Okumura-Hata Propagation Model Tuning Through Composite Function of Prediction Residual

Abstract

In this paper, an innovative composite function of prediction residual-based approach for tuning Okumura-Hata propagation model in the 800-900MHz GSM frequency band is presented. The study is based on empirical measurements conducted at University Of Uyo (UNIUYO) town-campus located at latitude and longitude of 5.042976, 7.919046 respectively. The proposed path loss tuning approach is compared with RMSE based tuning approach. According to the results, the composite function of prediction residual tuned Okumura-Hata model has the lowest RMSE value of 2.164, the highest Coefficient Of Determination (R^2) value of 0.967 and the highest prediction accuracy of 98.64%. On the other hand , the RMSE- tuned Okumura-Hata model has a higher RMSE value of 5.3, lower R^2 value of 0.814 and the lower prediction accuracy of 96.87%. Essentially, in all the three performance measures used , the composite function of prediction residual based tuning approach performed better than the RMSE based tuning approach. However, in pathloss tuning studies, RMSE value below 7dB is acceptable for the urban area. As such, the RMSE based tuning approach gave tuned model with acceptable RMSE value but with lower prediction accuracy than the model produced by the composite function of prediction residual based tuning approach.

Keywords

• Pathloss
• Residual
• Pathloss Model
• Okumura-Hata Model
• Model Tuning
• RMSE Based Tuning
• Composite Function
• Composite Function of Residual