中文摘要：

针对铁路信号在传输过程中存在时延与丢包，导致整个系统的控制性能下降甚至不稳定的问题，提出一种新的延时预测算法，首先用时间序列方法建立网络时延的状态空间模型，然后通过Kalman滤波方法对时延进行在线估计，得到延时预测值。仿真实验结果表明，该方法对传输时延具有预测效果，保证了较好的实时性。

英文摘要：

The existing of delay and the packet loss during the transmission led to deteriorate the control performance and stability of the network control systems, this paper proposed a new prediction algorithm. First of all, the time series method was used to establish the state-space model of network time delay. Then the Kalman filtering method was used for online time delay estimation and the predicted value was got. The simulation results showed that this method had a good effect in time delays prediction and real-time.