A Differentiable Policy for Shared Autonomy

In this paper we present a framework for the teleoperation of pick and place tasks. We define a shared control policy that allows to blend between direct user control and autonomous control based on user intent inference. One of the main challenges in shared autonomy systems is to define the arbitration function, which decides when to let the autonomous agent take over. In this work, we propose a differentiable policy model that integrates motion generation, user intent inference and arbitration. Full differentialbilty of the policy is desirable to further train the shared autonomy system using Reinforcement Learning (RL). We present initial results teleoperating a gripper in a virtual environment using pre-training and hand tuning of the arbitration function. Our results demonstrate the efficacy of the approach when the intent inference module is trained on a task similar to the one performed at test time. Our results also shed light on limitations that we believe demonstrate the need for a shared autonomy RL setup.