A Dynamic Fuzzy Neural Adaptive Control algorithm and its application

The uncertainties coursing by the changing of modeling parameters should be considered when modeling and controlling a kind of nonlinear dynamic systems. A Dynamic Fuzzy Neural Adaptive Control (DFNAC) algorithm is presented in the paper. The DFNAC combines a Dynamic Fuzzy Neural Networks (DFNN) with a PID controller. DFNN adjusts its structure and parameters online, and generates the fuzzy rules automatically when being trained. The algorithm conquers the disadvantage of either overfitting or overtraining in traditional static FNN-based control methods. Simulation results of the course control of a container ship validate the effectiveness of the proposed algorithm.