Trajectory Optimization for Quadrupedal Robot

This project implemented a trajectory optimization algorithm for a quadruped robot with 8 degrees of freedom (DOF), including 4 legs with 2 DOF each (only thigh and knee joints, no hip / abduction joint). The goal is to transfer various hand-designed / heuristic / kinematic trajectories into optimized trajectories that satisfy the simplified dynamics model with constraints, and to eventually transfer to physics-based simulation in MuJoCo.