A Particle Swarm Optimization-Based Online Optimization Approach for Virtual Coupling Trains With Communication Delay

Abstract

Virtual coupling allows multiple trains to run synchronously in a virtually coupled formation via train-to-train (T2T) wireless communication. The distance between virtual coupling trains can be reduced greatly to improve the line capacity. However, the time delay in T2T communication cannot be avoided, and the delay may fluctuate due to the uncertainties in the train operation, which will negatively impact the virtual coupling trains. In this article, a particle swarm optimization (PSO)-based online optimization approach for virtual coupling that considers the nondeterministic delay in T2T communication is presented. The optimization objectives include energy consumption, ride comfort, operation speed, and the distance between trains. The PSO algorithm is adopted to solve the problem online so that the train can be controlled according to the optimized traction/braking acceleration. The simulation was conducted with data from the existing Beijing subway line, and the results show that the proposed method reduces the impact of nondeterministic delay on virtual coupling significantly. The speed error of the follower train is below 0.3 m/s, and the tracking distance error is below 0.8 m. Moreover, the requirements of real-time control and ride comfort are satisfied.