Automatic tuning of Smith-predictor design using optimal parameter mismatch

The Smith-predictor is well known for its delay-free characteristic and suitable for regulating systems with an excessively long time delay. Our previous studies (1997) have found that by introducing an appropriate parametric or temporal mismatch can significantly improve the system performance. In this paper, the previous theoretical results are summarized to form an automatic tuning procedure based on the optimal values obtained from closed form integration solutions. The proposed procedure is verified experimentally on a fluid temperature control system of Stanford's hydraulic precision lathe. Test results show good practical feasibility and deserve more real world applications.