Longitudinal fuzzy control of a submerged vehicle

Abstract

Implementation of classical control techniques for the control of a submerged vehicle's movement is considered very complicated, due to the complexity of the motion equations, governing the submerged body's motion (nonlinear 6 degrees of freedom equations), the strong interaction between body's axes and the poor effectiveness of the control surfaces, compared to the vehicle's inertia. However, human operators control the submerged vehicle successfully, utilizing a set of heuristic “thumb rules”, which they acquire through a long and expensive training phase. In this paper we investigate the applicability of fuzzy control in imitating the human operator, in order to achieve a simple and compact representation of the control rules. In order to simplify the problem, we consider a special case – longitudinal motion control. We will show that an efficient and robust controller can be obtained, utilizing the fuzzy control method. We will demonstrate the effectiveness of a simple predictor on human operator's performance, and consequently on fuzzy controller synthesis. Finally, we will investigate the fuzzy controller's performance and compare it to the performance of a conventional controller.