Decision Support for GSM Co-location using Machine Learning
Kennedy Amadasun, James Agajo · 2022 International Conference on Electrical, Computer, Communications and Mechatronics Engineering (ICECCME) · 2022
Mobile phone communication systems is the fastest and easiest means of communication globally due to it conveniency and effectiveness. Nigeria telecommunication market is growing rapidly with a tele density 109.99 primarily intensified in built-up localities. Due to the ever-increasing number of subscribers, most mobile telecommunication operators in Nigeria are erecting new Base Transceiver Station (BTS)s, often without taking into cognisance of the current infrastructure on ground. This has led to a duplication of masts, causing unnecessary environmental pollution and operational issues. As an alternative approach, this work takes steps toward developing an AI-based decision support model for deciding the best position for siting co-located BTS for trio shared infrastructure for Global System for Mobile Communications (GSM). Five key parameters were considered for developing the AI-based model, namely Height of mast (HM), Utility cost (UC), Cell Coverage (CC), Antenna Radiation (AR) and Population density (PD). Collected data from a significant Telecommunication service provider. The data was used as a set of sequence for Artificial Neural Network (ANN) tool and to develop the hidden layers to adjust the various network weights. The regression result shows that training R was 0.999, Test R=0.99974, Validation R was 0.99954, and the overall regression result R was 0.99869, reflecting the model's good performance. Keywords-Telecommunication, Mast, Artificial Neural Network, Co-location, Transmitter