I built an autonomous parking system in Python capable of maneuvering three distinct vehicle types into a parking space: a differential drive delivery robot, an Ackermann-steered car, and an Ackermann-steered truck with a trailer. The planner uses a Hybrid A* search that explores continuous vehicle poses while maintaining a discretized grid for state tracking, with a cost function combining path length and a Euclidean distance heuristic.

The planner adapts its motion primitives as it goes: using coarser movements early on, then switching to finer motion primitives as the vehicle gets closer to the parking spot, enabling precise alignment where the whole vehicle fits inside the space.