| dc.description.abstract | Over the past few decades, researchers have worked towards developing autonomous systems that can be used in everyday transportation, and with the emergence of new sensor, hardware, and software technologies, the goal of self-driving vehicles is now on the brink of becoming a reality. In order for these systems to properly plan and react to their complex environments, they need to be equipped with the proper tools and algorithms to ensure safe deployment for all stakeholders. Navigating tight spaces with truck and trailer systems in dynamic environments can be a difficult task due to their nonlinear dynamics, delayed actuation, and large physical dimensions. This thesis presents a kinematic approach to local motion planning for truck and trailer vehicles in the forward motion. This approach was applied to the sample-based planning algorithms RRT* and RRTᵡ in order to adapt and replan in the presence of dynamic obstacles. A combined motion planning and control framework was then developed and deployed in both simulations, using American Truck Simulator, and on an International ProStar 122+ truck. After the feedback controllers were iteratively tuned, the motion planners were evaluated alongside a deterministic Hybrid A* approach using a lane change and seaport scenario with simulated static and dynamic obstacles. In both cases, the approach demonstrated the ability for the sample-based planner approach to provide real-time and feasible plans for the controller to execute at low speeds while maintaining a safe distance away from nearby obstacles. | en |
