Learning robotics from path planning and multi-agent systems to kinematics, dynamics, control and perception by implementing it hands-on.
Path planning and local control to reach a goal while avoiding obstacles. Global planners (BFS · A* · RRT*) and local planners (DWA · Pure Pursuit · MPC), implemented from scratch and benchmarked.
Rigid-body motion · forward/inverse kinematics · dynamics. A read-through study of Modern Robotics (Kevin M. Lynch).
The mathematics under robotics. Proofs, abstract linear algebra, least squares and matrix factorizations, through to the Kalman filter and optimization. A read-through study of ROB 501 (Jessy Grizzle, U. Michigan).
Multi-agent motion planning — robots passing around each other. The four-branch genealogy (priority and coupled search with CBS, sampling, decentralized negotiation with PIBT, kinodynamic) mirrored in C++/Python and run live in an in-browser sandbox.