Dec-MARVEL: Decentralized Multi-Agent Exploration without Communication under Budget Constraints

2026-07-09

AI Summary

This work addresses the challenging problem of multi-UAV cooperative exploration under stringent constraints—namely, no inter-agent communication, limited field of view, and strict travel budgets. The authors propose Dec-MARVEL, a decentralized framework that, for the first time, enables coordination using only local visual observations of teammates. It employs a graph attention policy to jointly reason about frontier geometry, teammate motion, and remaining budget to plan safe return-feasible waypoints and orientations. Key innovations include a phase-conditioned critic, a task-oriented privileged critic, and a hybrid budget-aware curriculum training scheme. Experiments across nine team-size and budget configurations demonstrate that Dec-MARVEL consistently achieves the highest (or tied-highest) exploration rates with the lowest perceptual overlap. Under a tight 720-meter budget, success rates reach 53%, 94%, and 100% for 2, 4, and 8 UAVs, respectively—substantially outperforming the strongest baseline—and the method successfully transfers from simulation to real-robot deployment.

Abstract

Multi-UAV exploration is often constrained by unreliable communication, limited field-of-view sensing (e.g., lightweight onboard camera), and finite travel budgets that require each robot to reserve enough budget to return to its base. We present Dec-MARVEL, a decentralized budget-aware exploration framework for communication-free teams with directional sensing. Rather than exchanging maps, goals, or messages, each robot coordinates through its incidental observations: any teammate trajectory within its field of view serves as a coordination signal. A graph-attention actor fuses local frontier geometry, teammate motion, and budget features to select return-feasible waypoint-heading actions. The actor is trained with phase-conditioned critics, a training-only task-oriented privileged critic, and a mixture-based budget curriculum. Across 900 held-out trials spanning three team sizes (2, 4, 8 robots) and three travel budgets (720, 800, 1024 meters) against four baselines, Dec-MARVEL achieves the highest or tied-highest exploration rate and lowest sensing overlap across all nine team-size budget configurations. Under our tightest 720m budget, it reaches 53%, 94%, and 100% success for 2, 4, and 8 robots, versus 37%, 83%, and 99% for the strongest baseline. Physical-robot experiments demonstrate successful sim-to-real transfer and real-world deployment of Dec-MARVEL.

Problem

• decentralized exploration
• multi-agent system
• budget-constrained planning
• communication-free coordination
• directional sensing

Innovation

• decentralized multi-agent exploration
• communication-free coordination
• budget-aware planning
• graph attention actor
• sim-to-real transfer