The LIMO's trajectory was generated using an A* path planning algorithm based on obstacle locations detected within the motion capture environment. The algorithm evaluated possible movements using both heuristic and actual path costs to determine the optimal route through the course.

To execute the generated path, I implemented a feedback control system that continuously adjusted the LIMO's steering angle based on the difference between its measured heading and the vector toward the next target waypoint. Once the robot reached a specified distance threshold from a waypoint, the controller transitioned to the next points until the goal location was reached.

The Bittle X quadruped was controlled using relative position and orientation data between the quadruped and the LIMO obtained from the motion capture system. Before implementing the physical system, I created a digital twin simulation to model and visualize the tracking behavior in real time.

The control strategy used simple directional commands ("left", "right", "forward", and "stop") to maintain a desired following distance and alignment relative to the LIMO. When the quadruped deviated significantly from the target heading, corrective turning commands were issued to reorient the robot toward the LIMO. If the Bittle X approached too closely, it stopped and waited until a safe following distance was restored.

This control logic enables the quadruped to reliably follow the LIMO throughout the obstacle course while adapting to changes in direction and obstacle-avoidance maneuvers.