基于神经网络的货运列车速度曲线实时规划算法

A Real-time Speed Profile Planning Algorithm for Freight Trains Based on Neural Network

为降低司机操纵难度、提高运输效能,文章提出一种基于神经网络模型的实时速度曲线规划算法来优化列车操纵指导。该算法首先通过分析二次规划的优化结果,确定神经网络模型的输入与输出形式;并以大量优化结果作为样本数据对神经网络模型进行训练,确定模型的结构与权重;最后将训练出的模型与输入构造模块和牵引计算模块相结合,设计出整个算法流程。为验证该规划算法的正确性、实时性和节能性,分别在虚拟复杂线路和实际线路上进行仿真。结果表明,在保证实时计算的前提下,利用该算法所得的速度曲线与基于二次规划的离线全局规划算法的速度曲线相吻合,且相比于优秀司机的操纵结果节能5.98%。

In order to reduce driver’s operation difficulty and improve transportation efficiency, this paper proposed a real-time speed profile planning algorithm based on neural network to guide the freight train driver’s operation. Firstly, the algorithm determines the input and output forms of the neural network model by analyzing the optimization results of the quadratic programming, and then trains the neural network model with a large number of optimization results as sample data to determine the structure and weights of the model.Finally, the trained model is combined with the input construction module and the traction calculation module to design the entire algorithm flow. In order to verify the correctness, efficiency and energy-saving of the scheduling algorithm, the virtual complex line test and the actual line test were simulated respectively. The results show that the speed profile calculated by this algorithm is identical with the speed profile calculated by the offline global planning algorithm based on quadratic programming, meanwhile, the calculation time can meet real-time requirements. The energy-saving rate is 5.98% as compared to the operation results of experienced drivers.