Scholarlink

<p>To address the decline in real-time tracking performance of moving targets in multi-UAV systems caused by uneven search coverage and low cooperation efficiency, the WGWO (Whale-Grey Wolf Optimization) model is proposed. The UAV swarm optimizes its flight paths using the spiral predation strategy of the WOA (Whale Optimization Algorithm) while employing a Kalman filter to process sensor data. Target information is aggregated through a wireless network, and a clustering-based routing protocol optimizes data transmission. GWO (grey wolf optimizer) further focuses on high-density regions for local search and real-time target position updates. Finally, a joint target detection model based on ConvNeXt and RetinaNet dynamically adjusts UAV positions to achieve high-precision continuous tracking. Experimental results show that compared with mainstream models such as ConvNeXt, ResNet, EfficientNet, MobileNetV3, and DenseNet, WGWO improves target recognition accuracy to 0.92 and reduces the target loss rate to only 2.35% for uniformly moving targets. This algorithm significantly enhances the real-time tracking capability of UAV swarms, achieving superior performance in tracking moving targets.</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

Bio-inspired swarm intelligence for enhanced real-time aerial tracking: integrating whale optimization and grey wolf optimizer algorithms

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年2月27日

查看全文
  • GaoFeng Han

摘要

To address the decline in real-time tracking performance of moving targets in multi-UAV systems caused by uneven search coverage and low cooperation efficiency, the WGWO (Whale-Grey Wolf Optimization) model is proposed. The UAV swarm optimizes its flight paths using the spiral predation strategy of the WOA (Whale Optimization Algorithm) while employing a Kalman filter to process sensor data. Target information is aggregated through a wireless network, and a clustering-based routing protocol optimizes data transmission. GWO (grey wolf optimizer) further focuses on high-density regions for local search and real-time target position updates. Finally, a joint target detection model based on ConvNeXt and RetinaNet dynamically adjusts UAV positions to achieve high-precision continuous tracking. Experimental results show that compared with mainstream models such as ConvNeXt, ResNet, EfficientNet, MobileNetV3, and DenseNet, WGWO improves target recognition accuracy to 0.92 and reduces the target loss rate to only 2.35% for uniformly moving targets. This algorithm significantly enhances the real-time tracking capability of UAV swarms, achieving superior performance in tracking moving targets.