GenSwarm: Scalable Multi-Robot Code-Policy Generation and Deployment via Language Models

2025-03-31

AI Summary

Conventional multi-robot control strategy development suffers from prolonged iteration cycles, poor adaptability to dynamic tasks, and labor-intensive, manual design of objective functions. Method: This paper introduces the first zero-shot, multilingual, agent-collaborative, end-to-end language model–driven framework for robotic control. It directly synthesizes interpretable, formally verifiable, and ROS-compatible white-box control code from natural-language instructions, enabled by a lightweight edge-aware compilation and deployment architecture ensuring consistent transfer across simulation and real-world platforms (e.g., UAV and ground robot swarms). Contribution/Results: The framework achieves zero-shot policy generation across 10+ dynamic task categories, with end-to-end latency under 30 seconds and real-system deployment success rate exceeding 92%. It delivers strong interpretability, high reproducibility, and scalable engineering integration—bridging the gap between high-level intent and low-level execution in multi-robot systems.

Technology Category

• Intelligent Robots: Multi-Robot Systems
• Natural Language Processing: Code Generation / Program Synthesis from Natural Language
• Planning, Routing, and Scheduling: Planning with Language Models

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems

Abstract

The development of control policies for multi-robot systems traditionally follows a complex and labor-intensive process, often lacking the flexibility to adapt to dynamic tasks. This has motivated research on methods to automatically create control policies. However, these methods require iterative processes of manually crafting and refining objective functions, thereby prolonging the development cycle. This work introduces extit{GenSwarm}, an end-to-end system that leverages large language models to automatically generate and deploy control policies for multi-robot tasks based on simple user instructions in natural language. As a multi-language-agent system, GenSwarm achieves zero-shot learning, enabling rapid adaptation to altered or unseen tasks. The white-box nature of the code policies ensures strong reproducibility and interpretability. With its scalable software and hardware architectures, GenSwarm supports efficient policy deployment on both simulated and real-world multi-robot systems, realizing an instruction-to-execution end-to-end functionality that could prove valuable for robotics specialists and non-specialists alike.The code of the proposed GenSwarm system is available online: https://github.com/WindyLab/GenSwarm.

Problem

• Automates multi-robot control policy generation via language models
• Eliminates manual objective function crafting for faster development
• Enables zero-shot learning for adaptable task execution

Innovation

• Leverages large language models for policy generation
• Achieves zero-shot learning for rapid adaptation
• Ensures reproducibility with white-box code policies