arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2602.02530cs.LGcs.RO
收藏

通过离策略评估制定人机协作的强化学习

Formulating Reinforcement Learning for Human-Robot Collaboration through Off-Policy Evaluation

Saurav Singh, Rodney Sanchez, Alexander Ororbia, Jamison Heard

AI总结:

本文提出通过离策略评估方法,利用记录数据自动选择状态空间和奖励函数,提升现实环境中人机协作的强化学习可行性与可扩展性。

摘要:

强化学习(RL)有潜力通过使自主代理从经验中学习来改变现实世界的决策系统。在现实环境中部署RL,尤其是在人机交互的背景下,需要定义状态表示和奖励函数,这对于学习效率和策略性能至关重要。传统RL方法往往依赖领域专业知识和试错,需要大量的人类参与以及与环境的直接交互,这在复杂和安全关键的应用中可能成本高昂且不切实际。本文提出了一种新的RL框架,利用离策略评估(OPE)来选择状态空间和奖励函数,仅使用记录的交互数据。这种方法消除了对实时访问环境或人类在环反馈的需要,大大减少了对昂贵实时交互的依赖。所提出的方法通过训练离线RL代理并应用OPE来估计策略性能,系统地评估多个候选状态表示和奖励函数。最优的状态空间和奖励函数是根据其在OPE指标下产生高性能策略的能力进行选择的。我们的方法在两个环境中得到验证:一个是OpenAI Gym的Lunar Lander环境,它为评估状态空间和奖励函数选择提供了受控设置;另一个是NASA-MATB-II人 subjects 研究环境,它评估了该方法在人机协作场景中的实际应用。这项工作通过数据驱动的OPE基于评估自动化关键的RL设计决策,提高了离线RL在现实环境中的可行性和可扩展性,从而为复杂的人机交互设置提供了更加可靠、有效和可持续的RL制定方法。

英文摘要:

Reinforcement learning (RL) has the potential to transform real-world decision-making systems by enabling autonomous agents to learn from experience. Deploying RL in real-world settings, especially in the context of human-robot interaction, requires defining state representations and reward functions, which are critical for learning efficiency and policy performance. Traditional RL approaches often rely on domain expertise and trial-and-error, necessitating extensive human involvement as well as direct interaction with the environment, which can be costly and impractical, especially in complex and safety-critical applications. This work proposes a novel RL framework that leverages off-policy evaluation (OPE) for state space and reward function selection, using only logged interaction data. This approach eliminates the need for real-time access to the environment or human-in-the-loop feedback, greatly reducing the dependency on costly real-time interactions. The proposed approach systematically evaluates multiple candidate state representations and reward functions by training offline RL agents and applying OPE to estimate policy performance. The optimal state space and reward function are selected based on their ability to produce high-performing policies under OPE metrics. Our method is validated on two environments: the Lunar Lander environment by OpenAI Gym, which provides a controlled setting for assessing state space and reward function selection, and a NASA-MATB-II human subjects study environment, which evaluates the approach's real-world applicability to human-robot teaming scenarios. This work enhances the feasibility and scalability of offline RL for real-world environments by automating critical RL design decisions through a data-driven OPE-based evaluation, enabling more reliable, effective, and sustainable RL formulation for complex human-robot interaction settings.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑