Control of underwater robot using ANFIS control system

Underwater robots are a common subject between robotic science and ocean engineering. Their applications are environmental monitoring, mapping and exploring the depths of the water. In this study, navigation and path tracking of an underwater robot by ANFIS and Memetic Algorithm are desired. Memetic Algorithm is a kind of evolutionary algorithms that uses combination of local searches and other evolutionary algorithms to optimize a problem to achieve to better solutions. In this study due to evolution of behavior in solving of an optimization problem, an adaptive strategy for Memetic Algorithm is presented. By determining an adaptive parameter, the number of population, mutation probability and permission to access local search algorithms are adaptive in each iteration. The algorithm can optimize static and dynamic systems. An ANFIS controller and Memetic are used to Offline control of an underwater robot. ANFIS parameters are trained by Memetic to navigate and control of the underwater robot with minimal error on a predefined path with disturbance condition. Also the ANFIS controller was trained by Genetic Algorithm. Comparison of results shows that Memetic Algorithm training is more stable and more precise.