Self-tuning control with a priori plant knowledge

Most adaptive control methods assume the plant to be completely unknown except for a parameterized structure. Often, however, the plant will have some dynamic modes which are known and do not vary. This paper presents two methods for including a priori plant knowledge in the explicit self-tuning controller (STC). The first method constrains the plant estimates to be consistent with the known information. The second method filters the measurements in an attempt to remove the known dynamics and to estimate only the unknown dynamics. The stability and convergence of these methods is studied. The first method, while stable, is quite numerically complex. The second method is superior. It gives a controller of reduced computational burden and increased robustness to non-persistently exciting signals.