摘要
针对农田无线传感器网络信号易受开放环境影响、节点定位存在区域空缺导致定位精度低的问题,提出一种基于指纹定位的高精度农田节点定位算法。通过采集农田网络信号数据,构建基于信号、坐标双尺度的K-means聚类指纹;针对开放农田环境因素造成的数据波动,利用XGBoost算法建立接收信号强度(RSSI)与节点位置之间的非线性映射模型,通过接收的信号强度指纹匹配粗定位节点位置;根据待测点与粗定位区域距离关系,进一步引入加权几何优化算法二次定位待测点,缩进定位区域实现节点精确定位。实验结果表明:在80%分位处,研究算法定位误差小于1.05 m,较对比算法定位精度更高、时效性更强,能够为农田无线传感器网络定位系统提供参考。
Aiming at the problem that the signal of farmland wireless sensor network(WSN)is susceptible to complex environment and the regional vacancies existing in node localization leads to low localization accuracy,a high-precision farmland node positioning algorithm based on fingerprint positioning is proposed.By collecting farmland network signal data,a K-means clustering fingerprint based on signal and coordinate double ruler is constructed;for data fluctuations caused by environmental factors in open farmland,a nonlinear mapping model between received signal strength(RSSI)and node location is established by XGBoost algorithm,and the received signal strength fingerprint matches the coarse localization node location;according to the distance relationship between the point to be measured and the coarse localization area,the weighted geometric optimization algorithm is introduced to locate the point to be measured twice and indent the positioning area to achieve precise positioning of the node.The experimental results show that the localization error of the research algorithm is less than 1.05 m at 80%quantile,which is more accurate and time-efficient than the comparison algorithm.The algorithm can provide a high-precision basic model for the farmland wireless sensor network positioning system.
作者
臧英凯
韩笑
陈金超
陈雯柏
吴华瑞
赵春江
ZANG Yingkai;HAN Xiao;CHEN Jinchao;CHEN Wenbai;WU Huarui;ZHAO Chunjiang(School of Automation,Beijing Information Science and Technology University,Beijing 100192,China;National Engineering Research Center for Information Technology in Agriculture,Beijing 100097,China;Key Laboratory of Digital Village Technology,Ministry of Agriculture and Rural Affairs,Beijing 100097,China;Nongxin Technology Limited,Beijing 100097,China)
出处
《重庆理工大学学报（自然科学）》
北大核心
2023年第10期1-8,共8页
Journal of Chongqing University of Technology：Natural Science
基金
科技创新2030——“新一代人工智能”重大项目(2021ZD0113605)
河北省重点研发计划项目(21327410D)
财政部和农业农村部:国家现代农业产业技术体系项目(CARS-23-D07)。
