Natural Gradient Shared Control
This addresses the challenge of balancing user autonomy with robotic assistance in shared control systems, though it appears incremental as it builds on existing natural gradient methods.
The authors tackled the problem of shared control between users and robots by proposing a formalism that blends user control with autonomous support using natural gradients from divergence constraints. Their user study on a manipulation task showed more efficient task completion while maintaining user control authority compared to baseline methods.
We propose a formalism for shared control, which is the problem of defining a policy that blends user control and autonomous control. The challenge posed by the shared autonomy system is to maintain user control authority while allowing the robot to support the user. This can be done by enforcing constraints or acting optimally when the intent is clear. Our proposed solution relies on natural gradients emerging from the divergence constraint between the robot and the shared policy. We approximate the Fisher information by sampling a learned robot policy and computing the local gradient to augment the user control when necessary. A user study performed on a manipulation task demonstrates that our approach allows for more efficient task completion while keeping control authority against a number of baseline methods.