A Distributionally Robust Optimization Model for Vehicle Platooning Under Stochastic Disturbances

• Published in: IEEE transactions on vehicular technology Vol. 73; no. 7; pp. 9666 - 9681
• Main Authors: Zhang, Peiyu, Tian, Daxin, Zhou, Jianshan, Duan, Xuting, Zhao, Dezong, Cao, Dongpu
• Format: Journal Article
• Language: English
• Published: New York IEEE 01.07.2024; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Chance constraints; Closed loops; communication uncertainty; Computational modeling; Control stability; Control systems; distributionally robust optimization; Disturbances; Feasibility studies; model predictive control; Noise control; Noise prediction; Optimization models; Platooning; Predictive control; Random variables; Real-time systems; Robust control; Robustness; Stability; State feedback; Stochastic processes; Uncertainty; vehicle platooning; Vehicles; Vehicular ad hoc networks
• ISSN: 0018-9545; 1939-9359
• DOI: 10.1109/TVT.2024.3375301

Inspired by connected and autonomous driving technologies, this paper proposes a closed-loop Distributionally Robust Model Predictive Control (DRMPC) method to address the problem of longitudinal platoon control disturbed by V2V communication noise. In particular, a Model Predictive Control (MPC)-based vehicle platoon control model subject to stochastic disturbances is first developed. Vehicle control and state are imposed with probabilistic chance constraints, and a state feedback structure is designed to ensure the stability of the platoon system, which poses a significant challenge to the platoon control system. To solve this computationally intractable DRMPC model, a Ball ambiguity set is constructed using the characteristic information (expectation and variance) of random variables. The original DRMPC model is reformulated into a computationally tractable robust counterpart approximation framework. Furthermore, the recursive feasibility of the proposed DRMPC and the string stability of the platoon vehicles are demonstrated by introducing an initialization strategy for nominal states. Finally, a simulation study in a platooning system consisting of six vehicles is performed to verify the validity of the DRMPC model under stochastic V2V noise disturbances.