Towards a Theoretical Foundation of PID Control for Uncertain Nonlinear Systems.

As is well-known, the classical PID control plays a dominating role in various control loops of industrial processes. However, a theory that can explain the rationale why the linear PID can successfully deal with the ubiquitous uncertain nonlinear dynamical systems and a method that can provide explicit design formulae for the PID parameters are still lacking. This paper is a continuation of the authors recent endeavor towards establishing a theoretical foundation of PID. We will investigate the rationale of PID control for a general class of high dimensional second order non-affine uncertain systems. We will show that a three dimensional parameter set can be constructed explicitly, such that whenever the PID parameters are chosen from this set, the closed-loop systems will be globally stable and the regulation error will converge to zero exponentially fast, under some suitable conditions on the system uncertainties. Moreover, we will show that the PD(PI) control can globally stabilize several special classes of high dimensional uncertain nonlinear systems. Furthermore, we will apply the Markus-Yamabe theorem in differential equations to provide a necessary and sufficient condition for the choice of the PI parameters for a class of one dimensional non-affine uncertain systems. These theoretical results show explicitly that the controller parameters are not necessary to be of high gain, and that the ubiquitous PID control does indeed have strong robustness with respect to both the system structure uncertainties and the selection of the controller parameters.