2009 Second International Symposium on Knowledge Acquisition and Modeling

Application of an Embedded System for the Car-Like Mobile Robot

Chao-Lin Kuo1, Yong-Ren Pu2, Shen-Min Liang3, Chia-Hung Lin4

1 Department of Electrical Engineering, Far East University, Tainan, Taiwan, R.O.C.
2 Department of Occupational Safety and Health, Chang Jung Christian University, Tainan, Taiwan, R.O.C.
3 Department of Computer Application Engineering, Far East University, Tainan, Taiwan, R.O.C.
4 Department of Electrical Engineering, Kao-Yuan University, Kaohsiung, Taiwan, R.O.C.
1
[email protected], [email protected], [email protected], [email protected]

Abstract—In this paper, a fuzzy controller is implemented by where ( xr , y r ) denote position of the rear wheel center of
an embedded system and applies to the wall-following
the car-like mobile robot, θ is angle between the vehicle
control of a car-like mobile robot. First, we introduce a
model of the car-like mobile robot. Next, we address the
frame orientation and the X-axis. The front wheel center of
design scheme of fuzzy control. According to the expert, the the car-like mobile robot is fixed relative to the rear wheel
fuzzy rules are constructed. Finally, the feasibility and center, thus the coordinate ( xr , y r ) is related to the position
robustness of the proposed controller are demonstrated by of the front wheel center ( x f , y f ) of the car-like mobile
experimental results.
robot
Keywords- embedded system; fuzzy control; car-like mobile
x r = x f − l cos θ
robot (2)
y r = y f − l sin θ .
1. INTRODUCTION
where l is the distance between the two wheel-axle
In the past few decades, the control of mobile robots centers. We can calculate derivatives of the velocity
has been studied [1-5]. Mobile robots offered widespread projections with respect to time.
applications, such as guidance tasks, Mars exploration,
material transportation, military missions, and security x r = x f + ω l sin θ
systems. Many functions of mobile robots have been (3)
discussed, such as parking control [1], trajectory-tracking y r = y f − ω l cosθ .
[2, 3], navigating [4], and obstacle detection and
avoidance. where ω = dθ / dt is angular velocity. Substituting (3) to
The fuzzy set theory was initiated by Zadeh [6]. The (1) and from Figure 1, we can get
fuzzy logic control schemes have been successfully x f = v cos(θ + φ )
applied to many fields [1, 3, 7, 8]. Li [3] has presented an (4)
adaptive sliding-mode dynamic control for the trajectory- y f = v sin(θ + φ ).
tracking control of nonholonomic wheeled mobile robots.
Kuo [7] has presented an adaptive fuzzy sliding mode where v is the driving speed and φ is relative steering
control for Sprott’s chaotic system synchronization. In his angle. Using (4), (3) and (1) one can obtain
paper, the wall following problem was addressed for a
care-like mobile robot. A design methodology which sin φ
ω=v . (5)
embedded system fuzzy controller was presented. l
This paper is organized as follows. In section 2, the
Summarizing (4) and (5), we have the dynamics of the axle
dynamic model of a car-like mobile robot is introduced.
center of the front wheels as follows
The hardware of the car-like mobile robot is briefly
described, in section 3. Section 4 describes the design of x f = v cos(θ + φ )
the fuzzy controller. Experimental results of the proposed
controller in a real time are shown in section 5. Finally, y f = v sin(θ + φ ) (6)
conclusions are given in section 6. sin φ
θ = ω = v .
2. MODEL OF A CAR-LIKE MOBILE ROBOT l
The system that we call a car-like mobile robot is a Substituting (6) to (3), we can get the dynamics of the
simplified model of the constraints imposed on the axle center of the rear wheels as follows
movement of a real car as shown in Figure 1 [1-3]. The
front wheels can turn to the left or right, but the left and xr = v cos θ cos φ
right front wheels must be parallel. The rear wheels are y r = v sin θ cos φ (7)
fixed to be parallel to the car body and allowed to roll or
sin φ
spin but not slip. The equations of motion for the car-like θ = ω = v .
mobile robot can be expressed by l

y r cosθ − x r sin θ = 0. (1)

978-0-7695-3888-4/09 $26.00
$25.00 © 2009 IEEE 140
136
DOI 10.1109/KAM.2009.70
Let u1 = v cos φ and u 2 = φ . The control system can be e1 = Dr − y f
(10)
rewritten as follows e2 = Dr − y r .
ª x r º ª cos θ º ª0 º The steering angel φc of the car-like mobile robot is
« y » « sin θ » « »
« r» = « » u + «0 » u (8)
chosen as
« θ » «(tan φ ) / l » 1
«0 » 2 φc = FC (e1 , e2 ) , (11)
« » « » « »
¬φ ¼ ¬ 0 ¼ ¬1 ¼
where FC denotes the functional characteristics of the
where u1 corresponds to the translational velocity of the fuzzy linguistic decision schemes. The controller inputs
rear wheels and u 2 the angular velocity of the steering are the error position of the front wheel center and rear
wheel center of the car-like mobile robot with the designed
wheels. position. The output of the controller is a relative steering
angle. The membership functions of input linguistic
3. HARDWARE CONFIGURATION OF THE CAR-LIKE
MOBILE ROBOT variables e1 and e2 , and the membership function of
output linguistic variable φc are shown in Figure 6. They
In this study, the hardware architecture of the car-like
mobile robot is shown in Figure 2. A car-like mobile robot are decomposed into seven fuzzy partitions expressed as
consists of servo motor, two infrared sensors, the batteries, negative big (NB), negative medium (NM), negative small
one DC motor, embedded system, and so on. Figure 3 (NS), zero (ZE), positive small (PS), positive medium
illustrates the block diagram of the car-like mobile robot. (PM) and positive big (PB). The fuzzy rule table for wall-
The power supply provides the energy for the control following control of the car-like mobile robot is designed
system servo motor, and DC motor. There are two infrared as listed in Table I. For example:
sensors that were mounted on the car-like mobile robot and IF e1 is negative small and e2 is positive medium,
used to detect the distant of car to the wall. The angle
between the vehicle frame orientation and the X-axis can THEN φc is positive medium.
be controlled by the servo motor. The velocity of the car-
like mobile robot is fixing. The embedded system The defuzzification of the control output is achieved by the
processes information of infrared sensor signal and adjusts weighted average method.
the direction of the car-like mobile robot according to the TABLE I. ROLE TABLE OF WALL FOLLOWING.
fuzzy control.
e1
4. FUZZY SLIDING MODE CONTROL FOR THE CAR-LIKE φc
NB NM NS ZE PS PM PB
MOBILE ROBOT
PB PB PB PM PS ZE NS NM
The object is to design a fuzzy controller for the car-
PM PB PB PM PS ZE NS NM
like mobile robot with the wall-following control that
keeps at a constant distance from staring at any distance PS PB PM PS PS NS NM NB

with wall. The dynamic behavior of fuzzy logic control is e2 ZE PB PM PS ZE NS NM NB

characterized by a set of linguistic rules based on expert NS PB PM PS NS NS NM NB
knowledge. The inference mechanism of fuzzy logic NM PB PM PS NS NM NB NB
control will be able to provide appropriate fuzzy control NB PM PS ZE NS NM NB NB
action. The fuzzy control scheme for the car-like mobile
robot is shown in Figure 4. 5. EXPERIMENTAL RESULTS
The variables of the car-like mobile robot are shown as The fuzzy controller has implemented by an embedded
Figure 5, where β denotes the angle of the wall with system and applied for the car-like mobile robot. This
respect to the car-like mobile robot, α the angle measured study used a look-up table technique to implement the
with respect to the Y-axis, Dr the desired distance fuzzy controller in the embedded system. The actual
between the care-like mobile robot and wall, Ds1 the appearance of the car-like mobile robot is shown in Figure
6. The objective of this study is the wall following control;
distance measured by a front sensor, and Ds 2 the distance the distance from the sensor of the car-like mobile robot to
measured by a rear sensor. In this study, angle α equals to the wall is 10 cm. For experimental part, the car-like
angle β . We can calculate the positions of y f and y r mobile robot was set at any position with the wall, and the
fuzzy controller was turn on. Experimental results are
from the front and rear sensors. These positions, y f and illustrated in Figure 8, where the car-like mobile robot can
yr be in wall following when staring at any distance with
, can be described as
wall.
y f = ( Ds1 + w) cos α − w1 sin α
(9) 6. CONCLUSIONS
y r = ( Ds 2 + w) cos α − w2 sin α .
In this paper, a fuzzy controller is proposed and
The error dynamic equations can be written as implemented by an embedded system for the car-like
mobile robot. The fuzzy control rules are determined by
systematic analysis and design. We have applied the fuzzy

137
141
rules to regulate the steering angle and to drive the car-like [3] C.-Y.g Chen, T.-H. S. Li, and Y.-C. Yeh, “EP-based kinematic
mobile robot followed to a trajectory. The practical control and adaptive fuzzy sliding-mode dynamic control for
wheeled mobile robots,” Information Sciences, vol. 197, pp. 180-
experimental results have demonstrated that the proposed 195, 2009.
fuzzy controller is correct and valid for the car-like mobile [4] S.-H. Kim, C.-W. Roh, S.-C. Kang, and M.-Y. Park, “Outdoor
robot. navigation of a mobile robot using differential GPS and curb
detection,” Proc. 2007 IEEE International Conference on Robotics
ACKNOWLEDGMENT and Automation, 10-14 April 2007, pp. 3414-3419, 2007.
This work was supported by the National Science [5] Y. Chen and S, Yasunobu, “Soft target based obstacle avoidance
for car-like mobile robot in dynamic environment,” Proc. IEEE
Council of R.O.C. under project NSC 98-2221-E-269 - International FUZZ-IEEE 2007 Fuzzy Systems Conference, 23-26
008. July 2007, pp. 1-6, 2007
[6] L. A. Zadeh, “Fuzzy logic,” IEEE Computer, vol. 21, pp. 83-93,
REFERENCES 1998.
[1] T.-H. S. Li and S.-J. Chang, “Autonomous fuzzy parking control of [7] C.-L. Kuo, “Design of an adaptive fuzzy sliding-mode controller
a car-like mobile robot,” IEEE Trans. Systems, Men, Cybernetics- for chaos synchronization,” International Journal of Nonlinear
Part A, vol. 55, no. 4, pp. 451-465, 2003. Sciences and Numerical Simulation, vol. 8, pp. 631-636, 2007.
[2] J.-.l Laumond, P. E. Jacobs, T. Michel, and R. M. Murray, “A [8] H.-T. Yau and C.-S. Shieh, “Chaos synchronization using fuzzy
motion planner for nonholonomic mobile robots,” IEEE Trans. on logic controller,” Nonlinear Analysis B: Real World Applications,
Robotics and Automation, vol. 10, no. 5, pp. 577-593, 1994. vol. 9, pp.1800-1810, 2008.

Figure 3. Block diagram of the car-like mobile robot

Figure 1. A car-like mobile robot.

Figure 4. The block diagram of fuzzy control for the car-like mobile
robot.

Figure 2. The drawing of the car-like mobile robot

Figure 5. The variables of the car-like mobile robot.

138
142
 (1) (5)

Figure 6. Membership function of the input-output variables of fuzzy

controller.
(2) (6)

(3) (7)

(4) (8)
Figure 7. Actual appearance of the car-like mobile robot.
Figure 8. Realized tracking control

139
143