2017 International Conference on Advanced Systems and Electric Technologies (IC_ASET)

Autonomous Wheelchair Navigation in Indoor

Environment Based on Fuzzy Logic Controller and
Intermediate Targets

Khaoula Maatoug Malek NJAH Mohamed JALLOULI

Research Laboratory Modeling, Analysis University of Sfax University of Sfax
and Control of Systems Engineering School of sfax Engineering School of sfax
National Engineering School of Gabès Email: njahmalek@yahoo.fr Email: mohamed.jallouli@enis.rnu.tn
Email: khaoulamaatoug@gmail.com

Abstract—The handicapped person needs more autonomy and mathematical tool that can be used when the system is difficult
flexibility in the home or in a hospital. The electric wheelchair to model and human expert knowledge is available. The fuzzy
can assist those with walking disabilities in getting around. In this logic controller is a useful tool when we want to emulate the
paper our purpose is to have an autonomous navigation for the human reasoning and to be closer to human behaviour. Two
electric wheelchair in the indoor environment. A method based on types of fuzzy logic inference systems exist: Mamdani type
the use of an intermediate targets before reaching the final target
and Sugeno type. To develop the fuzzy logic controller Zero-
is proposed. The used framework is considered as a structure of
a disabled person home. The kinematic model of the robot is order Sugeno model is used. Note that the Sugeno output
determined. Moreover, a fuzzy logic controller is developed. The membership functions are either linear or constant. For the
proposed method is successfully tested in simulations. zero-order Sugeno model, the output level is a constant.
In fact, to move from a room to another one, it is necessary
Keywords—Autonomous Navigation, Electric wheelchair, Fuzzy to take into consideration the positions of each door and the
Logic Controller, Intermediate target, Ultrasonic Sensor. obstacles in the environment. For that, due to complex or long
path between the robot and the final target, the use of an
I. I NTRODUCTION intermediate target is proposed to facilitate navigation towards
the goal. To prevent the robot from getting trapped in front of
The objective of this work is to have an autonomous nav- an obstacle or wandering indefinitely, the intermediate target
igation for the electric wheelchair in the indoor environment. is used as a solution for these cases in this work.
The goal of the wheelchair is to provide an assistance to the The rest of the paper is organized as follows: in section 2 an
user in indoor and outdoor environments and also to drive analysis of the kinematic model of the mobile robot is given.
it with more easily and efficiently. Since the wheelchair is Section 3 discusses the systems navigation behaviors and the
a mobile robot, our research is focused on the study of the fuzzy logic controller. Simulations and results are presented in
autonomous navigation of a mobile robot. The autonomous section 4. Finally section 5 summarizes the paper.
navigation of a mobile robot in unknown environments is one
of the most research area [1], [2], [3]. Nowadays the mobile II. T HE KINEMATIC MODEL
robot can be used in many applications such as indoor and
outdoor also they have been used in various fields such as In order to analyze the system and to design the controller,
space field, domestic field, hospital field and defense security. the mobile robot platform is studied. The employed wheelchair
For autonomous navigation the robot should be equipped is kinematically equivalent to a unicycle mobile robot type.
with perception system which is essentially consisted with Then, our mobile robot model is a unicycle type which
sensors that provide the robot with useful information of the equipped of two free rotating wheels and two independent
environment through visual images [4] or distance [5]. The driving wheels. The free rotating wheels ensures the static
most sensors used are distance sensors (ultrasonic, laser, etc). stability of the vehicle. By acting on the speed of each
They detect an obstacle and measure the distance to walls close wheel the independent driving wheels can be oriented and
to the robot path. By exploiting the sensors information the commended.
robot is capable to move through corridors, to follow walls, to Figure 1 shows the used variables in the kinematic model.
turn corners within indoor environments. An ultrasonic sensors The configuration of the mobile robot is characterized by the
is used in this work. position (𝑋, 𝑌 ) and the orientation 𝜃 in a cartesian coordinate.
For autonomous mobile robot many control methods have
been proposed to deal with the control problem. In literature,
several approaches have been developed using the artificial
intelligence techniques such as neural network [6], [7], [8],
fuzzy logic [9], [10], neuro-fuzzy [11], [12] and other methods.
Among these methods, the fuzzy logic controller is applied in
this work due to its simplicity. The fuzzy logic is a powerful
978-1-5090-6634-6/17/$31.00 ©2017 IEEE 55

Authorized licensed use limited to: Universidad Nacional de Colombia (UNAL). Downloaded on November 26,2020 at 09:09:09 UTC from IEEE Xplore. Restrictions apply.
 2017 International Conference on Advanced Systems and Electric Technologies (IC_ASET)

III. NAVIGATION A LGORITHM

The aim is to develop an efficient method that allows
controlling the wheelchair to reach its goal while avoiding
unknown obstacles on its way. The developed approach is
based on the fuzzy logic controller and on the intermediate
target concept which proposed to facilitate the navigation.
Figure. 2 shows the robot navigation environment which
supposed a planar surface. Five intermediate target are created
for each door. They are represented by an ellipse in figure 2.

20

Fig. 1. Scheme of the mobile robot. 15

10
Intermediate target
Vd,Vg : are respectively the linear speeds of the wheels
left and right.
5
L: is the distance between the two driving wheels.
𝜃 : is the angle of the orientation of the robot.

Y(m)
V : is the linear speed of the robot. 0

(X,Y) is the position of the robot.

A simplifying hypothesis considered for the modulation are : −5

∙ The wheel ground contact is a simple point. −10

∙ The rolling motion of the robot wheel without slip-

ping. −15

∙ The system evolves on horizontal ground. −20

−20 −15 −10 −5 0 5 10 15 20
X(m)
∙ The configuration of the mobile robot is described
with the three coordinates (𝑥, 𝑦, 𝜃).
Fig. 2. The navigation environment with the positions of the intermediate
The kinematic model [13] is given by these equations : targets.

𝑑𝑋 𝑉 𝐿 + 𝑉𝑅
= cos 𝜃 (1) The ability to sense the surrounding environment is a
𝑑𝑡 2
fundamental requirement to any autonomous system. The
appropriate sensors, able to specify distance to the walls, to
𝑑𝑌 𝑉 𝐿 + 𝑉𝑅 the obstacles and to the target, are mounted on the wheelchair.
= sin 𝜃 (2)
𝑑𝑡 2 Three steps are considered to create a fuzzy controlled (for
more details see [14]) :
𝑑𝜃 𝑉𝑅 − 𝑉𝐿
= (3) ∙ Fuzzification : transform each real value inputs and
𝑑𝑡 2 outputs into grades of membership for linguistic terms
The discret form of the kinematic model is : of fuzzy sets.

𝑉𝑅𝑘 + 𝑉𝐿𝑘 ∙ Rule evaluation : combine the facts obtained from the
𝑋𝑘+1 = 𝑋𝑘 + 𝑇 cos 𝜃𝑘 (4) fuzzification with the rule base and conduct the fuzzy
2 reasoning process.
𝑉𝑅𝑘 + 𝑉𝐿𝑘 ∙ Defuzzification : transform the subsets of the outputs
𝑌𝑘+1 = 𝑌𝑘 + 𝑇 sin 𝜃𝑘 (5) to obtain the actual results.
2
The navigation behavior consists of an avoidance behavior and
𝑉𝑅𝑘 − 𝑉𝐿𝑘 goal-seeking behavior.
𝜃𝑘+1 = 𝜃𝑘 + 𝑇 (6)
𝐿 Figure. 3 shows a schematic diagram of the navigation behav-
ior.
where T is the sampling time.
To build a model of the robot, the discret form is used. These
equations were used to simulate the robot in MATLAB. The
fuzzy logic controller has exploited the information given by
the discrete form.
56

Authorized licensed use limited to: Universidad Nacional de Colombia (UNAL). Downloaded on November 26,2020 at 09:09:09 UTC from IEEE Xplore. Restrictions apply.
 2017 International Conference on Advanced Systems and Electric Technologies (IC_ASET)

actual trajectory to avoid the obstacle. Therefore, the objective

of the fuzzy control system is to enable the robot to navigate
from a start position to a goal position without collisions.
Figure. 5 illustrates the block diagram of the robotic system.

Fig. 3. Navigation behavior of the mobile robot.

Fig. 5. Block diagram of the robotic system

If there is any obstacle in the way, the controller has two
inputs which are the distance (d) between the robot and the Before the robot moves ahead, the position of the desired
goal and the angle( 𝜑) between robot’s orientation and target’s target is checked. Afterward, an appropriate intermediate target
one. is chosen.
The following equations are the expression of the two inputs:
√ IV. S IMULATIONS AND R ESULTS
2 2
𝑑 = (𝑋𝑇 − 𝑋) + (𝑌𝑇 − 𝑌 ) (7)
The main goal for the robot is to be able to navigate
𝜑 = 𝜃𝑇 − 𝜃 (8) autonomously from a known start point, to a known goal
point, being aware of its position at all times. In simulation,
with : the starting position of the robot, the position of the target
𝑌𝑇 − 𝑌
𝜃𝑇 = tan−1 ( ) (9) and the position of the intermediate targets are given for each
𝑋𝑇 − 𝑋 navigation task. Figure. 6 and Figure. 7 illustrate a navigation
where 𝑋𝑇 and 𝑌𝑇 are the coordinate of the desired position. problem in the indoor environment. The execution time and
If the robot encounters an obstacle, the inputs of the proposed the trajectory are too long. Therefor, the mobile robot has a
fuzzy controller are the distances measured and the angle problem to reach the target in efficiently way.
between the robot and the goal.
The mobile robot should be equipped with sensor system to
detect obstacles. In our application, an ultrasonic sensors is
selected. The ultrasonic gives information that can be used to
calculate the distance between the robot and the obstacles. The
navigation control process start with the obstacle detection that
may collide the robot. This is accomplished by a set of three
ultrasonic sensors mounted on front of the robot. The sensors
are located at left, right, and front sides of the robot.
The distances obtained from these equations can be exploited
by the fuzzy controller to avoid obstacles.

Fig. 4. The distances measured by the ultrasonic sensor. Fig. 6. Trajectory of the robot without the use of the intermediate targets.

Hence, the inputs of the controller measured by the ultra-

sonic sensors are the left distance (LD), the right distance (RD)
and the front distance (FD) see Figure 4. The output variables
are the speed of the left and the right wheels.
The measured distances and angle is received by the fuzzy
controller. Based on these data, the controller can change the
57

Authorized licensed use limited to: Universidad Nacional de Colombia (UNAL). Downloaded on November 26,2020 at 09:09:09 UTC from IEEE Xplore. Restrictions apply.
 2017 International Conference on Advanced Systems and Electric Technologies (IC_ASET)

Fig. 7. Trajectory of the robot without the use of the intermediate targets Fig. 9. Trajectory of the robot with the use of the intermediate targets

Instead of moving ahead immediately, the robot choose

the appropriate intermediate target to allow the robot move
and to achieve the safe direction. After the intermediate target
have been chosen, the robot navigates towards the intermediate
target. Then, the intermediate target is considered as the start
position of the robot in the algorithm and the robot can move
towards the final target. Therefore, the wheelchair is able to
solve this problem successfully as shown in Figure. 8 and
Figure. 9. The robot moves directly towards the target along
the shortest path.

Fig. 10. Trajectory of the robot with the use of the intermediate targets and
more obstacle

To prove the effectiveness of the fuzzy controller in the

presence of more obstacle in our plan, an obstacles can be
added. Once the ultrasonic sensors detect any obstacle, the
fuzzy inference algorithm is activated to avoid it. As shown in
Figure. 10 and Figure. 11 the robot successfully navigate to
the desired target with autonomy and safety way.

Fig. 8. Trajectory of the robot with the use of the intermediate targets

58

Authorized licensed use limited to: Universidad Nacional de Colombia (UNAL). Downloaded on November 26,2020 at 09:09:09 UTC from IEEE Xplore. Restrictions apply.
 2017 International Conference on Advanced Systems and Electric Technologies (IC_ASET)

ACKNOWLEDGMENT
This work was supported by the Ministry of the Higher
Education and Scientific Research in Tunisia.

R EFERENCES
[1] R. Zhao, D. H. Lee, T. T. Li, and H. K. Lee, “Autonomous navigation
of a mobile robot in unknown environment based on fuzzy inference,”
in Automatic Control Conference (CACS), 2015 International. IEEE,
2015, pp. 19–24.
[2] M. Hamani and A. Hassam, “Mobile robot navigation in unknown
environment using improved apf method,” in the 13th international
conference in unknown environment using improved APF method, 2012,
pp. 0875–1812.
[3] A. Jazayeri, A. Fatehi, and H. Taghirad, “Mobile robot navigation in
an unknown environment,” in 9th IEEE International Workshop on
Advanced Motion Control, 2006. IEEE, 2006, pp. 295–299.
[4] M. Cao and E. L. Hall, “Fuzzy logic control for an automated guided
vehicle,” in Photonics East (ISAM, VVDC, IEMB). International
Society for Optics and Photonics, 1998, pp. 303–312.
[5] H. Yu and R. Malik, “Aimy: An autonomous mobile robot navigation
in unknown environment with infrared detector system,” Journal of
Intelligent and Robotic Systems, vol. 14, no. 2, pp. 181–197, 1995.
Fig. 11. Trajectory of the robot with the use of the intermediate targets and [6] R. Carelli, E. F. Camacho, and D. Patino, “A neural network based
more obstacle feedforward adaptive controller for robots,” IEEE transactions on
systems, man, and cybernetics, vol. 25, no. 9, pp. 1281–1288, 1995.
[7] A. Pandey, A. Pandey, D. R. Parhi, and D. R. Parhi, “New algorithm for
behaviour-based mobile robot navigation in cluttered environment using
neural network architecture,” World Journal of Engineering, vol. 13,
no. 2, pp. 129–141, 2016.
[8] M. Al-Sagban and R. Dhaouadi, “Neural based autonomous navigation
of wheeled mobile robots.” Journal of Automation, Mobile Robotics &
Intelligent Systems, vol. 10, no. 2, 2016.
[9] H. Surmann, J. Huser, and L. Peters, “A fuzzy system for indoor
mobile robot navigation,” in Fuzzy Systems, 1995. International Joint
Conference of the Fourth IEEE International Conference on Fuzzy
Systems and The Second International Fuzzy Engineering Symposium.,
Proceedings of 1995 IEEE Int, vol. 1. IEEE, 1995, pp. 83–88.
[10] M. M. Almasri, K. M. Elleithy, and A. M. Alajlan, “Development of
efficient obstacle avoidance and line following mobile robot with the in-
tegration of fuzzy logic system in static and dynamic environments,” in
Long Island Systems, Applications and Technology Conference (LISAT),
2016 IEEE. IEEE, 2016, pp. 1–6.
[11] B. Kosko, Neural networks and fuzzy systems: a dynamical systems
approach to machine intelligence/book and disk. Prentice Hall, Upper
Saddle River, 1992.
[12] P. Rusu, E. M. Petriu, T. E. Whalen, A. Cornell, and H. J. Spoelder,
“Behavior-based neuro-fuzzy controller for mobile robot navigation,”
IEEE Transactions on Instrumentation and Measurement, vol. 52, no. 4,
pp. 1335–1340, 2003.
Fig. 12. Trajectory of the robot with the use of the intermediate targets and [13] M. I. Ribeiro and P. Lima, “Kinematics models of mobile robots,”
more obstacle Instituto de Sistemas e Robotica, pp. 1000–1049, 2002.
[14] K. Maatoug, M. Njah, and M. Jallouli, “Autonomous navigation of
The results of the simulation show the success of the unicycle mobile robot based on fuzzy controller,” NNGT International
fuzzy logic controller for autonomous navigation wheelchair Journal on Artificial Intelligence, vol. 2, pp. 47–56.
in indoor environment. Fuzzy logic approach is considered as
a simple and a powerful technique for control problems.

V. C ONCLUSION
This paper study the autonomous navigation of an electric
wheelchair in indoor environment. The proposed algorithm
gives the shortest path and the control strategy that lead the
robot to the target. The kinematic model of the robot system
is determined. Moreover, the navigation fuzzy controller and
the concept of intermediate target are introduced. Finally, the
simulation results show that the proposed method can achieve
a successful navigation with short execution time and path.
59

Authorized licensed use limited to: Universidad Nacional de Colombia (UNAL). Downloaded on November 26,2020 at 09:09:09 UTC from IEEE Xplore. Restrictions apply.