Learning the optimal state-feedback via supervised imitation learning

Dharmesh Tailor, Dario Izzo

Imitation learning is a control design paradigm that seeks to learn a control policy reproducing demonstrations from expert agents. By substituting expert demonstrations for optima…