Prediction-Based Event-Triggered Platooning Control of Heterogeneous Connected Vehicles Under Stochastic Deception Attacks

Abstract

This article presents a new prediction-based dynamic event-triggered platooning control scheme for heterogeneous connected vehicles under stochastic deception attacks and constrained communication bandwidth. By introducing nominal powertrain coefficients, disturbed longitudinal dynamic models are constructed to account for vehicle heterogeneity. Vehicle state trajectories are predicted using both local dynamics and neighboring information. Stochastic deception attacks are modeled via a Bernoulli process. To alleviate the communication burden, a state-prediction dynamic event-triggered mechanism (DETM) is developed. The secure controller and event-triggered parameters are jointly designed by solving semidefinite programming problems. Input-to-state stability (ISS) theory is applied to establish the string stability of the platoon system under attacks. Cosimulations in MATLAB/CarSim validate the effectiveness of the proposed method, which reduces the triggering frequency by 57.8% relative to conventional platooning control approaches.