Network Control System, Scheduling And Simulation

This thesis is based on the project "Integrated design of Control and Real-TimeScheduling in networked control systems" sponsored by the National Science Foundation.Network control system (NCS) is a synthetic application which involves control science,computer science and network technology. The competition of network resources andpropagation delay raise a new scientific field where traditional control theories andmethodology can be applied. Network scheduling, for instance, is a essential part of thisnovel domain. The contributions of this dissertation are summarized as following:In the first section, the characteristic of three main types of control networks isanalysized, their advantages and disadvantages in the network control and theirperformances under certain networks are clarified.Then, the NCS platform is establishedby using TmeTime simulator. This thesis studies the network influences on the controlsystem via the simulation of the PID control system of DC motor;In the second section, four scheduling algorithms in protocol layer are investigated,and the corresponding simulation is carried on the platform. Moreover, the advantages anddisadvantages of the scheduling algorithms is clarified;In the third section, the former scheduling algorithms is applied to NCS with periodicdata and aperiodic data. The simulation results show that they can't satisfy the demands ofthe system. As the result, this thesis focuses on the Mixed Traffic Scheduling algorithm.MTS is an algorithm based on the task's real-time demands. This thesis analyses thefeasibility of the MTS, on the other side, it validates the feasibility and validity bysimulation;In this dissertation, four scheduling algorithms in protocol layer and the Mixed TrafficScheduling algorithm are studied thoroughly. An example is worked out in detail todemonstrate that the algorithm satisfies performance requirements for NCS, optimizesnetwork scheduling, and improves network efficiency.