Performance Improvement in Communication-Based Train Control (CBTC) Systems using Cognitive Control

Communication-Based Train Control (CBTC) Systems use bidirectional train-wayside wireless communications to ensure the safe operation of rail vehicles. Due to unreliable wireless networks, the uncertainties in train status information leads to imperfect traction/braking commands, which will affect CBTC performance. In this paper, we use recent advance in cognitive control to positively improve control performance as well as reduce information uncertainties. A cognitive control system composed of Networked Control System (NCS)-based state-space model and entropic-state model is presented. The performance optimization problem based on communication period decision and Medium Access Control (MAC)-layer parameters adaption is formulated as a stochastic control process. Minimizing the energy consumption and information uncertainties is the objective of our model. Simulation result shows that the proposed cognitive control scheme can improve CBTC performance.