Model reference adaptive control of nonlinear systems using RLLM networks

A model reference adaptive controller for nonlinear systems is implemented using rectangular local linear model (RLLM) network approach. An RLLM network is used to identify the plant as NARX model. Another network of the same kind consisting of linear controllers is tuned to achieve the closed-loop behaviour of a linear reference model. A master slave configuration is realised to decide the activation of candidate linear controllers in the controller network. The implemented control scheme is tested in simulations and real-time control on a hydraulic positioning system.