基于线性自适应神经网络的摆式列车横向加速度预测研究

Study on prediction of lateral acceleration of tilting train based on linear adaptive neural networks

阐述了用线性自适应神经网络对即将输入的控制参考信号进行多步在线自适应预测并编程实现的方法。对实测信号的仿真分析表明,线性自适应网络可以以满意的精度对摆式列车横向加速度进行多步预测,有效解决由于各种因素造成的滞后补偿问题。

The way is elaborated to predict multi-step online the reference control signal to be input soon with the linear adaptiveneural network and so is the way of realization by programming. The simulation analysis of measured signals show that the linear adaptivenetwork is able to predict multi-step the lateral acceleration of tilting train with satisfying accuracy, and it is able to solve efficiently lagcompensation problems caused by various reasons.