Research on Optimization Method for Minimum Following Distance of Virtual Coupled Trains Based on Polynomial Regression

This paper proposes an optimization method based on polynomial regression to reduce the minimum following distance of virtual coupled trains in response to the problem of uneven spatiotemporal passenger flow in urban rail transit. By constructing a dynamical model that considers braking force, basic resistance, and random interference factors, and applying polynomial regression algorithm to identify model parameters, accurate prediction of train braking trajectory has been achieved. On this basis, combined with safety constraints, the minimum following distance of virtual coupled trains is generated through simulation calculations. In comparison to traditional methods that rely on a fixed braking rate to achieve relative braking distance, the polynomial regression-based technique proposed herein demonstrates superior efficacy in decreasing following distance across varied speed scenarios. The simulation results show that under different speed conditions, this method can reduce the following distance by about 10%, demonstrates its potential in alleviating urban rail transit congestion.