An adaptive dynamic programming method for observer‐based sliding mode control of connected vehicles subject to deception attacks

• Published in: International journal of robust and nonlinear control Vol. 34; no. 6; pp. 3659 - 3678
• Main Authors: Xu, Yangguang, Guo, Ge, Yu, Shuanghe
• Format: Journal Article
• Language: English
• Published: Bognor Regis Wiley Subscription Services, Inc 01.04.2024
• Subjects: Acceleration; adaptive dynamic programming; connected vehicles; Control methods; cyber‐attacks; Dynamic programming; Dynamic stability; neural network; Neural networks; Nonlinear dynamics; Optimal control; Sliding mode control; Vehicles
• ISSN: 1049-8923; 1099-1239
• DOI: 10.1002/rnc.7155

This article investigates the problem of optimal observer‐based sliding mode control (SMC) of connected vehicles subject to deception attacks and disturbances with adaptive dynamic programming (ADP) method. For a group of vehicles with unknown nonlinear dynamics term and disturbance, this article aims to give a control methodology to achieve secure tracking of the desired spacing, velocity and acceleration. A neural network (NN) and an observer are constructed to estimate the unknown nonlinear term and the states, respectively. Then, a SMC scheme incorporating NN approximation is developed and an off‐policy ADP method is used to implement the optimal control of sliding mode dynamics. The proposed method can ensure individual stability and string stability of the set of vehicles. Numerical simulations are conducted to demonstrate the validity of the proposed controller.