The Improvement And Optimization Of A Deadband Scheduling Method

Networked Control Systems(NCS) is a real-time, closed-loop and feedback system which is composed of interconnected sensor, controller, actuator, plant and network. NCS is a new interdiscipline of Communication, Automatic Conrtol and Computer Network. NCS brings the system a lot advantages. However, involved by the network, the system could become extremely complicated to be analyzed and designed. Scheduling strategy can solve some of these problems.For the reason that each node serially transmits data and successively occupies the network bandwidth, both the network Quality of Service(QoS) and the Quality of Control Performance(QoP) of the multi-NCS will decrease for the sake of sharing limited bandwidth. Aiming at the shortcomings of a deadband scheduling method, this thesis improves and optimizes the predicting relation of the error change. Last but not least, based on the idea of Human-Simulated Intelligent, this thesis proposes a method to reduce the network loads by sending data package partly. It could incease the utilization of network bandwidth and improve the QoP to meet the demand of control system.By using Truetime1.5simulation toolbox in the CSMA/CD network evironment, the paper contrasts the QoP and utilization of network bandwidth between before and after using improved method. It takes three loops sharing the the same network with limited bandwidth as example. The result shows that the method improves the system dynamic performance(decrease time delay) and QoP of steady state(decrease overshoot). It decreases the network loads(decrease the amout of sending data package in each loop) and increases the utilization of network bandwidth. To sum above, the effectiveness of the method is verified by the results.