Modeling and Simulation of Mobile Robot Based on VREP

Zhenzhen Zhang, Jixuan Liu, Guoping Wang, Chen Zhang, Yuntao Liu and Lingqi Yang

Xi'an Jiaotong University City College, Xi'an, China

Keywords: VREP, Mobile Robot, Modeling, Simulation, Kinematics, Dynamics.

Abstract: In view of the problems of high computational complexity and poor repeatability in traditional mobile robot

modeling and simulation methods, this paper studied the rapid modeling and simulation of mobile robot

based on VREP to improve computational efficiency and repeatability. Firstly, the motion analysis and

modeling of mobile robot are introduced, and the advantages of its application in mobile robot simulation

are analyzed. Then, the modeling method of mobile robot based on VREP is described in detail, including

the selection of robot model, the setting of kinematics and dynamics parameters. Finally, the feasibility and

effectiveness of the proposed modeling method are verified through simulation experiments using VREP

software.

INTRODUCTION

Mobile robot is a kind of robot system with wide

application prospect, which can move and perform

tasks autonomously in complex environment.

However, due to the complexity of its mechanical

structure and control system, traditional

experimental methods are difficult to meet the needs

of systematic research and testing of mobile robots.

Therefore, using simulation technology to model and

simulate mobile robots is an effective research

method. Traditional mobile robot modeling and

simulation methods suffer from high computational

complexity and poor repeatability. In addition, most

of the existing mobile robot modeling and

simulation tools require programming knowledge,

which is difficult for non-programmers to operate.

The mobile robot based on VREP is a mechanical

device that can autonomously perform certain tasks.

Their invention is aimed at helping or replacing

humans to perform a series of repetitive and

laborious tasks, such as production and

manufacturing, civil construction, or hazardous

work

(Rohmer, 2013).

The significance of this research lies in the fact

that through the motion analysis and modeling of

mobile robots, and the development of a mobile

robot modeling and simulation system based on

VREP, the research and development efficiency and

quality of mobile robots can be improved, and the

development and application of mobile robot

technology can be promoted.

MOBIL ROBOT MOTION

MODEL BUILDING

This section elaborates on the modeling method of

mobile robots based on VREP, including the

selection of robot models, setting of kinematic

and dynamic parameters, addition of sensors and

controllers, and other related details. The design of

the robot model is also explained in detail to ensure

the accuracy and credibility of the simulation results.

2.1 Mobile Robot Modeling

Methodology

In the process of robot modeling, it is essential to

have a certain methodology, which can provide

clearer ideas and guidance. There are three

commonly used robot modeling methodologies:

1)Analytical Modeling Methodology

Analytical modeling is a modeling method widely

used in the field of robotics. It is based on physics

and mathematical theory and uses formulas and

equations to describe the robot. This method has a

clear idea and is suitable for modeling complex

robot systems. It can highly simulate robot behavior

and carry out corresponding control and

optimization applications. But correspondingly, the

mathematical and physical knowledge required for

analytical modeling is in-depth and comprehensive,

and the quality requirements for modelers are also

higher.

Zhang, Z., Liu, J., Wang, G., Zhang, C., Liu, Y. and Yang, L.

Modeling and Simulation of Mobile Robot Based on VREP.

DOI: 10.5220/0012274200003807

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 2nd International Seminar on Artiﬁcial Intelligence, Networking and Information Technology (ANIT 2023), pages 75-79

ISBN: 978-989-758-677-4

2)Experimental Modeling Methodology

Experimental modeling is a method based on

experimental data and simulation verification. It

usually involves many aspects such as robot

movement and perception ability, and data sources

are also unique, such as sensors, simulators, and

field environments. The experimental modeling has

good adaptability and practicability, and can

effectively reflect the actual behavior and working

environment of the robot. However, if the data

source is not reliable, the modeling accuracy will be

affected accordingly.

3) Deep Learning Modeling Methodology

Deep learning modeling is a new technology in the

field of robotics in recent years. It uses the idea of

artificial neural networks to extract key features

from large amounts of data for modeling. Deep

learning modeling has good intelligence and

adaptive ability, which can make rapid response and

decision according to different situations. However,

the large sample set and powerful computing power

required for deep learning modeling are also issues

that cannot be ignored.

To sum up, establishing a suitable robot modeling

methodology is an important means to ensure the

effectiveness of robot simulation and control. Only

by selecting the corresponding methodology for

modeling, can we better simulate the actual behavior

of the robot and carry out the corresponding

optimization and application. In this paper, the first

and the second are combined to promote the

experiment by theory, and the experiment then

improves the theory, which complements each other,

and provides the foundation for the tasks of

autonomous navigation, path planning, obstacle

avoidance and so on.

2.2 Mobile Robot Motion Model

Analysis

To improve the generality of mobile robot modeling,

ordinary tires with sliding steering are used in this

modeling. As shown in Figure 1, taking wheel A as

an example: it can be decomposed into

and

(that is, lateral velocity and longitudinal velocity),

where the lateral velocity is generated by the sliding

friction between the tire and the ground, and the

longitudinal velocity is generated by the rolling

friction between the tire and the ground. Therefore,

the turning motion is generated by sliding friction. It

can be inferred that the rolling friction is actively

generated by the motor output torque driving the

wheel rotation, while the sliding friction is passively

generated due to the inconsistent speeds of the four

wheels.

In order to simplify the model, the analysis is put

forward in the ideal state, that is: (1) no idling

phenomenon occurs when the robot wheel rolls; (2)

The mass of the robot body is evenly distributed and

located on the geometric longitudinal symmetry line

of the robot, but not necessarily on the geometric

transverse symmetry line. As can be seen from

Figure 1, the combined velocity direction of the

ideal contact point between the tire and the ground

(that is, the center of the wheels A, B, C, D) is

perpendicular to the radial direction of the rotation

radius, and can be decomposed into the longitudinal

component speed along the wheel rolling direction

and the transverse component speed along the motor

axis.

As can be seen from Figure 1, the two front

wheels are subjected to lateral forces to the left,

while the two rear wheels are subjected to lateral

forces to the right. The two sets of lateral forces are

opposite in direction, which just makes the robot

rotate around the center of mass. When forward

motion is superimposed, the robot appears to move

in a circular motion.

Figure 1: Motion model analysis of four-wheel drive

mobile robot.

From the longitudinal velocity analysis of the left

wheel, we can see:

11 1

22 2

cos

νν α













(1)

ANIT 2023 - The International Seminar on Artiﬁcial Intelligence, Networking and Information Technology

Formula ,

and

are respectively the linear

velocities of wheel A and wheel B;

and

are ∠ AOP and ∠ BOP

respectively, and P is the geometric center of mass.

From formula 1,

12xx

νν

that is, the

longitudinal component velocity of wheel A and

wheel B is the same, that is, the component velocity

in the X direction is the same. By the same token,

the longitudinal velocities of wheels C and D are the

same.

From the transverse velocity analysis of the front

wheel, it can be seen that:

11 1

44 4

sin

νν α













(2)

Formula ,

and

4 y

are respectively the linear

velocities of wheel A and wheel D;

and

are ∠AOP and ∠DOF respectively.

According to equation 2,

14yy

νν

, that is, the

longitudinal component velocity of wheel A and

wheel D is the same, that is, the component velocity

in the Y direction is the same. By the same logic, the

longitudinal velocities of wheels B and C are the

same.

To sum up:

xxL

xxR

yyF

yyB

νν ν

ννν

νν ν











(3)

Formula ,

and

represents the longitudinal

component speed of the left and right wheels

respectively,

and

represents the transverse

component speed of the front and rear wheels

respectively.

2.3 Mobile Robot Motion Model

Building

The motion model of a four-wheel drive mobile

robot is quite complex, so for ease of calculation, the

four-wheel motion is simplified and treated as an

equivalent two-wheel differential drive motion

model. As shown in Figure 2, a virtual two-wheel

simulation of four-wheel motion was established,

with OP as the horizontal axis and PQ as the vertical

axis. It was assumed that the virtual left and right

wheels were located outside the point vehicle and

close to the vehicle body respectively. The length of

virtual wheel spacing D2 was not necessarily equal

to that of real wheel spacing D1, and virtual wheel

spacing D2 was dynamically changing.

Figure 2: Motion model establishment of four-wheel drive

mobile robot.

From this analysis: if there is no rotational motion

(no slip), only

and

，

both

and

are 0,

then D1=D2; However, if there is rotational motion,

i.e. the velocities are not zero and there is a slip

phenomenon, the actual situation changes. This

means that the angular velocity calculated based on

the

and

and D1 is not the true angular

velocity. The degree of slip varies with

different rotational motions, which affects the actual

angular velocity differently, so the virtual wheel

base D2 is dynamically changing.

The simplified motion model can be expressed as:

Forward kinematics model:

222

DDD

νν

 

 

 



 

−





−

 



(4)

Inverse kinematics model:

Modeling and Simulation of Mobile Robot Based on VREP

ων





 



 

 



−−





(5)

VREP SIMULATION

EXPERIMENT DESIGN AND

RESULT ANAYLYSIS

A simulation experiment was conducted using

VREP software on the established mobile robot

model. The overall design scheme mainly

includes software design and experimental design.

The software design includes four aspects:

modeling, path planning, adding sensors, and

programming using Lua language. The specific steps

are as follows:

① Establish the body of the mobile robot. Add a

rectangular body with a length of 30cm, a width of

20cm and a height of 10cm into ADD, then add four

cylinders to the body as wheels, and add four

rotating joints to the four wheels, so that the cylinder

can drive the body to move.

② Build a simulation environment and plan the

path of the mobile robot. Click Add path, select the

circular path in the category, and then click Track

Edit to edit the 16 track points. The trajectory should

be as smooth as possible to ensure that the robot

does not encounter positional deadlocks during the

simulation.

③ Add sensors. Select Perspective sensor in the

sensor type to change the sensor position

coordinates. The same method adds two more

sensors and renames them left, center, and right.

Make sure the trajectory is within the detection

range of the sensors.

④ Write the motion program. Use the forward

and inverse kinematics equations of the robot in

Chapter 2 to model the programming. The

experimental design includes the correct connection

between each part of the experiment, reasonable

path curves, correct programming, and good

programming habits to achieve the final simulation

result. Figure 3 shows the simulation of the mobile

robot tracking the path.

Figure 3: Tracking simulation diagram of mobile robot.

In VREP, robot control programs can be written

using LUA scripts to achieve robot path tracking. By

using the grayscale sensors and the odometer multi-

sensor fusion algorithm, different sensor information

can be fused to obtain more accurate and

comprehensive robot environment perception

information, thus improving the robot's autonomous

navigation and obstacle avoidance capabilities.

As shown in Figure 3, the three pictures in the

lower right corner are the information collected by

the three grayscale sensors, that is, the information

of the three sensors on the left, middle and right.

When the robot deviates to the left, it can refer to the

information from the left sensor to move forward

and make corrections.Similarly, the right side is the

same as the left side; When the robot moves forward

with the middle sensor as the reference, it can

maintain a straight trajectory and reach the target

point. Through simulation experiments, this method

can accurately simulate the motion and sensor data

of mobile robots. Specific experimental design

includes robot path planning, motion control and

sensor data acquisition, etc. Ultimately,the robot can

navigate the experimental destination with high

precision and the controller has stable performance.

The experimental results show that the mobile robot

model can improve the accuracy of multi-sensor

fusion data and ensure the motion accuracy of the

simulated robot. Through the above experimental

simulation, the feasibility and effectiveness of the

proposed modeling method are verified.

CONCLUSION

This article studies the modeling and simulation

technology of mobile robots based on VREP.

Through the analysis of mobile robot modeling

methods and mobile robot motion analysis, a mobile

robot motion model is established. At the same time,

VREP software is used to simulate its modeling, and

multi-sensor fusion is added to achieve target point

tracking function and improve accuracy. The

ANIT 2023 - The International Seminar on Artiﬁcial Intelligence, Networking and Information Technology

research on multi-sensor fusion of mobile robots

based on VREP is an important research direction in

the field of robotics, which can improve the

perception and autonomous control capabilities of

robots and contribute to the development of robotics

technology. For example, in rescue missions,

multiple sensors can be fused to achieve precise

positioning and rescue operations for trapped

individuals, improving rescue efficiency and success

rate. VREP software can further explore the use of

other simulation software and the improvement of

robot modeling methods to meet more complex

application requirements. The research on multi-

sensor fusion of mobile robots can be applied to

fields such as robot autonomous navigation,

environmental monitoring, and rescue missions.

ACKNOWLEDGMENTS

This work is greatly appreciated for the strong

support from the Engineering Research Center of

Robot and Intelligent Manufacturing, Universities of

Shaanxi Province.

REFERENCES

Rohmer E, Singh S P, Freese M. VREP: a versatile and

scalable robot simulation framework[J]. Simulation,

2013, 89(3): 303-332. https://doi.org/10.1109/IROS.

2013.6696520

Pinrath N., Matsuhira N . Real-time Simulation System for

Teleoperated Mobile Robots using V-REP[C]//AIR

2019: Advances in Robotics 2019.2019.

https://doi.org/10.1145/3352593.3352598

Zhu Xiangbin, Cao Zuoliang, Feng Yubo. Kinematic

Modeling and Motion Control Simulation of Mobile

Robots[J]. Journal of Tianjin University of Technology,

2005(01):54-57.

Wang Yibo, Lv Wentao. A TRS Robot Automatic Control

Method Based on VREP Platform. [J] Science and

Technology Economics Guide. 2021.33-34.

Shamshiri R R. Simulation of Sweet Pepper Robotic

Harvesting in V-REP and ROS[J]. 2015. https://

doi.org/10.13140/RG.2.2.21273.62562

Ciszewski M , Mitka U , Buratowski T ,et al.Modeling and

simulation of a tracked mobile inspection robot in

MATLAB and V-REP software[J]. 2017(2). https://

doi.org/10.14313/JAMRIS_2-2017/11

Fabro J , Conter F , Oliveira A S .Integrating a Mobile

Robot Navigation Control, Visual Object Detection and

Manipulation Using ROS and V-REP[C]//Congresso

Brasileiro de Inteligência Computacional.2020.

https://doi.org/10.21528/CBIC2019-134

Ma Hongwei, Zhang Zhenzhen, Yang Lin, et al. Global

Positioning System and Odometry Fusion Localization

Method for Inspection Robot[J]. Science Technology

and Engineering, 2020, 20(23):9440-9444.

Modeling and Simulation of Mobile Robot Based on VREP