Optimal Parking
Optimal Parking is a small autonomous-parking system that connects geometric path planning with numerical trajectory optimization. The planner starts with an initial path from RRT*, then refines that path through a sequence of quadratic programs while respecting the vehicle model, input bounds, and obstacle margins.
The same C++ planner can be built as a native library or compiled to WebAssembly for the browser demo.
How to read this series
The chapters follow the planner's data flow:
- Problem and vehicle model defines the state, inputs, and parking objective.
- RRT initialization* finds a collision-free geometric route.
- Optimal control defines the objective, dynamics, and constraints for a feasible trajectory.
- Trajectory optimization turns that route into a smooth, dynamically feasible trajectory.
- Constraints and tuning explains obstacle avoidance, bounds, penalties, and convergence.
- Implementation and WebAssembly connects the C++ solver to the interactive demo.
- Series summary puts the complete planning loop into one picture.