Control of an autonomous underwater vehicle testbed using fuzzy logic and genetic algorithms

In this work, we applied fuzzy logic controllers on the control problem of an autonomous underwater vehicle testbed. Fuzzy logic controllers are rule-based system. Usually, one incorporate knowledge and experiences of experts into the controller design. In some applications, however, it is difficult to find an experienced expert or it is not so intuitive to incorporate expert's knowledge into the control system, particularly when many constraints are imposed on the controller design. In this study, genetic algorithms are applied to obtain a nearly optimal rule base for the fuzzy logic controller in the sense of fitness. Pool tests are conducted to show the testbed performance in its heading control mode. Our results lead to the conclusion that even in tripled sampling time and under parameter variations and disturbances, the fuzzy logic controller tuned by genetic algorithms can greatly enhance the control robustness, and improve the control quality from performance degeneration.