Abstract:The autonomous exploration of environments by multi-robot systems is a critical task with broad applications in rescue missions, exploration endeavors, and beyond. Current approaches often rely on either greedy frontier selection or end-to-end deep reinforcement learning (DRL) methods, yet these methods are frequently hampered by limitations such as short-sightedness, overlooking long-term implications, and convergence difficulties stemming from the intricate high-dimensional learning space. To address these challenges, this paper introduces an innovative integration strategy that combines the low-dimensional action space efficiency of frontier-based methods with the far-sightedness and optimality of DRL-based approaches. We propose a three-tiered planning framework that first identifies frontiers in free space, creating a sparse map representation that lightens data transmission burdens and reduces the DRL action space's dimensionality. Subsequently, we develop a multi-graph neural network (mGNN) that incorporates states of potential targets and robots, leveraging policy-based reinforcement learning to compute affinities, thereby superseding traditional heuristic utility values. Lastly, we implement local routing planning through subsequence search, which avoids exhaustive sequence traversal. Extensive validation across diverse scenarios and comprehensive simulation results demonstrate the effectiveness of our proposed method. Compared to baseline approaches, our framework achieves environmental exploration with fewer time steps and a notable reduction of over 30% in data transmission, showcasing its superiority in terms of efficiency and performance.

Toward a Generic Framework for Mission Planning and Execution with a Heterogeneous Multi-Robot System

Meeting-Merging-Mission: A Multi-robot Coordinate Framework for Large-Scale Communication-Limited Exploration.

Vision-Based Formation Control for a Heterogeneous Multi-Robot System

Heterogeneous Multi-Robot Collaboration for Coverage Path Planning in Partially Known Dynamic Environments

A Planning Framework for Persistent, Multi-UAV Coverage with Global Deconfliction

A Multi-Robot Task Assignment Framework for Search and Rescue with Heterogeneous Teams

A General-Purpose Application Platform For Multiple Heterogeneous Mobile Robots

3D Multi-Robot Exploration with a Two-Level Coordination Strategy and Prioritization

An Integrated Mission Planning Framework for Sensor Allocation and Path Planning of Heterogeneous Multi-UAV Systems

An Enhanced Hierarchical Planning Framework for Multi-Robot Autonomous Exploration

Tightly-Coupled Perception and Navigation of Heterogeneous Land-Air Robots in Complex Scenarios.

Enabling Large-scale Heterogeneous Collaboration with Opportunistic Communications

Collaborative Planning and Control of Heterogeneous Multi-Ground Unmanned Platforms

A Hierarchical Framework for Long-term and Robust Deployment of Field Ground Robots in Large-Scale Farming

A Framework for Autonomous Robot Deployment with Perfect Demand Satisfaction using Virtual Forces

A Hierarchical Mission Planning System for Multi-Uncrewed Ground Vehicles Using Fast Cost Evaluation and Ant Colony Optimisation

A pack hunting strategy for heterogeneous robots in rescue operations

A Multi-UAV System for Exploration and Target Finding in Cluttered and GPS-Denied Environments

A Framework for Safe Navigation of Unmanned Aerial Vehicles in Unknown Environments

Real-Time Path-Generation and Path-Following Using an Interoperable Multi-Agent Framework

Mission Planning and Execution in Heterogeneous Teams of Aerial Robots supporting Power Line Inspection Operations