Discrete Multi-Objective Switching Topology Sliding Mode Control of Connected Autonomous Vehicles With Packet Loss

Information flow topology is crucial in the control of connected autonomous vehicles. It substantially influences the platoon's performance. This study provides a real-time switching topology technique for improving the platoon's performance under poor communication. First, a discrete sliding mode controller with a double power reaching law is designed for a nonlinear heterogeneous vehicle dynamic model with packet loss. Then, Lyapunov analysis is applied to ensure the platoon's stability and string stability. Finally, a two-step switching topology framework is introduced. The first step is to search for Pareto optimal topology offline with predicted imperfect communication scenarios. Then, the platoon's overall performance is optimised using a multi-objective evolutionary algorithm. In the second step, the optimal topology is selected and switched in real-time to reduce the control cost. The proposed method maximises the advantages of information flow topology. It deals with poor communication, improves the platoon's performance, and ensures stability. Numerical simulations were conducted to validate the proposed approach. Compared to a standard robust sliding mode controller, the suggested technique enhances platoon tracking ability by 97.73 percent, fuel efficiency by 9.96 percent, and driving comfort by 20.18 percent, respectively.

Automated summary

This study presents a real-time switching topology technique to improve the performance of connected autonomous vehicles under poor communication. A sliding mode controller is designed for a nonlinear heterogeneous vehicle model with packet loss, and Lyapunov analysis is applied to ensure stability. A two-step switching topology framework is introduced to optimize the platoon's performance. Numerical simulations show significant improvements in platoon tracking ability, fuel efficiency, and driving comfort.