基于神经网络的时延预测算法研究

Research on Neural Network Based Prediction Algorithm of Time Delay

网络控制系统由于时延的存在 ,其稳定性通常难以保证 ,一般要求控制器能够准确预知系统的时延值 .网络形式的不同 ,时延的变化具有规律性或无序性 .针对网络时延变化的不同情况 ,提出了利用线性神经网络和RBF径向基函数网络技术对时延进行预测 .通过对二者预测仿真结果的对比分析 。

Stability is hard to ensure due to time delay in network-based control systems. Generally the accurate value of time delay is needed. The change is regular or not according to the type of network. Aiming at different situations of the change, linear neural network and RBF (Radial Basis Function) neural network are introduced to make time delay prediction. The terms of time delay prediction are obtained by comparing the results of the simulation.