A new approach for fuzzy controllers

A new type of real-time linguistic fuzzy controller (RTLFC), is proposed which uses support logic programming and evidential reasoning with uncertainty to model the controller and infer the actual control actions. Simplicity and adaptability are emphasised. The new controller takes only the intention of the action into account when choosing the control action. It uses as small amount of data as possible. Compared with the conventional fuzzy controller model, RTLFC obtains its basic knowledge base from the controlled process, the operators´ experience and actions and describes its control rules, inference rules and defuzzification rules in linguistic sentences. These rules, therefore, can be easily modified by operator and/or supervision systems