Human-in-the-Loop Optimization of Shared Autonomy in Assistive Robotics

Abstract

In this paper, we propose a mathematical framework which formalizes user-driven customization of shared autonomy in assistive robotics as a nonlinear optimization problem. Our insight is to allow the end-user, rather than relying on standard optimization techniques, to perform the optimization procedure, thereby allowing us to leave the exact nature of the cost function indeterminate. We ground our formalism with an interactive optimization procedure that customizes control sharing using an assistive robotic arm. We also present a pilot study that explores interactive optimization with end-users. This study was performed with 17 subjects (4 with spinal cord injury, 13 without injury). Results show all subjects were able to converge to an assistance paradigm, suggesting the existence of optimal solutions. Notably, the amount of assistance was not always optimized for task performance. Instead, some subjects favored retaining more control during the execution over better task performance. The study supports the case for user-driven customization and provides guidance for its continued development and study.

Game-Theoretic Modeling of Human Adaptation in Human-Robot Collaboration

“…Shared autonomy alleviates this issue by combining direct teleoperation with autonomous assistance [4][5][6][7][8]. In recent work by Javdani et al, the robot estimates a distribution of Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page.…”

Active Intent Disambiguation for Shared Control Robots

“…Shared control in assistive systems aims to reduce the user's cognitive and physical burden during task execution, typically without having the user relinquish complete control [3], [4], [5], [6]. In order to offset the drop in task performance due to shifting focus from the task at hand to switching between different control modes, various mode switch assistance paradigms have been proposed.…”

“…Heuristic approaches can incorporate domain-specific knowledge easily and are computationally inexpensive, though the trade-off for this simplicity is not being sophisticated enough to incorporate histories of states and actions, making them less robust to external noise. Instantaneous confidence functions for estimating the intended reaching target are employed with success on multiple robotic manipulation systems [5], [17]. In our work we develop an inference algorithm that updates the belief over goals using ideas from dynamic field theory in which the histories of states and actions are incorporated using a single time-scale parameter and robustness to noise is ensured via recurrent selfinteractions that stabilize the dynamical system.…”

Mode Switch Assistance To Maximize Human Intent Disambiguation

“…Even a simple time-optimal automatic mode switching system can significantly improve user satisfaction while maintaining quality task performance [14]. However, it also is not always the case that users are trying to optimize for time or effort during task execution [13]. Our present system therefore does not make a priori assumptions regarding the optimizing principles at work when a user operates a robot.…”