Sensors Fusion Technique for Mobile Robot Navigation Using Fuzzy Logic Control System

Abstract

In this paper the optimization of fuzzy rules using modified Fuzzy Associative memory (FAM) is designed and implemented. The behavior rules obtained from FAM is tested in a simulation environment and validated by conducting real world experiments on a popular robot. The experimental results clearly indicate the mapping of multiple inputs to outputs with optimum path in every control cycle of the robot navigation. This approach involves the natural way of dealing with the environments using simple linguistic logic rules without using any mathematical model. The knowledge base of each behavior rule is easy to comprehend, because it captures the behavior rules in a linguistic form. The strength of the proposed methodology is the mapping of the inputs to the output through compositional association of multiple input variables, thus reducing the number of rules without elimination of any of the sensor's input. The robot navigation used in this research is purely perception based and perceived data are optimized and fully used for building navigation rules. Utilization of the proposed FAM methodology for other applications requires minimum modifications during setting of input and output linguistic variables.