期刊文献+
共找到2篇文章
< 1 >
每页显示 20 50 100
基于地下无线传播信号的土壤含水率预测方法
1
作者 徐兴 刘国杰 +4 位作者 李碧芳 段洁利 杨洲 付函 金莫辉 《农业机械学报》 北大核心 2026年第4期347-354,共8页
及时准确地获取土壤水分信息对全面掌握土壤含水率、提高用水效率、节约宝贵的水资源具有重要意义。当前,土壤水分信息远程采集需同时使用土壤水分传感器及无线传输模块,但局限于离散点的数据获取,不能实现对区域尺度全面评估。而且土... 及时准确地获取土壤水分信息对全面掌握土壤含水率、提高用水效率、节约宝贵的水资源具有重要意义。当前,土壤水分信息远程采集需同时使用土壤水分传感器及无线传输模块,但局限于离散点的数据获取,不能实现对区域尺度全面评估。而且土壤水分传感器成本较高,往往无法进行批量部署。本文利用地下无线信号受土壤水分显著影响的工作特性及其无线传输通信优势,提出了一种低成本的基于地下无线传播信号对区域范围实现土壤含水率预测方法。基于LoRa技术和泛化能力强的快速机器学习方法实现,首先设计并制作了可用于地下环境工作的LoRa无线地下传感器网络,实现无线信号接收信号强度(Received signal strength indicator,RSSI)、信噪比以及土壤温度等数据的远程采集。接着,利用采集的数据,提出了基于多源数据的核极限学习机(Kernel extreme learning machine,KELM)模型和粒子群参数寻优计算土壤含水率,解决了土壤含水率与多源数据间的非线性关系复杂度高、大规模数据难以拟合、训练难收敛问题;试验结果表明,提出的基于粒子群参数寻优的KELM模型相对于极限学习机(Extreme learning machine,ELM)、支持向量机(Support vector regression,SVR)、长短记忆神经网络(Long short-term memory,LSTM)具有更高的土壤含水率反演精度。在训练集上,KELM模型平均绝对误差(MAE)为0.76%,均方根误差(RMSE)为1.02%,在测试集上,MAE为0.72%,RMSE为1.07%,能更准确地得到土壤含水率。提出的方法可在不依赖密集布设土壤水分传感器情况下实现土壤含水率有效预测和远程采集,为远程土壤水分区域探测提供了更低成本的解决方案。 展开更多
关键词 土壤含水率 预测方法 WUSN 预测模型
在线阅读 下载PDF
Interference of human's ankle to the channel characteristics of near surface WUSN
2
作者 LI Li WU Wei +1 位作者 CHEN Jian-ya LIU Yun-jie 《The Journal of China Universities of Posts and Telecommunications》 EI CSCD 2014年第1期36-42,共7页
Because of its potential applications in agriculture, environment monitoring and so on, wireless underground sensor network(WUSN) has been researched more and more extensively in recent years. The main and most impo... Because of its potential applications in agriculture, environment monitoring and so on, wireless underground sensor network(WUSN) has been researched more and more extensively in recent years. The main and most important difference of WUSN to terrestrial wireless sensor network(WSN) is the channel characteristics, which determines the design methodology of it. In this paper, the propagation character of electromagnetic(EM) wave in the near surface WUSN is analyzed, as well as the path loss model of it is given. In addition, the influence of human's ankle to the channel characteristics of near surface WUSN is investigated by electromagnetic theory analysis, simulation and experiment. A novel path loss model of near surface WUSN which takes the interference of human's ankle into consideration is proposed. It is verified that the existing of human above the WUSN system may cause additional attenuation to the signal of near surface WUSN which propagates as lateral wave along the ground. Moreover, the relation of the attenuation and operating frequency is deduced, which gives a reference to extend the frequency band applied in WUSN. 展开更多
关键词 WUSN lateral wave human tissue path loss model EM wave attenuation
原文传递
上一页 1 下一页 到第
使用帮助 返回顶部