基于神经网络预测的无线传感器网络田间射频信号路径损耗

为解决应用无线传感器网络技术监测农田信息时无法快速预测射频信号路径损耗的问题，基于神经网络理论研究了田间路径损耗与其影响因素间的关系。试验中选取915和2 470 MHz 2个载波频率，在冬小麦的不同生长阶段测量射频信号在田间各影响因素作用下的路径损耗，建立和验证基于神经网络的射频信号田间路径损耗预测模型。所建立模型模拟值与实测值的相关系数为0.92，应用建立的神经网络预测田间射频信号路径损耗并与实测值对比，最大预测误差绝对值为4.186 dB，最大预测标准差为2.759 dB，预测准确度为94.2%。所建立的BP网络可以对田间射频信号路径损耗进行预测。

Path-loss prediction for radio frequency signal of wireless sensor network in field based on artificial neural network

For solving the problem that path-loss of radio frequency signal could not be easily retrieved on the process of wireless sensor network (WSN) applications in field, the relationships between path-loss of WSN radio frequency (RF) signal in field and its impact factors were studied based on artificial neural network (ANN) theory. Two carrier frequencies, 915 MHz and 2 470 MHz, were selected. Path-loss prediction ANN model of WSN RF signal in field was achieved through measuring RF path-loss under the two carrier frequencies with different combinations of impact factors at different winter wheat growth stages. Correlation coefficient of the model was 0.92, by comparing the path-loss measured with predicted values, it was verified that the highest absolute prediction error was 4.186 dB, the highest prediction standard deviation was 2.759 dB and prediction accuracy was 94.2%. The designed BP ANN is suitable for path-loss prediction of the radio frequency signal in field.