用大模型自动设计机器人导航奖励函数，提升智能体表现。
EvoNav: Evolutionary Reward Function Design for Robot Navigation with Large Language Models
- 通过大模型生成奖励函数，再用渐进式评估筛选最优方案。
- 在真实环境测试中，导航成功率显著高于人工设计与现有方法。
- 适合需要快速迭代奖励策略的机器人研发团队使用。
机器人导航在动态人类环境中具有重要应用价值。尽管强化学习（RL）在此任务中展现出巨大潜力，但策略性能高度依赖奖励函数的设计。手工设计的奖励函数需大量领域知识，且嵌入难以审计或调整的先验偏见，限制了其有效性并导致次优表现。本文提出EvoNav，一种基于大语言模型（LLM）的进化框架，用于自动化机器人导航奖励函数设计。为克服策略训练成本过高的问题，EvoNav采用三阶段渐进式预热-增强评估流程：从低成本代理（如小数据集和解析规则）到轻量级回放，最终进入完整策略训练，实现高效探索与有效反馈。实验结果表明，EvoNav生成的导航策略优于人工设计的强化学习奖励及当前最先进的奖励设计方法。
原文摘要 · Abstract (English)
Robot navigation is a crucial task with applications to social robots in dynamic human environments. While Reinforcement Learning (RL) has shown great promise for this problem, the policy quality is highly sensitive to the specification of reward functions. Hand-crafted rewards require substantial domain expertise and embed inductive biases that are difficult to audit or adapt, limiting their effectiveness and leading to suboptimal performance. In this paper, we propose EvoNav, an evolutionary framework that automates the design of robot navigation reward functions via large language models (LLMs). To overcome prohibitively costly policy training, EvoNav evaluates each candidate proposal from the LLM via a progressive three-stage warm-up-boost procedure. EvoNav advances from analytical proxies with low-cost surrogates, such as small datasets and analytic rules, to lightweight rollouts and, finally, to full policy training, enabling computationally efficient exploration under effective feedback. Experiment results show that EvoNav produces more effective navigation policies than manually designed RL rewards and state-of-the-art reward design methods.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。