Abstract
This paper proposes a new predictive control system using recurrent RBF networks (RRBFN) and Fuzzy rules. This system is constructed from two kinds of prediction systems and a Fuzzy control system. The prediction systems are constructed from a short-term prediction system and a long-term prediction system. In the short-term prediction system, a RRBFN is inputted the current state of a controlled object, and learns to output the next state of it. In the long-term prediction system, the RRBFN is inputted the state of a controlled object, that was predicted by the short-term prediction system, and predicts the next state of it. The long-term prediction system is repeated this operation n times, and predicts the state of the controlled object on time t+n. The Fuzzy control system controls the controlled object based on the prediction results of the long-term prediction system. We test the proposed method under an outfielder problem in order to investigate its efficiency.
