1-4244-1355-9/07/$25.00 @2007 IEEE International Conference on Intelligent and Advanced Systems 2007 244 ~ Visual-Based Fuzzy Navigation System for Mobile Robot: Wall and Corridor Follower Balza Achmad, Mohd Noh Karsiti Department of Electrical and Electronics Engineering, Universiti Teknologi Petronas, Malaysia (balzach@ugm.ac.id, nohka@petronas.com.my) Abstract – This paper presents the development of a visual–based fuzzy navigation system that enables a mobile robot in moving through a corridor or following a wall. The system employs a camera to detect the existence of walls on the left, the right, and the front of the robot. A mamdani-type fuzzy logic controller uses the information gathered by the camera to determine the turning angle and the speed of the robot. The fuzzy system is tested using an OpenGL-based 3D simulator that capable in animating the movement of the robot as well as generating the images captured by the camera. The results of the test confirm that the controller shows a good performance in navigating the robot. Keywords: fuzzy controller, mobile robot, navigation, visual- based system I. INTRODUCTION Nowadays, robots play important roles in industries and manufacturing. Many robots have been built to help in enhancing productivity and capacity along production lines. They appear in various forms and sizes, including mobile robots. Among many different types of robots, mobile robots have unique characteristics that they are capable in moving from place to place. Therefore, they can be utilized to autonomously handle many human daily tasks, particularly for dangerous or monotonous tasks. In order to successfully move from one place to another, mobile robots require certain navigation system that incorporates several features. These features may include self-localization [1]-[3], path planning and tracking [1], [4], [5], corridor following [2], [4], [6], clearance maintenance [7], collision or obstacle avoidance [2], [5], [7], [8], [9], as well as target detection and recognition [8], [10], [11]. Mobile robots make use of various sensors to provide the above features, e.g. ultrasonic and infrared sensors, GPS device [3], as well as camera as visual sensor [1]-[2], [4]-[14]. The latter is more preferable since it can be used for almost all of the mentioned features. The type and the number of the camera for each application are also varied. Self localization is mostly performed using landmark detection, utilizing either ordinary planar camera [4], [11] or panoramic camera [1], [2]. Corridor following and clearance maintenance can be done using single camera [2], [4], [7] or 3 cameras comprising central and peripheral cameras [6], whereas to avoid obstacle some work used 2 cameras configures as stereo cameras [12]. The movement of the robot using is determined using certain control technique based on the above mentioned features [12]. Sometime, it needs to employ Artificial Intelligence techniques to solve the problem, such as Genetic Algorithm [9], Artificial Neural Network [8], [10], [12], [14], and Fuzzy Logic [7]. Fuzzy Logic based navigation system for mobile robots has been developed by many researchers; however most of them used several distance sensors, such as sonar and ultrasonic sensors [15]-[19]. This paper presents the use of vision means using camera for the input of the fuzzy controller. The architectures of the developed fuzzy logic controller mostly consisted of obstacle distance [15], [18], obstacle angle [17], [19], and error angle [18], [19] as inputs; and turning angle [15], [18], [19] or individual wheel speed [17] as output. This paper uses visual sceneries gathered from a camera to determine the existence of walls on the left, the right and the front of the robot as inputs and turning angle and speed as outputs. II. DESIGN CONCEPT A. System Design Fig. 1 shows the schematic diagram of the system which consists of two subsystems: a mobile robot and a fuzzy logic controller. A camera is attached heading forward on the front of the robot (Fig. 2). As the robot roaming around the field, the scenery captured by the camera will change. The image grabbed by the camera is processed. The existence of the walls are quantified and supplied into the fuzzy logic controller to calculate the turning angle and the speed of the robot. Fuzzification Inference Defuzzification Fuzzy Logic Controller Rule base Turning angle Speed Mobile robot Camera Image Quantification Field Scenery Movement left, right, and front wall area Fig. 1. Schematic of the system