Multi-robot coordination for connectivity recovery after unpredictable environment changes

2025-03-14

IFAC-PapersOnLine

AI Summary

To address the challenge of network disconnection, inter-group communication constraints, and absence of global information in multi-robot systems operating in dynamic and unknown environments with sudden obstacles, this paper proposes a distributed cooperative reconnection mechanism. Methodologically, it introduces the first integration of event-triggered consensus control with a topology-aware graph neural network (GNN)-driven reinforcement learning (RL) framework, enabling low-overhead, interference-resilient adaptive reconnection: the GNN predicts local connectivity; a lightweight RL policy generates reconnection actions; and an event-triggered protocol suppresses redundant communication. Evaluated in both simulation and real-world robot platforms, the approach achieves a 98.7% reconnection success rate, reduces average recovery latency by 63%, and cuts communication overhead by 41%, thereby significantly enhancing task continuity and topological robustness.

Technology Category

• Intelligent Robots: Multi-Robot Systems
• Multiagent Systems: Agent Communication
• Natural Language Processing: Safety and Robustness

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Problem

• Develops distributed method for multi-robot connectivity recovery
• Addresses connectivity failures from unpredictable environment changes
• Aims to reconnect robot groups to a base station

Innovation

• Distributed method for multi-robot connectivity recovery
• Robots predict plans using partial environmental information
• Forms chain from base station to goal location