Scholarlink

<p>Unmanned aerial vehicles (UAVs), due to their low cost, have found extensive applications in maritime search and rescue. However, challenges such as inaccurate path planning and low recognition accuracy persist in UAV-based maritime search and rescue operations. Therefore, this paper aims to enhance the efficiency and success rate of UAV search and rescue missions, focusing on collaborative path planning and visual perception. This paper proposes an improved diversity-based ant colony algorithm in collaborative path planning for multiple UAVs. Experimental results indicate that the enhanced ant colony algorithm outperforms traditional algorithms in terms of both mean and variance. This improvement translates to more vital search capabilities, enhanced cruising performance, and higher efficiency. Concerning visual perception, the paper incorporates a convolutional attention mechanism into the YOLOv8 network structure, thereby improving its performance. The proposed enhancement method significantly boosts detection performance, with a 19.1% increase in accuracy, a 22.8% improvement in average precision, and a reduction in the loss value from 1.45 to 0.99. These results demonstrate that the enhanced model exhibits superior detection capabilities. The methods presented in this paper significantly improve the efficiency and success rate of UAV search and rescue missions.</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

Multi-UAV path planning based on IACO and improved YOLOV8 perception for maritime search and rescue

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年11月18日

查看全文
  • Shuwei Sun,
  • Hao Zhang,
  • Enchong Dong

摘要

Unmanned aerial vehicles (UAVs), due to their low cost, have found extensive applications in maritime search and rescue. However, challenges such as inaccurate path planning and low recognition accuracy persist in UAV-based maritime search and rescue operations. Therefore, this paper aims to enhance the efficiency and success rate of UAV search and rescue missions, focusing on collaborative path planning and visual perception. This paper proposes an improved diversity-based ant colony algorithm in collaborative path planning for multiple UAVs. Experimental results indicate that the enhanced ant colony algorithm outperforms traditional algorithms in terms of both mean and variance. This improvement translates to more vital search capabilities, enhanced cruising performance, and higher efficiency. Concerning visual perception, the paper incorporates a convolutional attention mechanism into the YOLOv8 network structure, thereby improving its performance. The proposed enhancement method significantly boosts detection performance, with a 19.1% increase in accuracy, a 22.8% improvement in average precision, and a reduction in the loss value from 1.45 to 0.99. These results demonstrate that the enhanced model exhibits superior detection capabilities. The methods presented in this paper significantly improve the efficiency and success rate of UAV search and rescue missions.