An Effective Approach for Prediction of Exposure to Base Stations Using Wavelets to Fit the Antenna Pattern

Abstract

The antenna pattern representing the gain variation in space is a crucial factor for the prediction of exposure to the electric field emitted from antennas of base stations. In this communication, we propose an effective approach to fit the antenna pattern by using wavelets, which can conduct certain transformation for the antenna pattern in prediction models, to match the gradual change of the real exposure with respect to distance. We study the choices of the wavelet families and the support width, and investigate the performances of the prediction model at different scales. We also compare the performances of the prediction model using wavelets, traditional fitting functions (the modified Gaussian function and the cosine functions), and the original antenna pattern directly. Finally, we present a specific suggestion about using wavelets to fit the antenna pattern for the prediction of exposure. The results show that the model using wavelets can obtain a high correlation of 0.97 between the modeled and measured exposure, much greater than the previously reported value of 0.76.