Scholarlink

<p>Autonomous underwater exploration using Autonomous Mobile Robots (AMRs) is increasingly important as marine resource exploitation intensifies. However, limited computing power and the inherent complexity of unfamiliar environments often causes existing exploration frameworks to produce hazardous maneuvers or mission failures resulting from untimely planning. In this paper, we address bottlenecks in viewpoint generation and multi-goal path planning to improve the real-time performance of AMRs during exploration. We propose an exploration framework that integrates (1) an FOV-based viewpoint generation method to accelerate frontier clustering and viewpoint sampling, (2) a sparse graph construction technique guided by frontier clusters to reduce unnecessary node generation and minimize the search space, and (3) a two-stage multi-goal path planning algorithm that combines heuristic strategies for local rapid decision-making with Traveling Salesman Problem-based methods to balance real-time responsiveness with overall exploration performance. Experimental evaluations indicate that our framework reduces computation time by over 90% compared with state-of-the-art techniques while maintaining exploration performance comparable. The proposed framework significantly enhances the real-time capabilities of AMRs, thereby facilitating efficient and safe underwater exploration in complex environments. This work provides a robust and efficient solution for rapid underwater coverage and mapping tasks, while also being applicable to challenging aerial environments.</p>

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

Real-time autonomous underwater and aerial exploration with limited FOV sensors

  • 期刊论文

  • 发表日期：

    2025年5月23日

查看全文
  • Haiyu Huang,
  • Shu Zhang,
  • Hao Fan,
  • Ting Wang,
  • Yanguo Jing,
  • Junyu Dong

摘要

Autonomous underwater exploration using Autonomous Mobile Robots (AMRs) is increasingly important as marine resource exploitation intensifies. However, limited computing power and the inherent complexity of unfamiliar environments often causes existing exploration frameworks to produce hazardous maneuvers or mission failures resulting from untimely planning. In this paper, we address bottlenecks in viewpoint generation and multi-goal path planning to improve the real-time performance of AMRs during exploration. We propose an exploration framework that integrates (1) an FOV-based viewpoint generation method to accelerate frontier clustering and viewpoint sampling, (2) a sparse graph construction technique guided by frontier clusters to reduce unnecessary node generation and minimize the search space, and (3) a two-stage multi-goal path planning algorithm that combines heuristic strategies for local rapid decision-making with Traveling Salesman Problem-based methods to balance real-time responsiveness with overall exploration performance. Experimental evaluations indicate that our framework reduces computation time by over 90% compared with state-of-the-art techniques while maintaining exploration performance comparable. The proposed framework significantly enhances the real-time capabilities of AMRs, thereby facilitating efficient and safe underwater exploration in complex environments. This work provides a robust and efficient solution for rapid underwater coverage and mapping tasks, while also being applicable to challenging aerial environments.