基于水下云边协同架构的珊瑚礁监测新机制

珊瑚礁是极具研究价值的海洋生态系统, 水声传感器网络(underwater acoustic sensor networks, UASNs)是监测与保护珊瑚礁系统的有效手段。然而, 随着水下传感设备的大规模应用, 感知数据的类型及数量大幅增加, 传统UASNs架构将原始数据直接上传至水面数据中心的集中处理方式给网络能耗和通信效率带来严峻挑战。本文构建了一种基于边缘计算的水下端边云系统架构, 并提出一种适用于该架构的两级协同珊瑚礁系统监测机制。该架构将复杂处理任务从远程云中心分散至边缘端, 减轻了云端处理负荷。该机制由两级监测环节组成, 同时包含了端侧图像处理和端边协同数据检测策略, 实现了机器学习任务的边缘侧执行和数据的原位处理。实验结果表明, 本文研究能够明显减少网络数据流量, 有效降低网络能耗及传输时延, 显著延长网络生命周期。

A new mechanism for reef coral monitoring based on underwater cloud-edge collaborative architecture

Coral reefs are marine ecosystems of great research value, and underwater acoustic sensor networks (UASNs) are effective means to monitor and protect this system. However, with the extensive application of underwater sensing equipment, the types and quantity of sensing data have increased greatly. The traditional UASNs architecture uploads the raw data directly to the surface data center, which brings severe challenges to the network energy and communication efficiency. This paper constructs an underwater end-edge-cloud system architecture based on edge computing, and proposes a two-level collaborative coral reef system monitoring mechanism. This architecture sinks complex processing tasks from the remote cloud center to the edge, and reduces the cloud processing load. The mechanism includes end-side image processing and end-edge collaborative data detection strategies, which realizes the edge-side execution of machine learning tasks and the in-situ processing of data. Experimental results show that this study can significantly reduce network data traffic, effectively decrease energy consumption and transmission delay, and extend network life cycle.