Scholarlink

We propose a high-altitude platforms (HAP) and multiple autonomous aerial vehicles (AAVs) integrated multi-access edge computing architecture to address the limitations of traditional cloud computing in supporting low-latency, high-reliability services under 5G/6G networks. Our solution tackles AAV trajectory planning under dynamic workloads, latency assurance in computational task offloading, and privacy protection in distributed collaboration. The approach incorporates a spatio-temporal prediction model for proactive AAV trajectory planning, HAP-coordinated multi-AAV collaboration leveraging global information, personalized federated learning (PFL) for privacy-preserving knowledge sharing, and digital twin (DT) technology to enable real-time system adaptability. Performance evaluations demonstrate that, compared to state-of-the-art methods, our architecture achieves stronger privacy protection while reducing both latency and energy consumption.

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

Privacy-Preserving Task Offloading in DT-Enabled Air-Terrestrial Collaborative Edge Computing

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Junling Shi,
  • Jie Yang

摘要

We propose a high-altitude platforms (HAP) and multiple autonomous aerial vehicles (AAVs) integrated multi-access edge computing architecture to address the limitations of traditional cloud computing in supporting low-latency, high-reliability services under 5G/6G networks. Our solution tackles AAV trajectory planning under dynamic workloads, latency assurance in computational task offloading, and privacy protection in distributed collaboration. The approach incorporates a spatio-temporal prediction model for proactive AAV trajectory planning, HAP-coordinated multi-AAV collaboration leveraging global information, personalized federated learning (PFL) for privacy-preserving knowledge sharing, and digital twin (DT) technology to enable real-time system adaptability. Performance evaluations demonstrate that, compared to state-of-the-art methods, our architecture achieves stronger privacy protection while reducing both latency and energy consumption.