Obstacle Avoidance of a Mobile Robot Using Fuzzy Logic Control

During the past several years fuzzy logic control (FLC) has emerged as one of the most active and fiuitful areas for research in the application of intelligent system design. Presently, fuzzy logic has found a variety of applications in various fields ranging from industrial process control to medical diagnosis and securities trading. Most notably, a fuzzy logic system has been applied to control nonlinear, time-varying, ill-defined systems, to control systems whose dynamics are not exactly known, as servomotors position control, and robot arm control, and to manage complex decision-making or diagnostic systems. This project has the objective of designing a fuzzy logic controller, which will be used to control the navigation process of an autonomous mobile robot in a completely unstructured environment. The navigation algorithm is proposed for static obstacles and with no priori knowledge about the environment. In addition, an on-line path planning is used while navigation. The controller will have its inputs from the sensors that will be mounted on the robot. The number of sensors used is five where, three of them will be on the front side of the robot, whereas, one on the left side and one on the right side. The FLC was designed using three different fuzzifiers (triangular, trapezoidal and Gaussian) to represent the sensor readings values so that they can be interpreted by the inference mechanism. Moreover, two different implication methods (Mamdani minimum and Mamdani Product) implications are used in the interpretation of the IF-THEN rules in the rule-base. Depending on the number of fuzzy sets used to represent the sensor readings, the total number of control rules used in the design was 243 at the first stage and then reduced to 1 08. In other words, if the number of fuzzy sets used to represent each sensor reading is three (far, near, and very near) then the total number of rules is 243 which is (35). On the other hand, if the left and right sensors reading values were represented using only two fuzzy sets (far and near) then the total number of rules is 1 08 i.e. (33 *22 ). In addition, two defuzzification strategies (center of gravity and center average) were used to get the output of the FLC in a crisp value. It was observed that the triangular fuzzifier, center average defuzzification method, and the Mamdani minimum implication method with a total number of 108 rule are the best choices for the design.

Text FK_1999_6.pdf Download (1MB)

Actions (login required)

View Item