Multi-Agent Collaboration for Building Construction

TL;DR

This paper presents algorithms and a task planning framework for multi-agent UAV and UGV collaboration in construction tasks, specifically for the MBZIRC 2020 challenge, demonstrating practical implementation in simulation.

Contribution

It introduces a novel multi-agent collaboration approach for construction tasks using UAVs and UGVs, with implementation in ROS and Gazebo for the MBZIRC challenge.

Findings

Successful simulation of multi-agent construction tasks

Effective task planning algorithms for UAVs and UGVs

Framework adaptable to real hardware implementations

Abstract

This paper details the algorithms involved and task planner for vehicle collaboration in building a structure. This is the problem defined in challenge 2 of Mohammed Bin Zayed International Robotic Challenge 2020 (MBZIRC). The work addresses various aspects of the challenge for Unmanned Aerial Vehicles (UAVs) and Unmanned Ground Vehicle (UGV). The challenge involves repeated pick and place operations using UAVs and UGV to build two structures of different shape and sizes. The algorithms are implemented using the Robot Operating System (ROS) framework and visualised in Gazebo. The whole developed architecture could readily be implemented in suitable hardware.