Memory-based Control of Nonlinear Dynamic Systems Part II- Applications

In the first part of the paper we presented a memory-based approach for nonlinear system control. The method is not to assume the system is described by a linear model plus perturbations; not to linearize the system; not to estimate certain parameters based on the linear parametric assumption; not to determine the bounds on certain nonlinear functions; not to use infinite switch frequencies; not to involve ad hoc membership functions; not to run the system repeatedly for the same task. Instead, the control scheme is solely based upon certain memorized information such as current system response, previous system response and past control experience. Fundamentally, the desired control signal in the scheme is "learned" and generated from observing and processing the most recent experience stored in a memory. These features are confirmed via applying this method to via numerous examples considered by other researchers in the literature. Simulation results are presented here