Available online at www.sciencedirect.

com

ScienceDirect
Procedia Technology 24 (2016) 538 – 545

International Conference on Emerging Trends in Engineering, Science and Technology

(ICETEST - 2015)

Modelling and trajectory tracking of wheeled mobile robots

N.Leenaa, K.K.Sajub
a
Federal Institute of Science and Technology, India
b
Cochin University of Science and Technology, India

Abstract

Differential drive mobile robots are widely used due to their simplicity, easiness of control and flexibility. This paper discuss
a detailed modeling of a differential drive robot taking into account the kinematics, actuator dynamics and rolling resistances
of the wheels. Controllers have been designed for smooth trajectory tracking. Different trajectories similar to real life
scenarios have been created and the model and control algorithm are seen to give accurate trajectory tracking.
©©2016
2016TheTheAuthors.
Authors. Published
Published by Elsevier
by Elsevier Ltd. Ltd.
This is an open access article under the CC BY-NC-ND license
Peer-review under responsibility of the organizing committee of ICETEST – 2015.
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the organizing committee of ICETEST – 2015
Keywords: Differential Drive mobile robot; Kinematic controller; actuator dynamics; trajectory tracking

1. Introduction

Autonomous mobile robots are finding widespread application is many areas like mining, space exploration
and in service industry. The differential drive mobile robot (DDMR) is one such robot that has gained wide
popularity due to its simplicity and ease of control [2]. The differential drive robot consists of two individually
propelled wheels and a third wheel called castor wheel that can move freely in space. By adjusting the power
applied to motors, the robot can be operated to go forward, rotate in place or perform movement on any arbitrary
curve in plane.
Several research work on the modelling and control of such robots [1, 2, 3]. However most of these works
handle the two aspects separately. Detailed modeling of the wheeled robot taking into account the actuator
dynamics, rolling resistances and coupling constraints has not been attempted in most works when designing
control schemes.
This paper discusses a detailed kinematic and dynamic model for the mobile robot which includes the chassis
dynamics and actuator dynamics. Trajectory tracking is implemented using two controllers. A kinematic
controller has been used for outer loop control to generate the reference velocities whereas a proportional

2212-0173 © 2016 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the organizing committee of ICETEST – 2015
doi:10.1016/j.protcy.2016.05.094
 N. Leena and K.K. Saju / Procedia Technology 24 (2016) 538 – 545 539

controller has been used as inner loop controller to generate angular velocities for the wheels of the robot.
Finally, several test trajectories were created and the simulation results are presented are seen to show excellent
trajectory tracking.

2. Modeling of differential drive robot

The differential drive robot consists of a platform equipped with a front castor and a pair of rear co-axial drive
wheels for isostatic equilibrium. Each of these drive wheels are independently driven by a DC motor which is in
turn energized by a control voltage. By varying the power applied to the motors the differential wheeled mobile
robot can be made to move in straight line or trace different trajectories like curves, circles etc. Deriving a precise
mathematical model is a crucial part for designing any control system. Kinematic and dynamic model of the robot
has been discussed below.

2.1. Kinematic modelling

Kinematic modeling deals with the geometric relationships that govern the system and studies the
mathematics of motion without considering the affecting forces. For a differential drive which has two wheels
with radius r paced at a distance L/2 from robot center. with θ be the robot orientation angle measured from the
x-axis. From [1] and [2] the following equations are obtained:

r r

Fig. 1. Differential drive robot.

vrZ (1)
dT
w (2)
dt

The speed of each wheel in the robot frame is rZ , therefore the translational speed in the robot frame is the
average velocity

r (Z R  Z L )
vmob
2 (3)
vR  vL
Z
whereas the rotational velocity is given by L . (4)
The mapping between the inertia frame and the robot frame is done through the standard orthogonal
transformation. Hence the robot velocity in the inertial or global frame is given by
540 N. Leena and K.K. Saju / Procedia Technology 24 (2016) 538 – 545

ª.º ª r (Z R  Z L ) º
« x. » ªcosT  sin T 0º « »
« sin T » 2
« y» cosT 0» « 0 »
«.» « « r (Z  Z ) »
«T » «¬ 0 0 1»¼ « R L
»
¬ ¼ ¬« 2 ¼» (5)

2.2. Dynamic modelling

Dynamic modelling of the robot is the study of motion in which forces and energies are modeled and studied.
The actuator modeling is required to find the relationship between the control signal and the mechanical system
input.
2.2.1 Dynamics of the actuator: The actuating machines mostly used in mechatronics motion control systems
are DC machines (motors). The mobile platform motion control can be simplified to a DC motor motion control.
In modelling DC motors and in order to obtain a linear model, the hysteresis and the voltage drop across the
motor brushes are neglected, the motor input voltage, Vin is applied to the field or armature terminals.
The open loop transfer function of the dc motor [2] is given by
kt
Z m (s) n
V ( s) >( L J )s
a
2
 ( Ra J  BLa ) s  ( Ra B  k t k b ) @ (6)

where Ra -resistance and La -inductance of the motor

J –moment of inertia and B –viscous coefficient referred to the motor shaft
kt -torque constant ,kb -emf constant and n-gear ratio.
The left and right wheel motors are modelled using the above equation.
2.2.2 Dynamics of the chassis: Two methods are used for dynamic mode derivation.-the Newton Euler method
and the Lagrange method. The lagrange method has been chosen here due to its more systematic nature and
automatic elimination of workless and constraint forces.
From the lagragian equations of motion [6], the following relations are obtained:
. (TL  TR )
v  kv v (7)
m
. (TL  TR )l
Z  kZ Z (8)
r
Where TL and TR represent left and right motor torques, kv-resistance coefficient against linear motion kω-
resistance coefficient against rotary motion. The combined dynamic model of actuator and chassis is shown in
Fig.2.
 N. Leena and K.K. Saju / Procedia Technology 24 (2016) 538 – 545 541

Fig.2 Dynamic model of actuator and chassis

3. 3. Trajectory tracking of DDMR

Trajectory-tracking problem is posed as follows[1]: It is assumed that the main robot has the posture: P = (x, y,
ɸ)T , and that the reference robot (to be followed) has the posture: Pr = (xr, yr, ɸr)T . The objective is to find
control laws for the linear and angular velocities (v, w) of main robot such as: lim t→∞ |x(t)−xr(t)| = 0, lim t→∞
|y(t)−yr(t)| = 0, lim t→∞ |ɸ (t)−ɸ (t)| = 0.
Fig.3 shows the block diagram for trajectory tracking. For trajectory tracking and control loop has to be
established that will provide as output the required input voltage for the wheels. There are two controllers in the
control scheme. The outer controller is a kinematic controller that takes as input the errors in the x and y
coordinate’s between the reference position and actual position obtained from sensors along with the orientation
of the robot. The kinematic controller gives the reference linear and angular speeds.
xd ex wrref
vref ewr
REFERENCE KINEMATIC -1
TRAJECTORY CONTROLLER M INNER LOOP
GENERATOR yd ey CONTROLLER
wref
wlref ewl
x y

theta
vr
wr wl

KINEMATICS M ACTUATOR + MOBILE DYNAMICS

w

Fig.3 Block diagram for closed loop control

542 N. Leena and K.K. Saju / Procedia Technology 24 (2016) 538 – 545

The kinematic controller (outer loop) used in the study [5] is given by:

ª. kx º
ª v ref º ª cosT sin T º « x d  l x tanh l ( x d  x) »
«  1 * sin T  1 * cosT » « . »
x (9)
«Z » k
¬ ref ¼ «¬ » « y  l tanh y ( y  y )»
a a ¼« d y d »
¬ ly ¼
A simple P controller is used as the inner loop controller.
The following transformation has been used to change from linear and angular velocities to angular velocities
of the wheels.

(10)
Z
ªvº
«Z » >M @ª« r º»
¬ ¼ ¬Z l ¼
ªr / 2 r / 2 º (11)
M « r / l  r / l»
¬ ¼
4. 4. Simulation of trajectory tracking of DDMR

The control scheme for trajectory tracking given in fig.4 was implemented in MATLAB Simulink. Different
behaviors that need to be executed by a mobile robot like moving to a point, tracking a circle, tracking a line etc.
were checked by generating the required trajectories.

Fig.4. Simulation diagram for trajectory tracking

4.1 Simulation Results:

Case 1: Circular trajectory tracking of the robot.

As seen in Fig.5 the robot trajectory follows the reference circular trajectory of radius 1 cm. The robots initial
position is at (0,0).
 N. Leena and K.K. Saju / Procedia Technology 24 (2016) 538 – 545 543

Robot Trajectory
1.5
Reference
Actual

0.5
y [m]

-0.5

-1

-1.5
-1.5 -1 -0.5 0 0.5 1
x [m]

Fig.5 Circular trajectory tracking of mobile robot

Case 2: Straight line trajectory tracking of the robot.

Two straight lines with equation x=3 and one with unity slope are created. In both the cases, the robot moves
from its initial position (0,0) and tracks the given straight line trajectory as shown in Fig.6.a. and Fig.6.b.

Robot Trajectory Robot Trajectory

25 25
Reference
Actual
Reference
20 20 Actual

15 15
y [m]

y [m]

10 10

5 5

0 0

-5 -5
-0.5 0 0.5 1 1.5 2 2.5 3 -5 0 5 10 15 20 25

Fig.6.a. and b. Straight line trajectory tracking of the robot

Case 3: Moving to a point

The goal of the robot was chosen as (2, 3). The robot moves from its initial position (0,0) and reaches the
desired position as in Fig.7.
544 N. Leena and K.K. Saju / Procedia Technology 24 (2016) 538 – 545

Robot Trajectory
3.5
Reference
Actual
3

2.5

2
y [m]

1.5

0.5

-0.5
-0.5 0 0.5 1 1.5 2 2.5
x [m]

Fig.7 Robot moving to a desired position

Fig.s 8(a),(b) and 9 shows the error in x and y coordinates while the robot executed the behaviors: moving to
a point, circular trajectory tracking and straight line trajectory tracking respectively.
Error in x and y [m]
3.5
ex
ey
3

2.5

2
ex , ey (m)

1.5

0.5

-0.5
0 5 10 15 20 25 30 35 40 45 50
time [s]

Fig.8 (a) Error in x and y-moving to a point

Error in x and y [m]

1
ex
ey

0.8

0.6

0.4
ex, ey (m)

0.2

-0.2

-0.4
0 5 10 15 20 25 30 35 40 45 50
time [s]

Fig.8(b) Error in x and y-circular trajectory tracking

 N. Leena and K.K. Saju / Procedia Technology 24 (2016) 538 – 545 545

Error in x and y [m]

0.5
ex
ey
0.45

0.4

0.35

0.3
ex, ey (m)

0.25

0.2

0.15

0.1

0.05

0
0 5 10 15 20 25 30 35 40 45 50
time [s]

Fig.9 Error in x and y – straight line tracking

It was seen that for the same value of controller parameters, different trajectories have an influence on errors
in the x and y co-ordinates.

Summary

A kinematic and dynamic mode has been created for the differential drive mobile robot using
transfer function approach. A kinematic and simple dynamic controller was also done. The model with the
control schemes has been able to satisfactorily track the given trajectory. The control scheme is good enough for
basic tracking problems. More robust control algorithms will be studied in future.

References

[1] Egerstedt, M., 2013. Control of mobile robots.

[2] Ahmad A. Mahfouz, Ayman A. Aly, Farhan A. Salem, 2008, Mechatronics Design of a Mobile Robot System, IJISA, vol.5, no.3, pp.23-
3,2013.
[3] Nwe, A. A., Aung, W. P. and Myint, Y. M. Software implementation of obstacle detection and for wheeled mobile robot avoidance
system. World Academy of Science, Engineering and Technology, 42, 572–577.
[4] J. Craig,1989, Introduction to Robotics: Mechanics and Control,Addison-Wesley.
[5] Martins, F. N., Celeste, W. C., Carelli, R., Sarcinelli-Filho, M. and Bastos-Filho, T. F.,2008, An Adaptive Dynamic Controller for
Autonomous Mobile Robot Trajectory Tracking. Control Engineering Practice, v. 16, p. 1354–1363.
[6] Ali Gholipour, Yazdanpanah.M.T., Dynamic tracking control of Nonholonomic Mobile Robot with Model Reference Adaptation for
Uncertain parameters, European control Conference,2003, pp.3188-3122,2003.