Hierarchically Decentralized Heterogeneous Multi-Robot Task Allocation System

TL;DR

This paper presents a hierarchical decentralized auction-based task allocation system for multi-robot systems, enhancing scalability, reliability, and robustness for extraterrestrial resource utilization missions.

Contribution

It introduces a novel hierarchical decentralized coordination architecture with three policies for multi-robot task allocation in ISRU missions.

Findings

The system is scalable and robust in simulations.

Three policies are evaluated for effectiveness.

Decentralized architecture improves flexibility.

Abstract

With plans to send humans to the Moon and further, the supply of resources like oxygen, water, fuel, etc., can be satiated by performing In-Situ Resource Utilization (ISRU), where resources from the extra-terrestrial body are extracted to be utilized. These ISRU missions can be carried out by a Multi-Robot System (MRS). In this research, a high-level auction- based Multi-Robot Task Allocation (MRTA) system is developed for coordinating tasks amongst multiple robots with distinct capabilities. A hierarchical decentralized coordination architecture is implemented in this research to allocate the tasks amongst the robots for achieving intentional cooperation in the Multi-Robot System (MRS). 3 different policies are formulated that govern how robots should act in the multiple auction situations of the auction-based task allocation system proposed in this research, and their performance is evaluated in a 2D simulation called pyrobosim using ROS2. The decentralized coordination architecture and the auction-based MRTA make the MRS highly scalable, reliable, flexible, and robust.