Real-time optimal control of tracking running for high-speed electric multiple unit

High-speed electric multiple unit (HSEMU) is an advanced transportation system, which share some special characteristics such as nonlinear dynamics, complex working environment and time-varying operational conditions. To achieve a safe, punctual and automatic running of HSEMU, effective modeling methods of HSEMU and real-time optimal control schemes are useful to further improve the running performances. This paper aims to develop intelligent control techniques for speed tracking control of HSEMU. The well-known adaptive neuro-fuzzy inference system (ANFIS) is employed to model the dynamics, and an ANFIS-based generalized predictive controller is proposed with a stability analysis of the closed-loop system. Experimental results with comparisons on real-world running data demonstrate that our proposed techniques in this work can perform favorably in terms of both safety and punctuality.