The Feasibility Analysis of Cellphone Signal to Detect the Rain: Experimental Study

Microwave links (MLs), ranging from 10 to 30 GHz, have been widely applied for estimating rainfall; however, links below 10 GHz have rarely been applied, although they are more widespread via applications such as cellphone signals. This letter analyzes the feasibility of using cellular device like transmission signal to detect rain by establishing a 2-GHz ML and presents a detection method to classify dry/rainy periods by using statistical parameters from attenuation measurements. The detection model is trained using the C4.5 algorithm based on the combination of the average, standard deviation, minimum, and maximum of the attenuation measurements over the course of 1 min. The method is then applied for seven rain events, using a disdrometer to validate the results. The true positive rates of dry periods are all greater than 70%, and those of rainy periods are greater than 60%, indicating that the method performs well and could detect most of dry and rainy periods correctly. The results indicate the preliminary feasibility of using cellphone signals to detect rain.