A WAVELET GENETIC ALGORITHM BASED SIGNAL PATH LOSS MODEL FOR EVOLVED NODEB CELL SITES PLANNING AND EFFECTIVE COVERAGE OF WIRELESS COMMUNICATION SYSTEM
Keywords:wavelet; genetic algorithm; signal path loss; eNodeB cell sites; effective coverage wireless system
In this study, a long term evolution (LTE) path loss prediction models were developed using wavelet transform and genetic algorithm (GA) approach, based on denoised signal data (designated as wavelet-GA model) and unprocessed signal data (designated as GA model). The reference signal received power (RSRP) and distance were extracted through drive-test method in four selected evolve NodeB (eNodeB) cell sites deployed in Port Harcourt at 2.6GHz. Also, the maximum cell radius and coverage area of transmitters deployed in Port Harcourt were estimated. The developed wavelet-GA path loss model was compared with GA model, existing standard models such as COST231-Hata and Okumura-Hata path loss models, based on root mean square error (RMSE), mean absolute error (MAE), and correlation coefficient (R). Standard COST231-Hata and Okumura-Hata models were compared with the measured data for path loss estimation and it showed that the path loss values estimated using Okumura-Hata model is more closer to the measured data path loss values as such the best standard model than COST231-Hata model.
Result obtained showed the proposed wavelet-GA model as the most reliable and most suitable model to predict path loss within Port Harcourt, Nigeria. Wavelet-GA path loss model estimated the least RMSE and MAE of 2.90 dB, 3.87 dB, 4.66 dB, 7.33 dB and 2.20 dB, 2.71 dB, 3.39 dB, 6.09 dB respectively for all the sites as compared to GA model and COST231-Hata model. Also, the correlation coefficient when it was correlated with the measured data has the highest value of 90%, 63%, 79%, and 92% in all the sites. The COST231-Hata model evaluated the highest RMSE and MAE in all the sites. It showed that cell radius and coverage area evaluated to be suitable within Port Harcourt are 0.85 km and 1.39 km2 respectively. The proposed wavelet-GA path loss model has proven to be the best for cell
site planning within Port Harcourt and should be adopted for path loss evaluation during cell site planning.