Name: Yifan Cai
Date: Dec 12, 2022
Time: 1:00pm
Location: online
Supervisors: Stephan Smith
Title: Online Scheduling of Operator Assistance for Multi-Robot Teams with Uncertain Robot Capabilities and Environments
Abstract: In this study, we consider the problem of allocating human operator assistance in a system with multiple autonomous robots. Each robot is assigned with an independent mission, each defined as a sequence of tasks. While executing a task, a robot can either operate autonomously or be teleoperated by the human operator to complete the task at a faster rate. We are interested in finding a schedule of teleoperated tasks in order to minimize the system makespan. Both deterministic and stochastic models of robot task completion times are considered in this study. We first show that the deterministic problem of finding the optimal teleoperation schedule is NP-Hard. We then formulate the problem as a Mixed Integer Linear Program, which can be used to optimally solve small to moderate-sized instances. We also develop an anytime algorithm that makes use of the system structure to provide a fast and high-quality solution of the operator scheduling problem, even for larger instances. Our key insight in this algorithm is to identify blocking tasks in greedily-created schedules and iteratively remove those blocks to improve the quality of the solution. Through numerical simulations, we demonstrate the benefits of the proposed algorithm as an efficient and scalable approach that outperforms other common solution techniques. Expanding research to the stochastic setting, where task duration is exponential random variable to represent uncertainty of robot capabilities and environments, we developed a parameterized replanning policy. This policy selectively chooses to update the schedule based on task observations. The parameter can be used to control the trade-off between performance and efficiency. The resulting policy demonstrates good planning competence in both average and worst cases. Results also show significant reduction in resource requirements for replanning, with little to no compromise in performance, when compared to the policy that replans on every task completion.