Abstract
The rural and/or remote areas often face cellular network connectivity issues due to the challenges associated with infrastructure development and expansion for improving coverage and capacity, such as low population density, high infrastructure costs, complex terrain, which mobile operators are confronted with. This study addresses these challenges, particularly at Ga-Mampana (Masanteng) and parts of the N1 north highway in Limpopo, South Africa. By employing a mixed-method approach, this research explores cost-effective and energy efficient approaches for optimising cellular network coverage and capacity, concentrating on key performance metrics such as Reference Signal Received Power (RSRP), Reference Signal Received Quality (RSRQ), Channel Quality Indicator (CQI), Signal-to-Noise Ratio (SNR), Downlink (DL) and Uplink (UL) bitrates.
Leveraging Geographic Information Systems (GIS) tools such as QGIS, field measurements, and MATLAB simulations based on Longley-Rice propagation model, this research highlights most appropriate small cell placements and analyses signal propagation across varied terrains. Furthermore, a machine learning-driven optimization framework applying the Deep Deterministic Policy Gradient (DDPG) algorithm was formulated using Python programming, to dynamically adjust antenna tilt, transmit power and frequency in response to changing network demands and environmental conditions, substantially enhancing network performance while reducing energy consumption.
The research outcomes indicate that small cells deployment, combined with AI-driven parameter optimization, can significantly enhance connectivity in overlooked rural areas. Enhanced coverage and Capacity support vital services like telemedicine, remote education, working from home, and digital commerce, encouraging significant socio-economic benefits for these communities. This study also emphasizes the value of supportive policy frameworks for rural network infrastructure rollout and developments, proposing policy-driven reforms that foster public-private partnership. This study adds a practical framework for cost-effective rural network optimization and provides insightful strategies for enhancing cellular network connectivity solutions across comparable low-density regions in South Africa.