Fuzzy Logic And Modern Control: An Analysis Of Controller Performance Under Feedback Errors

Fuzzy logic is well suited to the control of non-linear, time-varying, and/or any system for which it is difficult to obtain an exact mathematical model. To design a good fuzzy controller, a thorough understanding of the desired process behavior is required. Many papers have suggested that fuzzy controllers are robust. If they are, it is reasonable to explore whether or not they respond better than conventional controllers to errors in the feedback signal due to imperfect measurement or sensor degradation. It may be the case that simpler or less expensive sensors could be used in some processes. The research presented here relies on an integrated array of computer programs that model and simulate (1) the process to be controlled, (2) common feedback instrumentation, and (3) both fuzzy logic controllers (FLC's) and conventional controllers. Various types of feedback sensor imperfections are imposed upon the system, and the performance of the two controllers is compared. Each controller is tuned to its optimal performance in response to a setpoint change using perfect feedback measurement.