Flexible Formation Control Using Hausdorff Distance: A Multi-agent Reinforcement Learning Approach

Pan, Chaoyi; Yan, Yuzi; Zhang, Zexu; Shen, Yuan

doi:10.23919/eusipco55093.2022.9909889

Cited by 6 publications

(2 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…1) Formation reward: In order to compute the maximum individual movement distance required for agents to form an ideal topology, the HD [10] is leveraged and a formation reward r (t) f can be obtained as…”

Section: B Reward Designmentioning

confidence: 99%

“…In particular, instead of centralized location assignment, we adopt Hausdorff-Distance (HD) [10]-oriented multipolicy-distilled ConsMAC for adaptive formation, which is capable to adopt to agent quantity changes. • We verified the effectiveness and superiority of our framework through extensive simulations in the multi-agent particle environment [11] and the quadcopter-physicalmodel-based UAV simulation environment [12].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Multi-Agent Reinforcement Learning-Based Decentralized Spectrum Access in Vehicular Networks With Emergent Communication

Pan

Shan

et al. 2023

IEEE Commun. Lett.

View full text Add to dashboard Cite

Adaptive multi-agent formation control, which requires the formation to flexibly adjust along with the quantity variations of agents in a decentralized manner, belongs to one of the most challenging issues in multi-agent systems, especially under communication-limited constraints. In this paper, we propose a novel Consensus-based Decentralized Adaptive Formation (Cons-DecAF) framework. Specifically, we develop a novel multiagent reinforcement learning method, Consensus-oriented Multi-Agent Communication (ConsMAC), to enable agents to perceive global information and establish the consensus from local states by effectively aggregating neighbor messages. Afterwards, we leverage policy distillation to accomplish the adaptive formation adjustment. Meanwhile, instead of pre-assigning specific positions of agents, we employ a displacement-based formation by Hausdorff distance to significantly improve the formation efficiency. The experimental results through extensive simulations validate that the proposed method has achieved outstanding performance in terms of both speed and stability.

show abstract

Section: B Reward Designmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%