DiffTOP: Differentiable Trajectory Optimization for Deep Reinforcement and Imitation Learning

This paper introduces DiffTOP, which utilizes Differentiable Trajectory OPtimization as the policy representation to generate actions for deep reinforcement and imitation learning. Trajectory optimization is a powerful and widely used algorithm in control, parameterized by a cost and a dynamics function. The key to our approach is to leverage the recent progress in differentiable trajectory optimization, which enables computing the gradients of the loss with respect to the parameters of trajectory optimization. As a result, the cost and dynamics functions …

algorithm control cost cs.ai cs.lg cs.ro differentiable dynamics function generate imitation learning key optimization paper policy reinforcement representation the key trajectory