用受力反馈感知植被阻力，让机器人安全穿行杂草丛
Feeling the Force: A Nuanced Physics-based Traversability Sensor for Navigation in Unstructured Vegetation
- 通过直接测量植物反推力判断可通行性
- 实验证明能捕捉细微力变化，精度高
- 适合野外机器人导航与智能学习算法开发
在众多应用中，机器人正越来越多地部署于非结构化自然环境中，面临各种类型的植被。植被作为可通行障碍物具有独特挑战：植物的力学特性会影响机器人是否能安全碰撞并克服障碍。传统方法难以评估其安全性与可通行性，而碰撞有时是安全且必要的。本文提出一种新型传感器，可直接测量植被作用于机器人的反推力，从而全面理解机器人与环境的交互过程。实验验证表明，该传感器能有效捕捉细微的力变化，提供可量化的通行决策依据，并为未来基于学习的导航算法奠定基础。
原文摘要 · Abstract (English)
In many applications, robots are increasingly deployed in unstructured and natural environments where they encounter various types of vegetation. Vegetation presents unique challenges as a traversable obstacle, where the mechanical properties of the plants can influence whether a robot can safely collide with and overcome the obstacle. A more nuanced approach is required to assess the safety and traversability of these obstacles, as collisions can sometimes be safe and necessary for navigating through dense or unavoidable vegetation. This paper introduces a novel sensor designed to directly measure the applied forces exerted by vegetation on a robot: by directly capturing the push-back forces, our sensor provides a detailed understanding of the interactions between the robot and its surroundings. We demonstrate the sensor's effectiveness through experimental validations, showcasing its ability to measure subtle force variations. This force-based approach provides a quantifiable metric that can inform navigation decisions and serve as a foundation for developing future learning algorithms.
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。