A Proposal for Path Loss Prediction in Urban Environments using Support Vector Regression

In the last few years, the mobile data traffic has grown exponentially making evident the importance of wireless networks. To ensure an acceptable level of quality of service for users in a wireless data network, network designers rely on signal propagation path loss models. To provide adaptability, the use of machine learning techniques has been considered to predict characteristics of the wireless channel. In this work, we propose a method for predicting path loss in an urban outdoor environment using support vector regression. Simulation results indicate that, depending on the employed kernel and its parameters, the performance obtained using support vector regression is similar and with lower computational complexity to that obtained by a multilayer perceptron neural network. Keywords—wireless networks, propagation models, machine learning, nonlinear regression.