Novel extensions of Smith predictors for MIMO systems with multiple I/O delays

Summary form only given. Despite of its enormous popularity and acceptance and many industrial applications and the thousands of articles and many book chapters written on it ( we apologize for mentioning here only very few references), the Smith Predictor (SP) control algorithm, (called also Dead Time Compensator-DTC) remains to date an advanced algorithm for systems with just a single delay. Its properties are well understood and suitable modifications (called modified Smith predictors (MSP)) for unstable SISO systems with delays are well known and documented. SPs structures appear in one form or another in optimal H 2 , L 1 and H ∞ solutions for systems with a single delay. Extending the SP to nonlinear systems with input delay is presently an active research area. The paper presents a brief review of the milestones in the development of the SP in both the academia and industry including the very few extensions of the SP to multiple delays suggested over the years. Through illustrative simple examples the role of the feedforward components will be discussed and clarified. Recent results on performance bounds in both the time and frequency domains, in static cases with multiple I/O delays, show the advantages of the new SPs over the GMDC.