Hybrid intelligent control algorithm to ship steering based on soft computing

A hybrid intelligent technique is used for ship steering control in this paper. GA optimization is used for the off-line learning period. According to the new definition of fitness function, the optimized result obtained is more suitable to the actual situation. In an online learning period, the reinforcement learning and neural fuzzy control are integrated. It removes the drawback of the conventional hybrid intelligent algorithm that learning must be provided with some sample data, and the ship control quality is effectively improved in the case of appending additional sea state disturbance.