Coalescent multi-robot teaming through ASyMTRe: a formal analysis

Abstract

This paper describes a general approach for automatically synthesizing task solutions for heterogeneous robot teams. In particular, our approach enables multiple robots to coalesce into teams to solve a task through tightly-coupled sensor sharing. Instead of designing special solution strategies for the team, our ASyMTRe approach enables the robot team to generate solutions autonomously according to the current robot team composition. In this paper, we first formulate the problems that the ASyMTRe approach addresses, and then present the anytime ASyMTRe configuration algorithm. We prove that the configuration algorithm is correct, and is guaranteed to find the optimal solution given enough time. Empirical results are also presented validating this analysis, and showing that the ASyMTRe configuration algorithm has good scalability and can quickly find a good solution with the solution quality increasing as additional planning time is available. By analyzing the configuration algorithm, we show that ASyMTRe is applicable to a large class of challenging multi-robot problems