A Fast Optimization Method For Automatic Train Stop Control

This paper studies a fast optimization method considering both stopping accuracy and riding comfort in Automatic Train Operation (ATO) systems. ATO system plays an important role in the operation of a train, but stopping accuracy is a difficult problem. We use a predictive model and minimize the stopping error under the constraint of gear switches. The braking acceleration is a function of velocity and gear. We propose a multi-step method which makes each step solvable with limited computing resources on trains. The method computes a sequence of action outputs at each time interval, and only adopts the first output as the gear action of the next time interval. The experiment shows that the stopping accuracy and riding comfort are better than the traditional PID method. The problem is motivated by a practical project for the automatic control of Dongguan-Huizhou intercity railway. Various experiments are conducted based on the real data and scenarios. It is demonstrated that this method is robust and easy to use in different railway lines.