Research on Cooperative Control for On-Ramp Merging with Multiple Lanes in Connected and Automated Environments

Abstract

This paper addresses the issue of on-ramp merging in multi-lane freeways and proposes a cooperative control method based on connected and automated vehicles. Focusing on a two-main-lane freeway scenario, the method consists of two key models: a merging sequence decision model and a motion planning model. The merging sequence decision model prioritizes collision avoidance by predicting the motion state and lane-changing trajectory of vehicles in the merging area. The motion planning model utilizes longitudinal and lateral cooperation to control the main lane vehicles to generate the merging gap through coordinated adjustments in the longitudinal speed or by performing lane changes. The optimal merging trajectory is determined using the entropy weight method, and a fast optimization method based on neural networks is employed. Through simulations considering different traffic density combinations, the proposed method is compared with traditional control schemes. Results demonstrate its superiority in improving traffic flow stability, rapidity, and overall efficiency.