Scalable Multi Agent Diffusion Policies for Coverage Control
Frederic Vatnsdal, Romina Garcia Camargo, Saurav Agarwal, Alejandro Ribeiro
AI summary
Problem
Existing decentralized multi-robot coordination strategies often fail to scale or adapt effectively as team size increases, struggling to capture complex inter-agent dependencies and diverse environmental demands.
Approach
Each robot runs a locally conditioned diffusion model that fuses its own sensor data with peer embeddings to sample adaptive control commands, trained via imitation learning from a centralized expert but executed fully decentralized.
Key results
- Consistently outperforms state-of-the-art decentralized baselines on coverage cost
- Generalizes robustly across varying robot densities and feature distributions
- Enables fully decentralized inference via permutation-equivariant spatial transformers
- Leverages diffusion stochasticity to generate diverse, adaptive trajectories in occluded environments
Why it matters
It advances scalable swarm robotics by providing a decentralized, diffusion-based control framework that reliably handles complex, high-dimensional coordination tasks in real-world conditions.
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, cap- turing 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.