Energy-Saving Automatic Train Regulation Using Dual Heuristic Programming

Issues regarding environment sustainability and energy saving have been receiving much attention in the worldwide railway society. Saving energy via train regulation is cost effective and has become an issue to be considered in railway operations. Automatic train regulation (ATR) plays an important role in maintaining metro service quality; however, designing ATR is an optimal control problem with large scale, high nonlinearity, heavy constraints, and stochastic characteristics. Considering issues regarding energy saving in the ATR design would further complicate the optimization problem. In this paper, traffic environment associated with energy consumption resulting from traffic regulation is investigated, and a model for describing the environment is developed; thereby, a dual heuristic programming (DHP) method for designing ATR with energy saving via coasting and station dwell control is proposed. The evaluation of the ATR design is conducted with field data. The evaluation shows that better traffic regulation with higher energy efficiency is attainable with the ATR design.