Special Issue - 2016 International Journal of Engineering Research & Technology (IJERT)

ISSN: 2278-0181
NCARMS - 2016 Conference Proceedings

Design, Simulation and Analysis of

Self-Reconfigurable Robot
P. Parthiban1, R. Mathiyazhagan2
Madras Institute of Technology
Anna University
Chennai, India

Abstract—This paper presents a hybrid robot that mobile robots in environments that is hazardous and
achieves locomotion through legs and wheels. The fixed mobile unapproachable to humans [18]. These robots are able to
robots not become accustomed themselves to the operate as a swarm and provide higher versatility [4]. They
environmental changes. To overcome this, the self have also categorized the reconfigurable robots based on
reconfiguring robots are able to purposely modify their
structural properties as Mobile Configuration Change and
individual shape by rearranging the connectivity of their parts
to meet particular operational demands. Recently, there has Whole Body Locomotion [17]. The speed achieved is same
been a great concern in using self-reconfigurable robots in when the robot walks on smooth as well as rough terrains
applications such as space exploration and rescue missions. [7]. The robot can navigate using wheels on both flat and
Designing and controlling of self-reconfigurable robots is a rugged terrains. The performance of the motor can be
difficult task. Hence, the research has mostly been focused on utilized by the introduction of motor driver which always
emergent systems that can function in a controlled provide maximum output current than input current of the
environment. The future work of this paper is to develop the motor [11].
robot with various modes to navigate in any terrain. Each robot’s mission is to reach the destination even in
unstructured terrains and avoiding collision with the other
Keywords—Mobile robots; hybrid robots; self-reconfigurable
robots [4]. The self reconfiguration robot should be aware of
robots; kinematics; robotics simulation
their environment and their own circumstances through a
I. INTRODUCTION sequence of sensors [3]. The fusion of information from
various sensors is a tedious process for autonomous decision
Reconfiguring robots will be most suitable for human
making [1]. The operations stated are to be achieved by
environment with greater effectiveness than any other type
using minimum number of actuators, thereby decreasing the
of robot. It is hoped that eventually reconfiguring robots can
mechanical complexity, programming complexity and cost
be used to perform tasks which are difficult, dirty and
[22]. So this project is undertaken to make use of the
dangerous (3D) for humans. This design is intended to
property of transforming on its own with help of various
operate in an uneven and unbalanced environment with the
sensor that could make it possible to help in the field of space
facility of sensing and communicating with other modules
exploration, rescue missions etc, where human faces a fatal
[2]. This includes working in extreme environmental
damage [9].
conditions or as an aide to humans in similar situations. The
Based on the analysis and study, the interaction of the
fixed mobile robots not turn out to be used to the
robot with the environment and robot locomotion plays a
morphological changes. Morphogenetic is the approach of
vital role in the selection of simulation [6]. The secondary
multi cellular organisms inspired by the emergent
criteria are close to reality, open source and usage of same
improvement [24]. These types of robots are well-suited for
code for simulation and real robot. V-Rep is an Integrated
locomotion in strange environments.
Development Environment and supports multiple physics
There are varieties of locomotion mechanisms to move
engine (ODE, Bullet and Vortex) [20]. The simulation
and hence robot’s locomotion approach plays a vital role in
results have confirmed that the proposed method can
mobile robot design. Some of them are walk, run, fly, slide
effectively direct the robot to change their locomotion [5].
and skate etc [8]. Generally the mobile robots can locomote
either through wheels or by leg. Self repairing, self II. QUADRUPED ROBOT
sustaining, self replication and self extension are some of the
The proposed approach implements a wheeled
challenges mentioned in these methods [16].
quadruped robot. There are various design considerations
Biological systems succeed in moving through a variety
when designing the robot. Self Reconfiguring robots require
of different environments. Therefore, it can be advantageous
unique controller and actuator for each module and then the
to replicate their choice of locomotion mechanisms.
modules are integrated [18]. The locomotion of robots has
Recently, there has been a noteworthy progression in the
been extensively analyzed and studied. From those studies
field of reconfiguring robots. A variety of robots that are
and surveys, a quadruped robot with mammalian gait pattern
designed to operate in environments populated by humans
with static stability is planned to design and certain
has already been developed [21]. The Robot is developed
transformations are employed to make it work like a
especially to move in any terrain i.e. both walk using legs
versatile reconfigurable robot [13]. At least three feet should
and move on wheels. The modular and reconfigurable

Volume 4, Issue 26 Published by, www.ijert.org 1

Special Issue - 2016 International Journal of Engineering Research & Technology (IJERT)
ISSN: 2278-0181
NCARMS - 2016 Conference Proceedings

be placed on the ground to achieve the static stability and are identical with servo motors are placed as frame motor,
maintaining the center of gravity within the support polygon. Hip motor, Knee motor and ankle motor [11]. The leg is
allowed to swing only when the neighboring legs perceive
ground contact.
The robot can be modeled by using rectangular frame,
links, motors and wheels. The battery which supply power
to the motor is placed to the slots provided in the mechanical
setup. This is the prototype model created using catia in
order to work out what could be the ultimate shape of the
robot. The robot has to be planned by considering the
following parameters such as wheel diameter, motor
dimensions of servo motor and DC motors.

Fig. 1. Block Diagram of the proposed Robot.

The block diagram of the proposed self reconfiguring

robot is shown in the figure 1. The power is supplied through
battery to various components. These components are
controller board, driver circuit board, sensor modules and
motors. The obstacle and variation in the environment is
sensed through sensing modules and send to the controller Fig. 2. Proposed CAD Model – Leg Configuration.
for the decision making. The data are processed by the The wheel configuration can be initiated by getting
controller and the required signals are sent to the motors feedback from the sensors and achieved by controlling the
through the driver board. The sensing modules continuously servo motor. The wheel configuration is suitable for flat and
monitor the environment for any change. smooth surfaces.
A. Conceptual Design
Primarily the reconfiguring robot was conceived with
sixteen degrees of freedom with four degrees of freedom per
leg (Fig 4). Due to difficulties faced in controlling greater
number degrees of freedom, this robot is redesigned with
three degrees of freedom per leg so as to have twelve degrees
of freedom. Lastly, a new design was arrived with the
knowledge gathered from developing previous models. The
new design has got twelve degrees of freedom with three
degrees of freedom per leg (Fig. 1). This design has more
stability with equal weight distribution on all the legs.

B. CAD Model for Proposed Robot Fig. 3. Proposed CAD Model – Wheel Configuration.
The joints of the legged robots have to operate in a
particular fashion as this is discrete systems [13]. The In second design, the number of actuators is reduced to
challenge of mobile robots is achieving controlled build the robot simpler [19]. In this project, 2 types of U-
locomotion in rough terrains. Walking robot may be a shaped bracket is used for joints formation. The servo motor
plausible solution to the problem [12]. But this makes the holder consists of two parts namely Servomotor bracket i
robot motion slow. On a flat terrain there is no need of legs and ii (Fig. 2). Servomotor can be fitted into the bracket i
for locomotion because wheeled motion is very much using the provision given in the bracket and the bracket ii
suitable for fast and less power consuming navigation. Thus, can be used to pass on the output of the servomotor. Bracket
wheeled robots are very energy efficient on hard surface i and ii are coupled using servomotor horn on one side. By
even at high speed. using the brackets there is a greater elasticity and other joints
In first design, 4 servo motors are used in each leg with cannot be disturbed while actuating the individual joint. The
DPST micro switch. The contact between the foot of the design of the servo motor holders and brackets are based on
robot and surface of the ground can be estimated for the the selection of the motor. The dimensions of bracket i and
control system of the robot. The model consists of a bracket ii are 32x40x75mm and 60x40x50mm respectively.
rectangular frame with 4 servo motors on each leg. These
motors are connected with the frame through links. All legs

Volume 4, Issue 26 Published by, www.ijert.org 2

Special Issue - 2016 International Journal of Engineering Research & Technology (IJERT)
ISSN: 2278-0181
NCARMS - 2016 Conference Proceedings

i, ii
Fig. 4. CAD Model of Servo Motor Brackets.

Each part of the robot body is designed by part Fig. 6. Co-Ordinate Diagrams for the Front Legs.
designing tools in CATIA and the parts are assembled to
form the body structure by using the assembling tool. The The D-H method allows four basic transformations lead
parts of the body are Chassis, Servo motor frame (brackets) to a unique homogeneous transformation matrix iLi-1 with
and leg (limb). Initially the components sizes were analyzed four variables representing the relationship between the
and the optimum size for the robot was determined. Robot links which is given by the equations.
size plays a major role. Based on the size of the robot, the
Cost of the Project and requirement of the actuators can be iLi-1 =[rot (z, θi-1) tran(0, 0, di-1) tran(ai-1, 0, 0) rot(x, αi-1)]
determined. The Fig. 6 shows the second design with
reduced number of servo motors and simpler mechanical (1)
design than the previous versions.
cosθi −cosαi sinθi −sinαi sinθi acosθi
sinθi cosαi cosθi sinαi cosθi sinθi
iLi-1= (2)
0 sinαi cosαi di
[ 0 0 0 1 ]

The DH convention of a 4-DOF robotic leg and

correspondingly the transformation matrices are computed
by using the above equations for the joints and these can be
explained by the following table.

TABLE 4.1 DH CONVENTION OF 4-DOF ROBOTIC LEG

Joint θi αi ai di
1 Θ1 -90º 0 0
2 Θ2 90º a1 0
Fig. 5. Proposed New design CAD Model. 3 Θ3 0º a2 0
4 Θ4 0º a3 0
C. Kinematics of the Robot
In this work, a four-legged and wheeled robot is The transformation matrix from foot to hip is the
developed especially to navigate in any kind of terrains i.e. concatenated transformation from link-to-link. Because both
both walk using legs and move on wheels [13]. The legs have the same mathematical representation, only one
controller will send the required position and orientation of matrix had to be calculated for both legs. This matrix,
the motor for each leg and robot body by using the kinematic considers as a base coordinate system, the origin of the hip
model [23]. The main role of these processors is to convert servos axis of each leg. By homogeneous transformation, the
the desired joint angle values for the joint positions into position and orientation of the base frame will be given by
accurate motor control pulses. The knee of each leg was the equations,
designed with a DC motor with the wheel. cosθ1 0 sinθ1 0
sinθ1 0 −cosθ1 sinθ1
0L1= (3)
0 −1 0 0
[ 0 0 0 1 ]

Volume 4, Issue 26 Published by, www.ijert.org 3

Special Issue - 2016 International Journal of Engineering Research & Technology (IJERT)
ISSN: 2278-0181
NCARMS - 2016 Conference Proceedings

cosθ2 0 sinθ2 a1 cosθ2 The above equations are solved to give certain angles
that can be employed for pure rotation of motors while
sinθ2 0 −cosθ2 sinθ2
1L2 = (4) reconfiguring of the robotic parts. The analog servo motors
0 1 0 0 have an integrated potentiometer in the position control
[ 0 ] mode to enable the closing of the feedback loop. Each servo
0 0 1
motor has to be manually calibrated to avoid the errors in
cosθ3 −sinθ3 0 a 2 cosθ3 position control.

sinθ3 cosθ3 0 sinθ3 III. SIMULATION MODEL

2L3 = (5)
0 0 1 0 Simulation software is a tool for testing behaviors before
trying them on the real robot. After the kinematic analysis is
[ 0 0 0 1 ] done, the V-Rep simulator is used to simulate the robot in an
integrated development environment. Robot walk can be
cosθ4 −sinθ4 0 a 3 cosθ4 evaluated by stability, velocity attained and torque required
sinθ4 cosθ4 0 sinθ4 at each joints.
3L4 = (6)
0 0 1 0
[ 0 0 0 1 ]

The matrices are multiplied to get 0T4 and then the

following link length given in the equations are submitted,

a = [0 0.40 0.52 1.2] (7)

The total transformation between the body of the robot

and foot of the robot leg is

0T4 = 0L1.1L2.2L3.3L4 (8)

=
nx ox a x px
Fig. 7. Inverse Kinematics in V-REP Simulator.
ny oy ay py
(9) The gait pattern of this robot is determined prior to the
nz oz az pz
design and fabrication of the model. The developed model
[0 0 0 1] is simulated for statically stable amble gait. The Fig. 8 shows
WHERE, the demonstration of the robot in leg mode using Virtual
Robot Experimentation Platform simulator environment.
nx=cθ1 cθ2 (cθ3 cθ4 − sθ3 sθ4 )+sθ1 (−sθ3 cθ4 + cθ3 sθ4 )
(10)
ox=cθ1 cθ2 (−cθ3 sθ4 −sθ3 cθ4 )+ sθ1 (−sθ3 sθ4 cθ3 cθ4 )
(11)
ax = cθ1 sθ2 (12)
px = cθ1 cθ2 (a 3 cθ3 cθ4 −sθ3 sθ4 + a 2 cθ3 ) + sθ1 (a 3 sθ3 cθ4
+cθ3 sθ4 + sθ3 ) + a1 cθ1 cθ2 (13)
ny = sθ1 cθ2 (cθ3 cθ4 −sθ3 sθ4 )-cθ2 (−sθ3 cθ4 + cθ3 sθ4 )
(14)
oy= sθ1 cθ2 (−cθ3 sθ4 −sθ3 cθ4 )-cθ2 (−sθ3 sθ4 + cθ3 cθ4 )

(15)
ay = sθ1 sθ2 (16)
py = sθ1 cθ2 (a 3 cθ3 cθ4 −sθ3 sθ4 + a 2 cθ3) – cθ2 (a 3 sθ3 cθ4
+cθ3 sθ4 + sθ3 ) + a1 cθ1 cθ2
Fig. 8. Leg mode GAIT Pattern in V-Rep.

(17)
The wheel mode can be obtained whenever the controller
nz = -sθ2 (cθ3 cθ4 −sθ3 sθ4 ) (18)
receives more values from the proximity sensor than the
oz = -sθ2 (−cθ3 sθ4 −sθ3 cθ4 )
reference values. The Fig. 9 shows the switching between
(19)
leg mode and wheel mode. The hybrid structure is used to
az = - cθ2
optimize the performance in terms of speed and stability
(20)
when the robot takes different kinds of transformations [15].
pz = -sθ2 (a 3 cθ3 cθ4 −sθ3 sθ4 + a 2 cθ3 ) - sθ2
(21)

Volume 4, Issue 26 Published by, www.ijert.org 4

Special Issue - 2016 International Journal of Engineering Research & Technology (IJERT)
ISSN: 2278-0181
NCARMS - 2016 Conference Proceedings

Wheels act as foot while walking motion and do rolling programming, there is a need for the change of programming
motion in wheel mode. language in a real environment.

IV. CONCLUSION
An extensive Literature Survey conducted for the work
gave profound insight on the requirements for building the
robot. Based on the inferences obtained from the survey, the
further development for the paper was carried out. CAD
model has been created and analyzed using virtual robot
experimentation platform. The angles obtained from the D-
H matrix can be used to give inputs to the servomotor. These
can be analyzed by the corresponding torques at various
times for the joints during simulation. The robot’s
locomotion can be obtained by placing three legs on the
ground thus follows stable amble gait. The angular velocity
for each joint can be recorded over the period of time.
V. REFERENCES
[1] Baarir S., Hillah L. M., Kordon F. and Renault E. (2010), ‘Self
Fig. 9. Wheel mode GAIT Pattern in V-Rep. Reconfigurable Modular Robots and Their Symbolic Configuration
Space’, 16th Monterey Workshop, Redmond, WA, USA, pp 103 –
Simulations are carried out for different configurations 121.
and the assumptions for the analysis are robots walk in [2] Behnam S., Mark M. and Wei-Min S. (2006), ‘SUPERBOT: A
amble gait pattern only. This can be shown in the Fig. 10. Deployable, Multi-Functional, and Modular Self-Reconfigurable
Robotic System’, IEEE International Conference on Intelligent
Similarly, the requirement of the torque at each joint for the Robots and Systems, pp 3636-3642.
actuators can be determined from the Fig. 11. [3] Bobovsky Z., Novak P., Krys V. and Kot T. (2014), ‘The Module for
a Self-Reconfigurable Robotic System’, IEEE International
Symposium on Applied Machine Intelligence and Informatics, pp
275-282.
[4] Bojan J., Benjamin M., Erik M. (2008), ‘Swarm Intelligence for Self
Reconfiguring Walking Robot‘, IEEE International Symposium on
Swarm Intelligence, pp 1 – 8.
[5] Chun-Hsu K., Kuu-Young Y. and Yi-Hung H. (2015), ‘Optimized
Trajectory Planning for Mobile Robot in the Presence of Moving
Obstacles’, IEEE International Conference on Mechatronics, pp 70 –
75.
[6] Cobano J. A., Vera S., Heredia G. and Ollero A. (2015), ‘Safe
Trajectory Planning for Multiple Aerial Vehicles with Segmentation-
Adaptive Pseudospectral Collocation’, IEEE International
Conference on Robotics and Automation, pp 29 – 34.
[7] Dongil C. and Jun-Ho O. (2011), ‘Four and Two Wheel
Transformable Dynamic Mobile Platform’, IEEE International
Conference on Robotics and Automation, pp 1-4.
[8] Hristo B., Arancha C. And Tad H. (2000), ‘Emergent Structures in
Fig. 10. Angular Velocity (rad/s) vs Time(s). Modular Self-Reconfigurable Robots’, IEEE International
Conference on Robotics and Automation, pp 1734-1741.
[9] Jie Z., Xindan C., Yanhe Z. and Shufeng T. (2011), ‘A New Self-
Reconfigurable Modular Robotic System UBOT: Multimode
Locomotion and Self-Reconfiguration’, IEEE International
Conference on Intelligent Robots and Systems, pp 1020-1026.
[10] Jingfu J., Yoon G., Sung-Gil W. and Nicholas G. (2015),
‘Decentralized Cooperative Mean Approach to Collision Avoidance
for Nonholonomic Mobile Robots’, IEEE International Conference on
Robotics and Automation, pp 35 – 41.
[11] Junichi U., Yuto N., Kei O. and Masayuki I. (2010), ‘Design of High
Torque and High Speed Leg Module for High Power Humanoid’,
IEEE International Conference on Intelligent Robots and Systems, pp
4497-4503.
[12] Keith K. and Daniela R. (2005), ‘Efficient Locomotion for a Self-
Reconfiguring Robot’, IEEE International Conference on Robotics
and Automation, pp 2963-2969
[13] Kenta N. and Yasutaka F. (2015), ‘The Stable Wheeled Locomotion
Fig. 11. Torque (Nm) vs Time (s). in Low Speed Region for a Wheel-Legged Mobile Robot’, IEEE
International Conference on Mechatronics, pp 404-409
There will always be a difference between simulation [14] Gor M.M., Pathak P.M., Samantaray A.K., Yang J.M. and Kwak S.W.
(2015), ‘Control Oriented Model – Based Simulation and
and reality and hence the gap should be taken into account. Experimental Studies on a Complaint Legged Quadruped Robot’,
As this simulator software supports script language for cornell university”.

Volume 4, Issue 26 Published by, www.ijert.org 5

Special Issue - 2016 International Journal of Engineering Research & Technology (IJERT)
ISSN: 2278-0181
NCARMS - 2016 Conference Proceedings

[15] Mark Y., Paul W., Michael P. and Jimmy S. (2009), ‘Modular Self [21] Sprowitz A., Moeckel R., Vespignani M., Bonardi S. and Ijspeert A.
Reconfigurable Robots’, School of Engineering and Applied Science, J. (2014), ‘Roombots: A Hardware Perspective on 3D Self-
University of Pennsylvania, Philadelphia, USA Reconfiguration and Locomotion with a Homogeneous Modular
[16] Mark Y., Wei Min S., Behnam S., Daniela R., Mark M., Hod L., Eric Robot’, Robotics and Autonomous Systems, Volume 62, Issue 7, pp
K. and Gregory S.C. (2007), ‘Modular Self Reconfigurable Robot 1016–1033.
Systems’, IEEE International Conference on Robotics and [22] Tom L. and Subramanian R. (2010), ‘An Efficient Algorithm for Self
Automation Society, pp 43 – 52. Reconfiguration Planning in a Modular Robot’ IEEE International
[17] Paul M. and Pinhas B. (2012), ‘Modular and Re-Configurable Conference on Robotics and Automation.
Robotics’, Robotics and Automation Systems, pp 1648-1666. [23] Xinjie W., Liangwen W., Chunmei P., Yuxiao Z., Weigang T. and
[18] Ren Z. and Zhu Y. (2011), ‘Self Reconfiguration Planning of Self- Xiaohui Z. (2010), ‘Inverse Kinematics Analysis of Multi-Legged
Reconfigurable Robot based on Windmill like Crystal Cell Group’, Walking Robots Based on Hand-Foot Integration Mechanism’, IEEE
IEEE International Conference on Management Science and International Conference on Mechatronics and Automation.
Industrial Engineering, Harbin, China, pp 1218 – 1222. [24] Yan M., Yuyang Z., Abhay S., Yaochu J. and Bernhard S. (2011),
[19] Robert F., Zack B. and Daniela R. (2005), ‘Reconfiguration Planning ‘Cross Ball: A New Morphogenetic Self Reconfigurable Modular
Among Obstacles for Heterogeneous Self-Reconfiguring Robots’, Robot’, IEEE International Conference on Robotics and Automation,
IEEE International Conference on Robotics and Automation. Shanghai, China, pp 267 – 2
[20] Serena I., Vincent P. and Francesco N. (2014), ‘Tools for Dynamics
Simulation of Robots: A Survey Based on User Feedback’, Cornell
University Library.

Volume 4, Issue 26 Published by, www.ijert.org 6