The Research Of Scheduling And Control Co-Design For Networked Control Systems

With the rapid development of control technology, network technology and computer applied technology, networked control system (NCS) are becoming a new kind of control systems and becoming an important issue in automation fields in recent years. In the dissertation, control algorithm, scheduling scheme and co-design of NCSs are studied based on some existing research results both at home and abroad, the main contents are as follow:1. Based on the research of basic structure, characteristics and fundamental problems of NCSs, by focusing on the differences of information networks and control networks, the time-delay characteristic of typical networks is analyzed, and the advantages and disadvantages of different networks are obtained. Then the mathematics modeling is set up under different node-driven modes, an optimal modeling method is presented, which can improve the model accuracy of NCSs. The using method of TrueTime toolbox is introduced, a simulation platform called CAE_NCS based on Ethernet is designed, and the design framework, operation principle and program flow of the platform is described.2. The state space prediction model with time delay characteristic is proposed in NCSs which sensor is time-driven while controller and actuator are event-driven, the application of prediction function control algorithm to NCS is studied, which can better overcome the adverse affects of stochastic long time delay. The application of grey control theory to NCS is studied by regarding NCSs with uncertain time delays and data dropouts as a gray system.3. The information scheduling scheme is studied. Based on the prioritized network environment, by using intelligent control theory, a network architecture which sensor and controller are time-event mixing driven is designed, a two-level scheduling scheme based on BP neural network and fuzzy feedback inference is proposed. The first level scheduling reduces the network load by node-driven method and dead-zone technique, the second level scheduling dynamically schedules priorities of the nodes by using the information scheduler designed based on fuzzy feedback inference, which ensures the fairness of data sending for inferior initial priority nodes. Simulation results show the validity of the scheme. Two-level scheduling scheme can reduce network traffic and optimize information priority scheduling, which improves the control quality.4. The scheduling and control co-design of NCS is studied. The multi-layer design based on predictive control and optimizing sampling frequency is presented. A NCSs model is designed, which the controller is redundant and controller and actuator are event-driven while sensor is time-variable sampling. The NCS is regarded as a gray system, transmission network and plant are regarded as a generalized plant, in order to ensure the stability of the system and adequacy of the network load, a model with optimal sampling frequency scheduling of the sensor is set up, and the optimal sampling frequency is obtained, which ensures the overall performance of the NCS...