Thesis In Partial Fulfillment of the Requirements for the Degree of Doctor of Sciences / 26 April 2024
Analysis and Synthesis Methods for the Optimal Design of Control Systems in Automated Vehicles
Automated vehicles pose various system analysis and control synthesis challenges on the levels of vehicles, vehicle interactions and transportation systems. Although the complexities and structures of these levels are different, methods with which stability and performance requirements through optimal control systems are guaranteed must be developed. This thesis provides new robust control synthesis frameworks based on the Linear Parameter-Varying (LPV) theory, in which the designed robust controller is in cooperation with unconventional control elements, e.g., learning-based agents. The frameworks developed are used for providing energy-optimal controllers to various control problems. In the thesis the analysis and synthesis methods for the longitudinal control of individual automated vehicles, for the coordination of automated vehicles in intersection, and for the coordination of automated vehicles on the global transportation level are elaborated. The contributions of the thesis on each level are formed. From the viewpoint of control theory, design frameworks for control systems with learning-based agents for different control structures have been formulated. On the level of automated vehicles control, a novel method for achieving energy-optimal motion profile has been provided and implemented, in which several road signals on the forthcoming road horizon have been incorporated. Moreover, the guaranteed frameworks and motion profile methods for handling safety critical interactions of automated vehicles have been transformed, i.e., the complex multivehicle coordinated control problem of cruising vehicles in intersections has been solved. Finally, the coordination of automated vehicles in macroscopic traffic context has been developed, i.e., simulation-based and polynomial analysis methods for exploring the impact of automated vehicles on the traffic flow have been elaborated. The results of the analysis in the control synthesis for improving macroscopic traffic performances have been incorporated.
