Scholarlink

This paper presents a novel topology-driven method for planning recovery paths amidst dynamic obstacles. Our framework integrates Discrete Morse Theory principles from prior research and expands to provide congestion-aware solutions for mobile robot and manipulator alike. To enhance efficiency and reliability, our algorithm balances local segment deflection and topologically diverse paths, aiming to reduce dynamic obstacle interference and optimize path length. We introduce a new approach to path deflection that efficiently navigates around dynamic obstacles while avoiding congested areas. Our methodology is validated through comprehensive simulations, demonstrating its improvement over existing models in terms of efficiency, adaptability, and scalability. We perform experiments in 2D and 3D environments and with 3–14 degrees of freedom robots. Results show comparable path length and improved computation time for a varying number of dynamic obstacles even when compared with existing method.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Topology-Driven Recovery Path Planning in Dynamic Obstacle Environments

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Mukulika Ghosh,
  • Aakriti Upadhyay,
  • Chinwe Ekenna

摘要

This paper presents a novel topology-driven method for planning recovery paths amidst dynamic obstacles. Our framework integrates Discrete Morse Theory principles from prior research and expands to provide congestion-aware solutions for mobile robot and manipulator alike. To enhance efficiency and reliability, our algorithm balances local segment deflection and topologically diverse paths, aiming to reduce dynamic obstacle interference and optimize path length. We introduce a new approach to path deflection that efficiently navigates around dynamic obstacles while avoiding congested areas. Our methodology is validated through comprehensive simulations, demonstrating its improvement over existing models in terms of efficiency, adaptability, and scalability. We perform experiments in 2D and 3D environments and with 3–14 degrees of freedom robots. Results show comparable path length and improved computation time for a varying number of dynamic obstacles even when compared with existing method.