Reinforcement Learned Multiagent Cooperative Navigation in Hybrid Environment With Relational Graph Learning | IEEE Journals & Magazine | IEEE Xplore

Reinforcement Learned Multiagent Cooperative Navigation in Hybrid Environment With Relational Graph Learning


Impact Statement:For MCNP, the centralized methods consume huge computational resources and have limited performance when facing unknown dynamic objects. The decentralized methods face th...Show More

Abstract:

The multirobot cooperative navigation problem (MCNP) is an essential topic in multiagent control. This article proposes a distributed approach named GAR-CoNav to solve th...Show More
Impact Statement:
For MCNP, the centralized methods consume huge computational resources and have limited performance when facing unknown dynamic objects. The decentralized methods face the nonstationarity of the environment, and it is difficult to seek an actual cooperative navigation policy. The centralized training and decentralized execution framework shows potential but further research is needed. In summary, this work addresses the challenges of collision avoidance during decentralized execution in hybrid environment, actual cooperative navigation policy facing multiple destinations, and scalability for the changing number of destinations or agents. The insights this study provides can potentially inform future research directions.

Abstract:

The multirobot cooperative navigation problem (MCNP) is an essential topic in multiagent control. This article proposes a distributed approach named GAR-CoNav to solve the navigation problem of multiagent to multiple destinations in the face of static and dynamic obstacles. Agents are expected to travel to different destinations without conflicting with each other to achieve maximum efficiency. That is, cooperative navigation in hybrid environment. The velocity obstacle encoding is combined with a graph to build a global representation, which helps the agent capture complex interactions in hybrid environment. GAR-CoNav processes and aggregates environmental features through the graph attention network and has scalability for the changing number of entities in the graph. A novel reward function is developed to train agents to achieve an actual cooperative navigation policy. Extensive simulation experiments demonstrate that GAR-CoNav achieves better performance than state-of-the-art methods.
Published in: IEEE Transactions on Artificial Intelligence ( Volume: 6, Issue: 1, January 2025)
Page(s): 25 - 36
Date of Publication: 14 August 2024
Electronic ISSN: 2691-4581

Funding Agency:


References

References is not available for this document.