关键词: 多功能车辆总线控制器, 车门控制, 增量式 PID 控制, 长短期记忆网络, 车门防夹

Abstract: In order to improve the reliability and safety of subway door control, a subway door control system based on LSTM and incremental PID algorithm was proposed. Among them, multifunction vehicle bus ( MVB) was adopted to build the subway door central control system, train communication network ( TCN) was used to build the gating control unit, the incremental PID algorithm was utilized to control the door motor and long short-term memory ( LSTM) neural network was used to achieve door anti-pinch obstacle avoidance. The results show that control accuracy of incremental PID algorithm for motor speed was 98. 36%, which was significantly higher than that of traditional fuzzy control. Compared with traditional ripple method, speed comparison method and current method, the anti-pinch force of door anti-pinch obstacle avoidance based on LSTM was the smallest. After conducting 10 experiments on opening and closing the door, the success rate of opening and closing the door was highest and the time taken was shorter, when the the opening and closing time was controlled in 2. 2 ~2. 4 s. It is concluded that this method can meet the requirement of subway door control and has high safety and reliability.

Key words: multifunction vehicle bus ( MVB ) controller, door control, incremental PID control, long short-term memory( LSTM), door anti-pinch