A Murdoch-Based ROS Package for Multi-robot Task Allocation

Abstract

Due to the rapid development of robotics application in a wide range of industries and the rapid growing complexity of tasks coordination, researches in Multi-Robot Task Allocation (MRTA) systems have increased swiftly in recent years. This work implemented an auction-based MRTA system using the MURDOCH architecture on ROS. Murdoch is an architecture which allows robots to place bids on tasks that they are capable of executing, and it selects the most fitted robot based on bids values calculated by a fitness function. The aim of this work is to present an adapted MURDOCH architecture that focused on proving a more generic MRTA solution, which may be applied to wider range of problems. Using ROS and a few other tools, our architecture was successfully validated, and it implemented this MURDOCH architecture in a readily scalable way. A few possible approaches for future improvements were also suggested.