Distributed Model Predictive Control for Two-Dimensional Electric Vehicle Platoon Based on QMIX Algorithm
Open Access
Abstract
In this paper, an improved distributed model predictive control (IDMPC) method for the platoon consisting of electric vehicles is put forward. And the motion of the platoon is performed in two dimensions, which contains longitudinal motion and lateral motion. Firstly, a platoon model is built based on the car-following model for a single following vehicle. Then, the IDMPC strategy is designed with the consideration of multiple objectives. The symmetrical weight matrices in the IDMPC are important for the final control effect. To control each following vehicle in the platoon coordinately, the weights for the IDMPC are optimized based on the QMIX algorithm in multi-agent reinforcement learning. The QMIX can fully consider the global information in the multi-vehicle following process; therefore, the IDMPC can get optimal control variables. Finally, the simulation and experimental results verify the IDMPC. Compared to the comparison strategies, the IDMPC has the better lane tracking, stability in lateral direction and economic performance.Keywords
Funding Information
- National Natural Science Foundation of China (U1713213, U1913202, U1813205)
- Key-Area Research and Development Program of Guangdong Province (2019B090915001)
- Shenzhen Technology Project (JCYJ20180507182610734, JSGG20191129094012321)
This publication has 29 references indexed in Scilit:
- Predictive energy-saving optimization based on nonlinear model predictive control for cooperative connected vehicles platoon with V2V communicationEnergy, 2019
- Energy savings from an Eco-Cooperative Adaptive Cruise Control: a BEV platoon investigation2019
- Anti-jerk model predictive cruise control for connected electric vehicles with changing road conditions2017
- Model predictive control-based efficient energy recovery control strategy for regenerative braking system of hybrid electric busEnergy Conversion and Management, 2016
- Optimal battery charging, Part I: Minimizing time-to-charge, energy loss, and temperature rise for OCV-resistance battery modelJournal of Power Sources, 2016
- Human-level control through deep reinforcement learningNature, 2015
- Model predictive control for improving operational efficiency of overhead cranesNonlinear Dynamics, 2014
- Practical String Stability of Platoon of Adaptive Cruise Control VehiclesIEEE Transactions on Intelligent Transportation Systems, 2011
- Model Predictive Multi-Objective Vehicular Adaptive Cruise ControlIEEE Transactions on Control Systems Technology, 2010
- Multi-Agent Reinforcement Learning: Independent vs. Cooperative Agents1993