两级模糊神经网络在高速列车ATO系统中的应用研究

列车自动驾驶系统(ATO)作为替代司机、实现铁路运输自动化的重要设备,受到国内外铁路科技工作者的广泛关注,它的性能关系到铁路运输系统的安全和效率。一种高效的高速列车控制方法可以很好地满足人们对高速铁路"安全、正点、舒适和快捷"的要求。由于模糊神经网络具有自动提取经验和进行推理的特点,在前人研究的基础上,本文用两级模糊神经网络对高速列车运行过程进行控制。参照当前列车运行速度、线路状况、列车编组、列车时刻表、距目标点距离以及目标点所允许的速度,前一子网模拟优秀司机的操纵可获得在当前状态下应采取的最佳列车运行工况,后一子网根据运行工况、线路状况、列车在线路上位置、列车编组、列车时刻表、距目标点距离和目标点允许速度,得到列车在该工况及当前条件下的运行速度。仿真实验结果表明该方法正确有效,达到了期望的效果。

Research on Application of Two-degree Fuzzy Neural Network in ATO of High Speed Train

As the important equipment to take the train driver's place and realize railway transportation automation,the automatic train operation(ATO) system draws extensive attention among domestic and overseas scientific and technical workers.ATO performance concerns safety and productivity of railway transport systems.A good high-speed train control method can satisfy people's demands for safety,punctuality,comfort and fast convenience of high speed railways.On the basis of preceding research and the fact that fuzzy neural networks have the characteristics of abstracting experience and implementing deduction,two-degree fuzzy neural networks are used to control high speed train running processes.The first sub-network simulates on excellent train driver's operation to obtain the optimal train running conditions with reference to the train travelling speed,railway line conditions,train formation,train timetable,distance to the object point and corresponding permissible speed at the present time.The next sub-network is used to deduce the train running speed from the obtained train running conditions and the present railway traffic conditions,such as the railway line conditions,train location on the line,train formation,train timetable,distance to the object point and permissible velocity at the object point on the track.Simulation results show that the solution to the high speed train control is correct and effective as anticipated.