ADAPTIVE PREDICTIVE CONTROL OF SYSTEMS WITH TIME-VARYING TIME-DELAY

This paper discusses three different methodologies for performing adaptive predictive control in the presence of unknown or time-varying time delays. The first methodology is the well-known extended B-polynomial. The second methodology consists of describing the process dynamics by a network of Laguerre functions. The third methodology, named the variable regressor estimator (VRE), consists of an explicit time delay estimator running in parallel with a standard recursive estimator. Theoretical as well as practical aspects of these methodologies will be discussed. Finally, our industrial experience with these is discussed.