Imitation on BiconvexMPC

This project focuses on imitation learning of policies generated in simulation by BiconvexMPC. The goal of this project was to imitate the larger horizon policy generated by BiconvexMPC in simulation which are infeasible to run on the real robot due to large computational requirements. The idea of the work was to learn these policies in the simulation using imitation learning and transfer it to the real robot using robust sim2real transfer approach.