Distributed Mobile Robots

Overview

A ROS2-based research platform combining validated distributed auction algorithms with custom rover hardware to enable research in autonomous task allocation and collaborative robotics. The system enables multiple robots to coordinate without central control.

Key Features

Decentralised task allocation without central coordination
Multi-robot coordination and inter-robot communication
Navigation stack integration via Nav2
GPU-accelerated bid calculations in simulation
Failure recovery with time-weighted consensus protocols

Technical Implementation

Algorithm Foundation

Built upon validated distributed auction algorithms offering:

O(K² · b_max/ε) convergence guarantees
≤ 2ε optimality gap for task assignment
Robust consensus protocols for distributed decision-making

Software Stack

ROS2 Humble for robot middleware
Gazebo Garden for simulation
Nav2 for autonomous navigation
Python 3.10+ for algorithm implementation
Docker for containerised deployment

Development Roadmap

The project follows a structured four-phase development:

Foundation: Core infrastructure and single-robot validation
Algorithm Port: Distributed auction implementation in ROS2
Multi-Robot Coordination: Fleet coordination and communication
Validation: Hardware testing and performance benchmarking

Technologies

ROS2 Humble on Ubuntu 22.04 LTS
Gazebo Garden simulation
Python with PyTorch for algorithm validation
Docker for reproducible environments