Single input fuzzy logic controller tuning for steering control of autonomous underwater vehicle: Genetic algorithm approach

Abstract

Autonomous underwater vehicles (AUV) are robotic devices which perform tasks underwater without operator interference. The paper presents, a simple genetic algorithm (sGA) is employed to tune gains for single input fuzzy logic controller (SIFLC) used for steering control of AUV. SIFLC is a computationally optimized form of conventional fuzzy logic controller (CFLC). In contrast to PD-like CFLC which requires two inputs, error and change in error, SIFLC uses only one input, signed distance. The universe of discourse for the input is tuned using a simple genetic algorithm (GA). GA is an optimization algorithm which mimics biological evolution to find the optimum solution to the problem. Population of likely solution is initialized and fitness of the population is calculated. The fit member of population reproduce while the unfit die out. The reproduction is replicated with help of arithmetic crossover. The population is subjected to mutation, to bring diversity to the population. The elite members of the population are preserved from extinction. The optimum value of universe of discourse are used to simulate yaw control of steering subsystem of NPS AUV II.