Research On Control Methods Based On CMAC For Network Control Systems

In recent years, with the development of computer network communication technology,its application has been involved in all walks of life. The networked control systems (NCS) isa new product at the cross-point of computer network technology, communication technology,sensor technology and control science. It is the computer network technology in the controlfield of extension and application. The network is used as a transmission medium in NCS torealize the exchange of information between the various components of the system such assensors, controllers and actuators, enabling resource sharing, remote monitoring andcontrolling. NCS has become one of the hotspots of the international academic community inrecent years. However, with the network introduced in control system, the existence oftime-delay and data packet loss will inevitably deteriorate the performance of the controlsystem, resulting in deterioration of the control performance. Worse more, sometimes, it willmake the stability domain small or even not stable.The plant and network are regarded as one time-varying system in this paper, and twocontrol methods for each data packet loss and random delay network control system aredesigned based on the cerebellar model articulation controller (CMAC) to achieve the bettercontrol performance.The first method is designed for network control systems with data packet loss. Thecerebellar model articulation controller and a PD controller are combined to achieve theforward feedback control. And the PD controller parameters are adjusted adaptively by fuzzyreasoning mechanism. And the switching control system between fuzzy PD and PD control isproposed in this paper. According to the size of the deviation, PD control in the smalldeviation is applied to obtain higher static control accuracy, and fuzzy PD control in largedeviations to obtain faster dynamic response and smaller overshoot.The second method is designed for network control systems with time-delay. Smithcompensator is used to compensate the system. And because the time-delay is time-varying,the parameters of the system are time-varying. Forgetting factor recursive least square(FFRLS) is used to estimate the parameters on line. Considered the existing of predictionerror caused by RRFLS, Smith compensator can’t compensate the delay completely.Therefore, at the same time, the CMAC-PD complex controller including a cerebellar model articulation controller and a PD controller is applied to the system. If parameters andnonlinear function are chose properly, it can greatly improve the control performance of thesystem.