on developing dynamic interpretation of
potential obstacles that may arise such as variations in the data collected from the environment, enabling robots to
elevation. A potential solution for those limitations can be adapt and respond to continuously changing conditions. The
seen in the legged robots designed to traverse uneven surfaces final goal is to propel working robots towards greater
and surmount obstacles. The platform possesses the ability to adaptability autonomy and overall performance.
navigate and access spaces and locations that are inherently B. Current State
inaccessible to larger robotic counterparts being very
modular and customizable while using cost-effective parts. In the present industry landscape various types of walking
The development of the biomimetic hexapod can be divided robots exist each showcasing a distinct number of legs. For
into three different chapters, each having its own set of bipeds, Atlas by Boston Dynamics [3] is a notable example,
associated sub-chapters. The initial chapter focuses on the while Spot also by Boston Dynamics [4] represents the
mechanical structure of the robot, drawing its inspiration from tetrapod category. Lauron V, the hexapod walking robot by
nature’s living organisms for design insights and functionality FZI (Forschungszentrum Informatik) [5] represents the
purposes. The second major development phaseencompasses hexapod category being engineered to navigate and traverse
the creation of the necessary electricalframework required for diverse environments while showcasing impressive stability
the seamless operation of the entire assembly. The final and adaptability. This robot designs allows it to handle
chapter explores the contour structures and compassing challenging terrains, making it suitable for applications
elements such as algorithm implementation methods, it ranging from exploration in rough landscape to tasks that
considers both direct and inverse kinematics realtime control
and the necessary application and hardware components
essential for remote platform control.

Authorized licensed use limited to: Zhejiang University. Downloaded on December 04,2024 at 14:25:20 UTC from IEEE Xplore. Restrictions apply.
demand enhanced mobility and stability such as search and A. Mechanical Structure
rescue missions. a) Components
In recent years, stepping robotics research has focused on The hexapod design incorporates three degrees of
refining essential aspects in terms of performance. These freedom for each of its six legs, requiring a total of
include developing optimal algorithms to enhance efficiency, eighteen servomotors (MG995). To facilitate
creating precise trajectory determination programs for simultaneous control of these servomotors, a
accurate movement and enabling dynamic interpretation of development board with the necessary number of
the environmental data. Those endeavors endowed those connections have been chosen (Arduino Mega 2560).
platforms with the capability to navigate through areas strewn Powering the platform is a set of two accumulators
with obstacles in the shortest time possible. Another great providing the required energy for the whole assembly. For
focus in their development was the modularity of the platform, remote control of the platform a Bluetooth module (HC-
those not having a single predefined role but rather a large set 05) is integrated into the platform. The selection of these
of functionalities and domains of activity in which those could components has been methodical, taking into
operate. consideration the choices made in similar projects. Each
As for the hexapod robots, these strike a good balance component has been chosen based on its potential
between speed and equilibrium, distinguishing from both performance and inherent limitations, ensuring a balanced
biped and tetrapod robots due to the wide range of movement and effective integration into the overall hexapod design.
options offered by their six legs. This type of platform
provides superior stability, as at any given time, a minimum b) Hexapod CAD Modeling
of three legs supports the robot. Additionally, it offers the In the creation of the robot parts, the design process
capability to swiftly alter its direction of movement. In most was facilitated by Autodesk's Fusion 360, a computer-aided
cases, each leg of this type of robot is designed as a robot of
design (CAD) program. Offering a versatile array of tools,
type RRR (three degrees of rotation robot), with individual
motors serving as articulations, forming pairs to facilitate this software enables users to design, model, render, and
movement. Examining the legs symmetry, two types of simulate both 2D and 3D objects. The program adheres to a
hexapod robots come into focus: hexagonal and rectangular. top-down approach, breaking down a complex system into a
The latter, characterized by vertical leg symmetry, draws series of subsystems (components). This iterative process
parallels with specific insects [6]. Platforms of these types can continues until the fundamental pieces (bodies) that form the
be encountered at various stages of development while having complete assembly are successfully designed. (Fig. 1)
various applications. Some are undergoing testing for
industrial purposes (Spot by Boston Dynamics) while others
serve as learning platforms (Lauron) providing valuable
insights and hands-on experience for researchers, engineers
and students. Among the commercially available advanced
legged robots, some showcase abilities to move objects, sort
different items based on various criteria (color/shape) and to
determine the exact position of the end effector. Achieving
this precision involves employing algorithms that compare the
torque of the servomotors with a set of values obtained from
different simulations.
I. MATERIAL AND METHOD
The development of the hexapod robot involves three
pivotal phases: mechanical structure, electrical structure, and Fig. 1. Final variant of the Hexapod
the control structure. In this approach, the characteristics of each part are
The initial phase centers around component selection defined in relation to the entire assembly. This ensures that
along with 3D modeling of the platform. The model is then each element is intricately connected to the blueprint of the
transferred into physical components through 3D printing. design or the geometry of other parts, thereby preemptively
The assembly process follows, with through testing ensuring mitigating potential conflicts. The initial phase involved
mechanical performance and structural integrity. The second designing the first leg of the robot starting from a 3D model
phase focuses on the creation of the electrical infrastructure of one of the hexapod’s servomotors. This approach allowed
beginning with a detailed electrical diagram and finishing with for a seamless integration of the individual leg components,
the printed circuit board resulting in a robust electrical ensuring combability with the chosen components while also
structure. The third and final phase revolves around the establishing a foundation for the overall hexapod design.
control structure employing direct kinematics. The integration After completing the design of the first leg, the entire
of the direct kinematics with mechanical and electrical assembly was imported into a separate project and replicated
components along with control algorithm ensures a cohesive six times. The intention was to create interchangeable
and precise system, contributing to the overall functionality of components, and to achieve this, the robot’s body consists of
the hexapod robot.
three identical plates connected by a series of clips. This
design choice ensures uniformity and interchangeability in
the overall structure of the platform, streamlining the
manufacturing process and allowing for efficient assembly.

Authorized licensed use limited to: Zhejiang University. Downloaded on December 04,2024 at 14:25:20 UTC from IEEE Xplore. Restrictions apply.
 The plates are equipped with a series of slots to accommodate advanced functionalities in subsequent stages of
the clamps, designated spaces for servomotors, and a series development.
of ribs on the lower part aimed at reinforcing the plates. In
addition to the plates, two trays have been incorporated to a) Direct Kinematics
house the electrical components. These trays include screw To attain real-time control of the robot’s movements,
holes and supports for precise positioning and fastening of direct kinematic principles were used in which each leg of the
electronical elements. Furthermore, they feature a set of holes robot is treated as an independent kinematic chain, with the
and channels intended to streamline cable management for tip of the leg representing the final effector [8]. Each
the electrical structure. The design was meant to ensure the kinematic chain being composed by a series of three
structural integrity, the modularity, and the functionality of segments connected with three rotation couplings, each
the platform. having a single degree of freedom (each servo rotates around
a) Printing, Assembly And testing a single axis). Therefore, the rotation of these segments will
Once modeled the parts, they were brought to reality using directly determine both position and orientation of the end
a 3D printer after the optimal internal structure in order to effector. Hence, direct kinematics involves determining the
achieve a favorable weight-to-structural resistance ratio. position and orientation of the final effector, knowing the
Following the printing process, a preliminary test was displacement from each rotational coupling and the direction
conducted on a singular leg at a reduced scale. Based on the of this displacement. This led to the analyzation of the
outcome of this test, adjustments were made to the design to required movements of the platform, breaking them into a
enhance support and stability during movements. This series of simple motions and determining the rotation angles
iterative process ensured that the final 3D-printed in each coupling as well as their sequence. Prior to executing
components would meet the desired standards of both any movement, it was imperative to establish the minimum
structural integrity and operational efficiency in the full-scale stability condition. This condition ensures that the platform
hexapod platform. maintains stability by keeping its center of weight within the
polygon described by the tip of its legs on the ground.
B. Electrical Structure Consequently, indifferent on the type of movement, three legs
The core of the electrical structure revolves around the must remain in contact with ground at all times to fulfill the
PCB (printed circuit board), created using Autodesk Eagle. minimum stability requirement. The entire structure operates
This process unfolds in two stages: the development of the in a cyclic manner, as the platform needs to continue taking
electrical diagram and the subsequentPCB design. Beginning steps as long as the movement commands are received from
with the chosen developmentboard, which, in this case, is the the control application. This cyclic nature ensures the
Arduino Mega, the electrical design commences by adjusting continual responsiveness and adaptability of the hexapod
the voltage to meet the desired specifications for each robot.
component based on its respective datasheet. Once that the b) Inverse Kinematics
connections were mapped out and the diagram completed, it Compared to the direct kinematics, in inverse kinematics
proceeds to the next stage. The finalized model is then we already know the position and the orientation of the final
forwarded to a CNC (Computerized Numerical Control) effector while the values for each kinematic couple have to be
millingmachine used to create the PCB prototype. (Fig. 2) determined [9]. The values of the rotation angles are
determined for each individual kinematic couple using a
geometric approach, involving the creation of right-angled
triangles through a series of projections. This geometric
process allows for the calculation of rotation angles,
employing principles such as Pythagorean theorem and cosine
rule. Both the method used and the formula for the first
rotation couple obtained using tangent formula can be
observed in Fig. 3.

Fig. 2. The final scheme of the PCB

C. Control Structure
The control structure of the hexapod is intricately
designed, incorporating principles of direct kinematics while
laying the groundwork for future development by
incorporating the inverse kinematics module for more
versatile movements and overall better control [7]. The
current design ensures a robust foundation that combines
immediate control capabilities with the potential for Fig. 3. Determination of the rotation angle θ1 for thefirst kinematic
couple.

Authorized licensed use limited to: Zhejiang University. Downloaded on December 04,2024 at 14:25:20 UTC from IEEE Xplore. Restrictions apply.
 (1)

The same method is applied to the following segments of

the considered leg, creating a new right-angled triangle with
one vertex at θ2 and the center at θ3. (Fig. 4)

Fig. 6. Determining the θ2 angle.

(
)
(5)

, ∙
Fig. 4. The new considerate triangle for θ2 and θ3. * + .
- ∙
(6)

Knowing the sides AB, AC (Pythagorean theorem), and ( * (7)

BC (the difference between z and segment L1), it is observed , ∙
+ .
- ∙
that θ3 can be determined from the triangle with the vertices
)
(8)
at point A, D, C by applying the cosine rule. (Fig. 5)
To
The resulting set of equations obtained from this geometric
approach forms the basis for implementing the inverse
kinematic model. These equations allow the calculation of
rotation angles for each coupling, starting from the dimensions
of the segments and the coordinates of the final effector. The
initial positioning of the legs is chosen keeping in mind to
describe a Polygon which will confer the platform with
enough stability. Initially the platform runs the calibration
code which sets the robot in the above-mentioned start
position. Subsequently, the code runs a loop continuously
checking for input from the application, while each input
corresponding to a different type of movement assigned to the
hexapod platform.
c) Control Algorithm
The control algorithm underwent a detailed study
involving the examination of similar platforms [10] in order
Fig. 5. The ADC triangle and the associated equations for the θ3 angle to assess their complexity and problem-solving capabilities.
were determined using the cosine formula. Initially, the robot was controlled through serial
communication, chosen for its ability to facilitate quick error
∙ 2∙ ∙ ∙ corrections during the development phase. In the final code,
cos 180
(2) communication shifted to the Bluetooth module combined
)
with a mobile application. To facilitate the remote control of
cos 180 the platform, a mobile application was developed using MIT
2∙ ∙
(3)
App inventor web application (Fig. 7). The application first
!" ! #$%& $ % $ %'
cos (4)
checks for the existence of a communication channel between
∙% ∙%' the mobile device and the hexapod’s Bluetooth module. Upon
confirmation, it then awaits input from the user, enabling
determine the θ2, it is divided into a difference of two angles, seamless and intuitive control of the hexapod robot through its
and a new point E in the plane is considered to create a new interface. Derived from a comparable scenario, the decision
right-angled triangle formed by the D, C, E points. was made to calibrate the servo motors in order to ensure that
Subsequently, arctangent is applied to both angles from the all motors commence from a reference position determined by
difference in order to determine their values, after which θ2 is factors like the initial platform position (hexapod standing on
found by substitution. (Fig. 6) its own legs), the direction of rotation (considering the mirror
positioning of hexapod legs) and the individual position of
each leg within the structure.

Authorized licensed use limited to: Zhejiang University. Downloaded on December 04,2024 at 14:25:20 UTC from IEEE Xplore. Restrictions apply.
 II. CONCLUSION AND FUTURE DEVELOPMENT
The platform presents several advantageous
characteristics lending it significant potential for diverse
applications within the realm of robotics. Its modular
architecture offers scalability and flexibility, allowing for
integration of additional components and customizations
meant to suit specific operational requirements. Moreover, its
utilization of cost-effective components further expands its
possible applications. Its compact dimensions make it able to
operate in constrained spaces, while maintaining relative
stability and precision in task execution. As for the limitations
of the platform, its modest speed and limited operational
range impose constraints on its suitability for applications
requiring extensive locomotion. Another limitation could be
represented by the lack of adaptability to variations in the
terrain along with the lack of protection for outdoor use
exposes it to environmental factors.
Fig. 7: Interface of the control application In terms of further development directions, the focus will
involve a series of steps aimed at enhancing the prototype
I. RESULTS autonomy. A pivotal step will be transitioning from direct
The present paper outlines the realization of a remotely kinematics to inverse kinematics in the movement algorithm,
controlled biomimetic platform based on the principles of utilizing the previously calculated model. Subsequently,
direct kinematics. The platform exhibits the capability to incorporating the Raspberry Pi 4 board in a master-slave
execute various spatial movements, including forward and configuration with the Arduino Mega 2560 board is planned
backward motions, as well as rotations to a fix point in both in order to provide the platform with increased processing
directions. These movements are dictated by commands power. Once this step is accomplished, the platform could
received through the Bluetooth module from an external leverage the camera module along with an image processing
application. The platform’s versatility extends to its ability to algorithm developed in Python using the OpenCV library
traverse different terrains, even while bearing additional loads [11]. This would enable the hexapod to exhibit a degree of
such as an external battery for enhanced autonomy, Moreover, autonomy, for instance, using image processing to recognize
its modular structure become obvious, allowing for the swift and categorize different objects or potential obstacles. This
replacement of a deformed part or a broken segment without capability would empower the platform to navigate its
necessitating a complete disassembly of the platform or of one surroundings, recognize potential entrances, and adapt to
of its legs. Noteworthy contributions include the design and various environmental conditions. Regarding potential
the mechanical structure, the creation of a dedicated PCB for applications, the platform could offer significant value in
the electrical structure to facilitate optimal movement search and rescue missions, particularly in earthquake-
freedom, and the symmetrical control algorithm along with affected areas. It has the potential to navigate through
the associated application. Additionally, the inverse kinematic hazardous environments without risking human lives,
model will further enhance the overall functionality and facilitating the search for survivors amidst rubble and debris.
performance of the platform. (Fig. 8) It’s essential to underscore that these capabilities may become
feasible once the platform is developed enough.
REFERENCES

[1] J. Wang, W. Chen, X. Xiao, Y. Xu, C. Li, X. Jia, Meng and M. Q.-H,
"A survey of the development ofbiomimetic intelligence and
robotics," Biomimetic Intelligence and Robotics, 2021.

[2] Silva, M. &. T. Machado and José., "A Historical Perspective of

Legged Robots," Journal of Vibrationand Control, vol. 13, pp.
1447-1486, 2007.

[3] S. Feng, X. Xinjilefu and C. G. A. a. J. Kim, "Optimization based

controller design and implementation for the Atlas robot in the
DARPARobotics Challenge Finals," IEEE-RAS 15th
International Conference on Humanoid Robots (Humanoids),
Seoul, Korea (South), 2015.

[4] Guizzo, "By leaps and bounds: An exclusive lookat how Boston
dynamics is redefining robot agility,"IEEE Spectrum, vol. 56, no.
12, 2019.

Fig. 8. Hexapod Platform final configuration. [5] A. Roennau, G. Heppner and M. N. a. R. Dillmann,"LAURON V:
A versatile six-legged walking robotwith advanced
maneuverability," International Conference on Advanced
Intelligent Mechatronics, pp. 82-87, 2014.

[6] T. Fukuda, F. Chen and Q. Shi, "Special Feature on Bio-Inspired

Robotics," Applied Sciences, vol. 8, no.5, p. 817, 2018.

Authorized licensed use limited to: Zhejiang University. Downloaded on December 04,2024 at 14:25:20 UTC from IEEE Xplore. Restrictions apply.
[7] OscarLaing, "Inverse Kinematics for Hexapod and [10] J. Coelho, F. Ribeiro, B. Dias, G. Lopes and P. Flores,"Trends in
Quadruped Robots," 2012. [Online]. Available: the Control of Hexapod Robots: A Survey," 2021.
https://oscarliang.com/inverse-kinematics- implementation-
hexapod-robots/. [Accessed 1 1 2024]. [11] J. Hong and K. T. a. C. Chen, "Obstacle avoidance ofhexapod
robots using fuzzy Q-learning," 2017 IEEE Symposium Series on
[8] Ding, X. &. Wang, Z. &. Rovetta and A. &. Z. J.M., Computational Intelligence (SSCI), 2017
"Locomotion Analysis of Hexapod Robot," 2010.

[9] RoboticsLearning, "Inverse Kinematics," [Online].Available:

https://www.rosroboticslearning.com/inverse- kinematics.
[Accessed 3 March 2022].

Authorized licensed use limited to: Zhejiang University. Downloaded on December 04,2024 at 14:25:20 UTC from IEEE Xplore. Restrictions apply.