小波支持向量机船舶物联网监测拓扑结构设计  

Topological structure design of ship IoT monitoring based on wavelet support vector machine

作　　者：李静[1] 韦志鹏 LI Jing;WEI Zhi-peng(School of Information Engineering,Kaifeng University,Kaifeng 450046,China;Personnel Department of Kaifeng University,Kaifeng 450046,China)

机构地区：[1]开封大学信息工程学院,河南开封475004 [2]开封大学人事处,河南开封475004

出　　处：《舰船科学技术》2022年第7期170-173,共4页Ship Science and Technology

基　　金：开封大学科研基金项目(KDQN-2020-GK009);开封市科技发展计划项目(2101004)。

摘　　要：在船舶物联网监测拓扑结构的设计中,由于监测方向众多导致较大通信能耗与较高网络丢包率,设计一种基于小波支持向量机的船舶物联网监测拓扑结构。通过2种数据融合算法帮助该拓扑结构实现船舶监测机能。其中对于同质传感器的数据融合,使用的处理算法是一级数据融合算法,通过二级数据融合算法实现非同质传感器之间的数据融合。基于小波支持向量机设计一种交换机数量预测模型,确定物联网监测拓扑结构中需要的交换机数量,完成拓扑结构的设计。在某船舶上测试设计拓扑结构的通信性能,测试结果为该结构的吞吐量较高,通信能耗较低,各节点的平均能耗最低可达2.05 mJ,网络丢包率也较低,证明了设计结构具有优越的通信性能。In the design of ship IoT monitoring topology, due to the large number of monitoring directions resulting in greater communication energy consumption and higher network packet loss rate, a ship IoT monitoring topology based on wavelet support vector machine was designed. Two data fusion algorithms are used to help this topology realize the ship monitoring function. Among them, for the data fusion of homogeneous sensors, the processing algorithm used is the firstlevel data fusion algorithm, and the data fusion between non-homogeneous sensors is realized through the second-level data fusion algorithm. Design a switch quantity prediction model based on wavelet support vector machine, determine the number of switches needed in the IoT monitoring topology, and complete the topology design. The communication performance of the designed topology is tested on a ship. The test results show that the throughput of the structure is high, the communication energy consumption is low, the average energy consumption of each node can be as low as 2.05 mJ, and the network packet loss rate is also low. It is proved that the designed structure has superior communication performance.

关 键 词：小波支持向量机 舰船物联网监测 分配式机制 拓扑结构 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]