Enhancing Path Loss Predictions for 2.6GHz 4G LTE in Urban Areas: A Case Study of Ibadan

Accurate path loss estimation is crucial for the design and performance optimization of mobile communication networks. However, existing empirical models often fail to deliver reliable predictions across diverse environments. This study focuses on developing an optimized path loss prediction model for 2.6 GHz 4G Long Term Evolution (LTE) networks in urban Ibadan, Nigeria. Three widely used empirical models COST-231 Hata, Stanford University Interim (SUI), and ECC-33 were evaluated by comparing their predicted values with real-world measurement data. Performance metrics, including Root Mean Square Error (RMSE), Mean Absolute Error (MAE), and Standard Deviation (SD), identified the ECC-33 model as the most accurate base model, yielding an RMSE of 6.68dB, MAE of 4.89dB, and SD of 5.80dB. An improved ECC-33 model was developed by training a linear regression with 70% of the collected field data to obtain optimised ECC-33 parameters. When validated with the remaining 30%, it demonstrates an enhanced accuracy, achieving an RMSE of 2.88dB, MAE of 2.46dB, and SD of 3.20dB. This optimized model can be a reliable tool for planning and optimizing 4G LTE networks in urban Ibadan, thus improving voice and data service quality.