网络控制系统的动态权重变采样周期调度算法

A dynamic weight variable sampling periodical scheduling algorithm for networked control system

摘要:

为提高网络控制系统的性能，提出一种用于网络控制系统的动态权重变采样周期调度算法.首先，通过二次平方根映射函数，根据控制回路数据传输误差对各个控制回路赋予不同的动态权重；然后，根据当前网络运行状况，通过比例控制预测新的网络利用率，同时利用最小二乘支持向量机算法预测各回路数据包传输时间；最后，在网络带宽资源受限的情况下，利用回路动态权重、网络利用率预测值并结合控制回路数据包传输时间，动态的调整回路的采样周期，使得网络控制系统的性能得到优化.仿真实验表明，提出的变采样周期调度算法能够使得网络利用率收敛到设定值，减少了控制回路的数据传输时延，提高了系统的输出响应与控制性能.

Abstract:

In order to improve the performance of networked control system, this paper proposed a dynamic weight variable sampling period scheduling algorithm for networked control system. Firstly, each controlled loop weight is dynamically adjusted according to controlled loop data transmission error through quadratic square root mapping function, the different weight is given for each loop. Then the prediction of the new network utilization through proportional control is given based on the current network status, and least square support vector machine algorithm (LSSVM) is used for controlled loop packet transmission time prediction. Finally, under certain conditions of the network bandwidth limitation, variable sampling time for each controlled loop is determined by using dynamic weight of the controlled loop, network utilization and data transmission time, thus the system performance is optimized. The simulation results show that under the proposed variable sampling periodical scheduling algorithm, the network utilization converges to the setting value, the delay of control loop data transmission is reduced, and the output response performance and control performance of the system are improved.