Distributed Model Predictive Control for Two-Dimensional Electric Vehicle Platoon Based on QMIX Algorithm

In this paper, an improved distributed model predictive control (IDMPC) method for the platoon consisting of electric vehicles is put forward. And motion performed in two dimensions, which contains longitudinal and lateral motion. Firstly, a built based on car-following single following vehicle. Then, IDMPC strategy designed with consideration multiple objectives. The symmetrical weight matrices are important final effect. To each vehicle coordinately, weights optimized QMIX algorithm multi-agent reinforcement learning. can fully consider global information multi-vehicle process; therefore, get optimal variables. Finally, simulation experimental results verify IDMPC. Compared to comparison strategies, has better lane tracking, stability direction economic performance.