基于常识推理与线性时序逻辑规划的机器人多目标高效搜索方法

燕洁 夏(芜湖职业技术大学，中国)
鸿芳 程(芜湖职业技术大学，中国)

摘要

在部分信息已知的环境中实现多目标的高效搜索，是机器人领域面临的一项重要挑战，其核心制约是环境仅提供地标名称、位置等基础信息。现有搜索算法存在明显局限：未能充分利用环境信息，多目标场景下搜索效率低。为此，本文提出基于常识推理的多目标反应式搜索框架，以COMET推理模型为核心，通过推断目标与已知地标的空间共现概率，结合线性时序逻辑（Linear Temporal Logic，LT L）的状态约束，构建具备动态修正能力的高效多目标搜索算法。大量仿真实验验证表明，常识推理与线性时序逻辑规划的融合，可显著提升多目标搜索效率，为复杂环境下的机器人搜索问题提供了有效解决方案。

关键词

常识推理；线性时序逻辑；多目标搜索

参考

T. Kollar and N. Roy, “Utilizing object-object and object-scene context when planning to find things,” in IEEE Int. Conf. Robot. Autom. IEEE, 2009, pp. 2168–2173.

K. Zheng, R. Chitnis, Y. Sung, G. Konidaris, and S. Tellex, “Towards optimal correlational object search,” in IEEE Int. Conf. Robot. Autom. IEEE, 2022, pp. 7313–7319

W. Ge, C. Tang, and H. Zhang, “Commonsense scene graph-based target localization for object search,” arXiv preprint arXiv:2404.00343, 2024

T. Kollar and N. Roy, “Utilizing object-object and object-scene context when planning to find things,” in IEEE Int. Conf. Robot. Autom. IEEE, 2009, pp. 2168–2173

L. Holzherr, J. Förster, M. Breyer, J. Nieto, R. Siegwart, and J. J. Chung, “Efficient multi-scale pomdps for robotic object search and delivery,” in IEEE Int. Conf. Robot. Autom. IEEE, 2021, pp. 6585– 6591.

KOLVE E, MOTTAGHI R, HAN W, et al. Ai2-thor: An interactive 3d environment for visual ai[A]. 2022. arXiv: 1712.05474.

BOSSELUT A, RASHKIN H, SAP M, et al. Comet: Commonsense transformers for automatic knowledge graph construction[A]. 2019. arXiv: 1906.05317.

WIJMANS E, KADIAN A, MORCOS A, et al. Dd-ppo: Learning near-perfect pointgoal navigators from 2.5 billion frames[A]. 2020. arXiv: 1911.00357.

CHAPLOT D S, GANDHI D, GUPTA S, et al. Learning to explore using active neural slam [A]. 2020. arXiv: 2004.05155.

XIA F, ZAMIR A R, HE Z, et al. Gibson env: Real-world perception for embodied agents [C]//Proceeding

DOI: http://dx.doi.org/10.12345/xdjyjz.v4i2.36113