Fuzzy function estimators as basis on learning from experience

Abstract

This paper describes an alternative method to neural networks for description of high nonlinear systems identification as well as systems in which being nonlinear there are variables with very high time varying rate with respect to the other system variables. It consists in a learning algorithm to be applied in process control, covering several topics of control applications such as system identification, observer design and adaptive control in a simple and useful way which make the method reliable to be applied on industrial process control. Knowledge acquired by means of a proposed learning algorithm is stored into a DAM or FAM (deterministic or fuzzy associative memory) for finally be applied on controller mapping, state observer mapping or model parameter mapping. With such mappings, control design techniques may be applied included the adaptive/learning or hybrid control algorithms.