Degree Name

Doctor of Philosophy


School of Electrical, Computer and Telecommunications Engineering


Information flow topology plays a crucial role in connected autonomous vehicles (CAVs). It describes how CAVs communicate and exchange information with each other. It predominantly affects the platoon's performance, including the convergence time, robustness, stability, and scalability. It also dramatically affects the controller design of CAVs. Therefore, studying information flow topology is necessary to ensure the platoon's stability and improve its performance. Advanced sliding mode controllers and optimisation strategies for information flow topology are investigated in this project.

Firstly, the impact of information flow topology on the platoon is studied regarding tracking ability, fuel economy and driving comfort. A Pareto optimal information flow topology offline searching approach is proposed using a non-dominated sorting genetic algorithm (NSGA-II) to improve the platoon's overall performance while ensuring stability.

Secondly, the concept of asymmetric control is introduced in the topological matrix. For a linear CAVs model with time delay, a sliding mode controller is designed to target the platoon's tracking performance. Moreover, the Lyapunov analysis is used via Riccati inequality to guarantee the platoon's internal stability and input-to-output string stability. Then NSGA-II is used to find the homogeneous Pareto optimal asymmetric degree to improve the platoon's performance. A similar approach is designed for a nonlinear CAVs model to find the Pareto heterogeneous asymmetric degree and improve the platoon's performance.

Thirdly, switching topology is studied to better deal with the platoon's communication problems. A two-step switching topology framework is introduced. In the first step, an offline Pareto optimal topology search with imperfect communication scenarios is applied. The platoon's performance is optimised using a multi-objective evolutionary algorithm based on decomposition (MOEA/D). In the second step, the optimal topology is switched and selected from among the previously obtained Pareto optimal topology candidates in real-time to minimise the control cost. For a continuous nonlinear heterogeneous platoon with actuator faults, a sliding mode controller with an adaptive mechanism is developed. Then, the Lyapunov approach is applied to the platoon's tracking error dynamics, ensuring the systems uniformly ultimately bounded stability and string stability. For a discrete nonlinear heterogeneous platoon with packet loss, a discrete sliding mode controller with a double power reaching law is designed, and a modified MOEA/D with two opposing adaptive mechanisms is applied in the two-step framework.

Simulations verify all the proposed controllers and frameworks, and experiments also test some. The results show the proposed strategy's effectiveness and superiority in optimising the platoon's performance with multiple objectives.

FoR codes (2008)




Unless otherwise indicated, the views expressed in this thesis are those of the author and do not necessarily represent the views of the University of Wollongong.