Iterative Learning Control Approach to Fuzzy Control Systems Development

The paper suggests a new development method for fuzzy control systems involving iterative learning control (ILC) algorithms in terms of two-level hierarchical control system structures. The method employs a simplified version of the norm-optimal ILC on the higher hierarchical level and fuzzy controllers on the lower level. The fuzzy controllers on the lower hierarchical level replace the initial linear controllers in order to obtain the improvement of control system performance indices based on the transfer of results from the linear development case to the fuzzy one in terms of the modal equivalence principle. The new method and original fuzzy control structure are validated by the development of Mamdani PI-fuzzy controllers dedicated to speed control of a nonlinear DC drive accompanied by real-time experimental results.