GA-based adaptive fuzzy logic controller for a robotic arm in the presence of moving obstacle

The design and implementation of a fuzzy logic controller (FLCs) for intelligent control of robotics is very complicated and time consuming process due to articulated robot platforms. In order to automate the controlling and design process, the machine learning techniques are used. In this paper, FLCs and GAs are integrated to control the robot arm in existence of an obstacle. Here, genetic algorithms are employed to optimize the fuzzy knowledge base in three different ways concurrently: select the optimal number of linguistic levels, optimize the parameters of membership functions and optimize the rule base. In this way, evolutionary strategy can be used to minimize the number of mandatory rules and maximize the performance of the fuzzy logic controller by searching for a subset of rule, number of linguistic levels, parameters of membership functions from a given knowledge base. The simulated results depict the validity of the fuzzy logic control system and better performance is obtained.