关键词: 农田无线传感器网络, 指纹定位, ＸＧＢｏｏｓｔ, ＲＳＳＩ

Abstract: 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.