Análise de modelos de predição de perdas de propagação em redes de comunicações LTE e LTE-Advanced usando técnicas de inteligência artificial

Abstract

The perfect functionality of the 3rd and 4th generation of wireless systems requires, among other parameters, knowledge of the numerical values of the prediction of loss of propagation of propagation signals in urban, suburban and rural environments. Therefore, the study of propagation conditions in any environment will always be a concern of design engineers. The analysis and development of robust propagation loss prediction models in Long Term Evolution (LTE) and Long Term Evolution Advanced (LTE-A) communications networks using Artificial Intelligence techniques is performed in this work. The methodologies used were applied to improve the prediction of loss of empirical propagation SUI, ECC-33, Ericsson 9999, TR 36.942 models and the Free Space model applied in LTE and LTE-A networks in the frequencies of 800 MHz, 1800 MHz and 2600 MHz, for suburban environments in mid-sized cities in northeastern Brazil. Thus, in these thesis two models of Artificial Neural Networks (RNA) are proposed: (i) the neural network model with inputs based on error (RNBE) using as main feeder of the network the error between measured and simulated data, and (ii) the neural network model with land-based inputs (RNBT). The performance of these models was compared with the models of propagation considered in the work and also the versions optimized using Genetic Algorithms (AG) and the Least Square Method (LMS). Comparisons were also made with measured values, obtained from a measurement campaign carried out in the city of Natal, state of Rio Grande do Norte. The final results obtained through simulations and measurements presented good metric concordances, with emphasis on the performance of the RNBE model. Thus, the main contribution of this thesis is that, by using these techniques that make more efficient use of empirical propagation models, we can estimate realistic propagation signals, avoiding errors in the planning and implementations of LTE and LTE- A wireless networks in suburban areas.