VDCWorkbench: A Vehicle Dynamics Control Test & Evaluation Library for Model and AI-based Control Approaches

Abstract

The ability to systematically compare and evaluate diversecontrol strategies is essential for the development ofeffective control algorithms in autonomous driving. Thiscontribution presents the VDCWorkbench Modelica Library, aunified platform designed to support the development,testing, validation and verification of vehicle dynamicscontrollers and energy management strategies. The presentedlibrary is an extension of the IEEE VTS Motor VehicleChallenge 2023 models and offers multi-physical componentmodeling, including a hybrid energy storage system (fuelcell & hydrogen tank and battery with aging model), as wellas vehicle dynamics control for autonomous driving researchprojects. Two path-following approaches are featured: anopen-loop lateral controller with a static inversion of asingle-track model, and a closed-loop state-dependentgeometric path-following controller with static controlallocation. The library may also serve as the foundationfor development of vehicle control methods, such astwo-degree-of-freedom control approaches concepts. Oneexample is given for this combination of a feedforwardcontroller with residual reinforcement learning, where alearned agent improves the performance of the open loopcontroller. The entire library will be released as opensource on GitHub in September 2025.