Scheduling Analysis For Tasks With Uncertain Characteristics In Networked Control System

Networked control system (NCS) is an integrated application of control science, computer and network technology. The traditional control theory and methods have new meaning and content due to the competitiveness of network resources and network delay. So the research on scheduling in network control system is an important part. Based on the present results on control performance and scheduling algorithm in the protocol layer, this paper analyses the scheduling algorithm in the application layer in NCS. In order to simulate the multi-task networked control system, the virtual platform is set up.The introduction of bottom control network and network technology lays the foundation of system's model. Two classical scheduling algorithms, RM algorithm and EDF algorithm, are described in the following step. The plat of NCS is set up to simulate. The differences of two algorithms can be presented during the simulation.Secondly, the characteristics of tasks are supposed to be certain, however, it is uncertain. On the basis of the classical EDF, this paper proposes the static extension EDF scheduling algorithm according to the uncertainty of the characteristic. This algorithm can get the distribution function of the schedulability variable according to the distribution functions of tasks with the fuzzy theory. Then the value of schedulability variable which is satisfied the scheduling of the system can be determined by the possibility and necessity of the schedulability variable. The suitable execution time and period of each task can be calculated to decide the priority of each task. So the order of data can be decided. This algorithm is very sample and overcomes the disadvantage of not deciding priorities by fuzzy time in classical scheduling algorithm. Also, it keeps the advantage of high occupancy rate of the resources. It is proved that this algorithm can schedule well under uncertainty.Otherwise, this paper sets up the multi-task simulation model of networked control system. RM, EDF and extension EDF are applied into the model. Simulation results show that extension EDF algorithm can schedule well under uncertainty.