Scalable Multi Agent Diffusion Policies for Coverage Control

2025-09-21

AI Summary

This paper addresses decentralized multi-robot coverage control—a canonical distributed navigation problem—by proposing MADP, the first framework to integrate diffusion models into decentralized multi-agent cooperative decision-making. MADP employs a spatial Transformer to fuse ego-observations with neighbor-perceived information, modeling the joint action distribution for fully decentralized inference; it is trained end-to-end via imitation learning from oracle expert trajectories. Its core contribution lies in pioneering the use of diffusion models for multi-robot coordinated action generation, thereby overcoming the restrictive Gaussian assumptions of conventional approaches and significantly enhancing expressiveness for complex, multimodal action distributions. Experiments demonstrate that MADP consistently outperforms state-of-the-art baselines across diverse environments and agent densities, exhibiting strong robustness, generalization, and scalability.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Planning, Routing, and Scheduling: Planning with Markov Models (MDPs, POMDPs)
• Intelligent Robots: Multi-Robot Systems

Application Category

• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

We propose MADP, a novel diffusion-model-based approach for collaboration in decentralized robot swarms. MADP leverages diffusion models to generate samples from complex and high-dimensional action distributions that capture the interdependencies between agents' actions. Each robot conditions policy sampling on a fused representation of its own observations and perceptual embeddings received from peers. To evaluate this approach, we task a team of holonomic robots piloted by MADP to address coverage control-a canonical multi agent navigation problem. The policy is trained via imitation learning from a clairvoyant expert on the coverage control problem, with the diffusion process parameterized by a spatial transformer architecture to enable decentralized inference. We evaluate the system under varying numbers, locations, and variances of importance density functions, capturing the robustness demands of real-world coverage tasks. Experiments demonstrate that our model inherits valuable properties from diffusion models, generalizing across agent densities and environments, and consistently outperforming state-of-the-art baselines.

Problem

• Addressing coverage control in decentralized robot swarms
• Generating coordinated actions from complex multi-agent distributions
• Enabling robust generalization across varying agent densities and environments

Innovation

• Diffusion models generate multi-agent action distributions
• Policy sampling uses fused local and peer observations
• Spatial transformer enables decentralized imitation learning inference