The adaptive minimal control synthesis algorithm with integral action

The minimal control synthesis (MCS) algorithm is an adaptive control strategy which, for a wide class of plant structure, requires no knowledge of parameter values in order to achieve a stable and robust control performance. In addition, no a priori values for the controller gains are required (they are often set to zero), and no self-tuning component is necessary. This paper describes a significant extension to MCS which incorporates integral action in an explicit manner. The motivation for this new strategy (called MCSI) is threefold: (i) to permit integral control when the controller gains are locked; (ii) to compensate for nonzero mean bias terms on the plant; (iii) to ensure that MCS is structured in an analogous fashion to conventional PID control. The main content of the paper, in addition to the description of MCSI itself, is a stability analysis of the closed-loop dynamics, together with a set of comparative implementation studies on a typical SISO servo control problem. The implementations show that the assertions of the analytic studies are borne out in practice