iCORPP：机器人上的交错常识推理与概率规划
iCORPP: Interleaved Commonsense Reasoning and Probabilistic Planning on Robots
AI总结：
本文提出 iCORPP 算法，将常识推理与概率规划交错耦合，使机器人能同时理解世界状态、主动收集信息并完成长期服务任务。
摘要：
现实世界中的机器人序贯决策是一项挑战，因为它要求机器人同时推理当前世界状态和动态，并规划行动以完成复杂任务。一方面，声明性语言和推理算法很好地支持了常识知识的表示与推理，但这些算法不擅长在漫长且未指定的时域内规划行动以最大化累积奖励。另一方面，概率规划框架，如 Markov decision processes（MDP，马尔可夫决策过程）和 partially observable MDP（POMDP，部分可观测马尔可夫决策过程），很好地支持在不确定性下为实现长期目标进行规划，但它们在表示或推理与行动没有直接关系的知识方面能力不足。在本文中，我们提出一种名为 iCORPP 的新型算法，用于同时估计当前世界状态、推理世界动态并构建面向任务的控制器。在此过程中，机器人决策问题被分解为两个相互依赖的（较小的）子问题，分别专注于“理解世界”的推理和“实现目标”的规划。上下文知识在推理组件中表示，这使规划组件具备认知性，并能够主动收集信息。所开发算法已在仿真和真实机器人上使用室内导航、对话管理和物品递送等日常服务任务实现并评估。结果表明，与包括手工行动策略在内的强基线相比，该算法在可扩展性、效率和适应性方面均有显著提升。
英文摘要：
Robot sequential decision-making in the real world is a challenge because it requires the robots to simultaneously reason about the current world state and dynamics, while planning actions to accomplish complex tasks. On the one hand, declarative languages and reasoning algorithms well support representing and reasoning with commonsense knowledge. But these algorithms are not good at planning actions toward maximizing cumulative reward over a long, unspecified horizon. On the other hand, probabilistic planning frameworks, such as Markov decision processes (MDPs) and partially observable MDPs (POMDPs), well support planning to achieve long-term goals under uncertainty. But they are ill-equipped to represent or reason about knowledge that is not directly related to actions. In this article, we present a novel algorithm, called iCORPP, to simultaneously estimate the current world state, reason about world dynamics, and construct task-oriented controllers. In this process, robot decision-making problems are decomposed into two interdependent (smaller) subproblems that focus on reasoning to "understand the world" and planning to "achieve the goal" respectively. Contextual knowledge is represented in the reasoning component, which makes the planning component epistemic and enables active information gathering. The developed algorithm has been implemented and evaluated both in simulation and on real robots using everyday service tasks, such as indoor navigation, dialog management, and object delivery. Results show significant improvements in scalability, efficiency, and adaptiveness, compared to competitive baselines including handcrafted action policies.
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。
