Traversing Supervisor Problem: An Approximately Optimal Approach to Multi-Robot Assistance

Abstract

The number of multi-robot systems deployed in field applications has increased dramatically over the years. Despite the recent advancement of navigation algorithms, autonomous robots often encounter challenging situations where the control policy fails and the human assistance is required to resume robot tasks. Human-robot collaboration can help achieve high-levels of autonomy, but monitoring and managing multiple robots at once by a single human supervisor remains a challenging problem. Our goal is to help a supervisor decide which robots to assist in which order such that the team performance can be maximized. We formulate the one-to-many supervision problem in uncertain environments as a dynamic graph traversal problem. An approximation algorithm based on the profitable tour problem on a static graph is developed to solve the original problem, and the approximation error is bounded and analyzed. Our case study on a simulated autonomous farm demonstrates superior team performance than baseline methods in task completion time and human working time, and that our method can be deployed in real-time for robot fleets with moderate size.

Informing Real-Time Corrections in Corrective Shared Autonomy Through Expert Demonstrations

“…autonomy that sequences the execution of two robots based on regions where they may require assistance. Previous work includes investigations of elements of multirobot shared autonomy, such as interfaces for operator attention management [6], supervisor allocation across a fleet of agents (e.g., mobile robots) [3,15,4,17,11,12], and scheduling of agent subtasks and supervision [19,7,1,20]. However, the allocation methods focus on enabling an operator to temporarily teleoperate an agent needing assistance and in scheduling, little to no work focuses on coordinating agents around operator intervention.…”

Fleet-DAgger: Interactive Robot Fleet Learning with Scalable Human Supervision

“…Zheng et al [30] propose to compute the time until stopping for each robot based on its estimated risk and prioritize the robots accordingly. Ji et al [31] consider the setting where physical assistance is required to resume tasks for navigation robots and formalize multi-robot, single-human allocation as graph traversal. Dahiya et al [32] formulate the problem of multi-robot, multi-human allocation during robot execution as a Restless Multi-Armed Bandit problem.…”