Learning Visuomotor Policy for Multi-Robot Laser Tag Game

2026-03-12

AI Summary

This work proposes an end-to-end visuomotor policy to overcome the performance bottlenecks of classical modular approaches in multi-robot laser combat, which suffer from limited observability and reliance on depth maps and inter-agent communication. The method leverages multi-agent reinforcement learning to train a teacher policy and transfers its knowledge via distillation to a student policy that operates solely on monocular image inputs. To enhance generalization and coordination under partial observability, the architecture incorporates a permutation-invariant feature extractor and a depth heatmap encoder. Experimental results demonstrate a 16.7% improvement in hit accuracy and a 6% increase in collision avoidance capability compared to baseline methods. Furthermore, the proposed policy has been successfully deployed and validated on a real-world robotic platform.

Technology Category

• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Computer Vision: Multi-modal Vision
• Machine Learning: Multimodal Learning

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems

Abstract

In this paper, we study multi robot laser tag, a simplified yet practical shooting-game-style task. Classic modular approaches on these tasks face challenges such as limited observability and reliance on depth mapping and inter robot communication. To overcome these issues, we present an end-to-end visuomotor policy that maps images directly to robot actions. We train a high performing teacher policy with multi agent reinforcement learning and distill its knowledge into a vision-based student policy. Technical designs, including a permutation-invariant feature extractor and depth heatmap input, improve performance over standard architectures. Our policy outperforms classic methods by 16.7% in hitting accuracy and 6% in collision avoidance, and is successfully deployed on real robots. Code will be released publicly.

Problem

• multi-robot
• laser tag
• limited observability
• depth mapping
• inter-robot communication

Innovation

• visuomotor policy
• knowledge distillation
• multi-agent reinforcement learning
• permutation-invariant feature extractor
• depth heatmap