Prediction of signal attenuation value caused by weather changes on cellular communication networks using backpropagation algorithm

Abstract

The value of signal attenuation by the resulting weather changes may differ at any time. The collection of signal power data with different times, weather, humidity, rainfall, and temperatures using the drive test method in Malang area will be processed using machine learning methods and backpropagation algorithms. The process is carried out using Matlab software. In this study, data collection is carried out on four BTS ranges. In addition to these data, it is also necessary to calculate the value of signal attenuation by weather changes in order to find out whether the weather category is good or bad for telecommunications activities. When the weather is sunny and cloudy it has an RSSI range value of -85 dBm to -75 dBm, while in cloudy and rainy weather it has an RSSI range of -104,2 dBm to -87 dBm. Data from the results of the drive test measurements obtained the signal attenuation value by the largest weather change of 40.49718 dB and the largest rainfall of 681.8 mm / hour. Based on the test data, the signal attenuation value when the weather is sunny and cloudy is worth 0.096164 dB to 8.61604 dB, and in cloudy and rainy weather it has a greater attenuation value, from 12.3466 dB to 21.0098 dB. Using the backpropagation algorithm, the accuracy rate in this prediction reaches 99.7 %.

Keywords

• backpropagation
• drive test
• machine learning
• rssi
• signal attenuation
• weather changes