arXiv：cs.LG· Isaac Peterson, Christopher Allred, Jacob Morrey, Mario Harper·· 7 小时前AI 评分33
HARL-A：面向 IsaacLab 异构多智能体对抗强化学习的可扩展基准框架
HARL-A: An Extensible Benchmark Framework for Heterogeneous Multi-Agent Adversarial Reinforcement Learning in IsaacLab
AI 导读
研究者发布开源框架 HARL-A，基于 IsaacLab 支持异构机器人团队的多智能体对抗强化学习训练与基准测试，可支持任意数量队伍及每队任意机器人形态组合。该框架扩展了 HARL 算法库，提供模块化架构、Sumo/Soccer/3D Galaga 三套基准环境和十多个预训练策略，已公开于 Hugging Face。代码、环境和文档均已开源。
正文
Abstract:Progress in adversarial multi-agent reinforcement learning (MARL) for robotics has been hampered by a lack of shared, extensible infrastructure that supports heterogeneous agent morphologies in high-fidelity physics simulation. Existing frameworks either focus on cooperative tasks, rely on simplified physics engines, or provide isolated implementations that are difficult to extend. We present HARL-A, an open-source, actively maintained framework built on IsaacLab that enables scalable training and benchmarking of adversarial policies across morphologically diverse robot teams with any number of teams and any mix of robot morphologies per team. HARL-A extends the HARL algorithm library and IsaacLab with adversarial multi-agent support and contributes three components: (1) a modular software architecture that reduces the engineering overhead of defining new heterogeneous adversarial environments, (2) a suite of three benchmark environments---Sumo, Soccer, and 3D Galaga---spanning contact-rich pushing, ball-skill competition, and pursuit/evasion, (3) over ten pretrained policies spanning homogeneous and heterogeneous team configurations, released publicly on Hugging Face to enable immediate exploration of adversarial learning dynamics without retraining from scratch. We demonstrate the framework across multiple competitive scenarios, showing that it reliably produces learned adversarial policies and emergent role specialization. All code environments, trained policies, and documentation are openly available at this https URL.
| Comments: | 8 page, 9 figures, code this https URL |
| Subjects: | Machine Learning (cs.LG); Robotics (cs.RO) |
| MSC classes: | 68T42 (Primary) 68T40, 68T05 (Secondary) |
| Cite as: | arXiv:2510.01264 [cs.LG] |
| (or arXiv:2510.01264v2 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2510.01264 arXiv-issued DOI via DataCite |
Submission history
From: Isaac Peterson [view email]
[v1]
Fri, 26 Sep 2025 03:16:48 UTC (4,375 KB)
[v2]
Mon, 5 Oct 2026 18:10:49 UTC (4,471 KB)
来源：arXiv：cs.LG · arxiv.org