Driankov, Hellendoorn, and Reinfrank's An Introduction to Fuzzy Control
Kevin M. Passino · Journal of Intelligent & Fuzzy Systems · 1995
Fuzzy control has seen some success in challenging applications; however, despite the hype and frequent extraordinary claims about the capabilities of fuzzy control, this new technology is still in its infancy and hence is frequently in need of careful engineering cost-benefit analysis to determine its advantages and disadvantages so that it is not indiscriminately applied.There is a particular need for comparative analyses with conventional control approaches (which have a much longer, more well established track record, and are much more widely used) and the need for the use of nonlinear analysis techniques for studying the stability, controllability, observability, and robustness properties (the fuzzy controller is a nonlinear controller and is therefore amenable to such analysis).Such mathematical analysis has significant value for helping to verify that a control system will behave properly (especially if safety is a concern), but just like for conventional control systems the analysis techniques do not tend to be applicable (e.g., due to structural assumptions on the model of the plant) for very complex real-world control problems.In these cases, we often heuristically construct a controller, use approximate models and hence approximate mathematical analysis, and invest a significant effort into validation via extensive simulations and experimental evaluations.Regardless, for many serious applications there is a clear need for careful engineering analysis by control engineers who understand the implications of nonlinearities, disturbances, noise, and complex time-varying dynamics.Even though fuzzy control is easy to understand, the problems