Evaluating Soft Computing Techniques for Path Loss Estimation in Urban Environments

  • Authors:
  • Filippo Laganà;Matteo Cacciola;Salvatore Calcagno;Domenico De Carlo;Giuseppe Megali;Mario Versaci;Francesco Carlo Morabito

  • Affiliations:
  • University Mediterranea of Reggio Calabria, DIMET, Via Graziella, Feo di Vito, 89100 Reggio Calabria, Italy;University Mediterranea of Reggio Calabria, DIMET, Via Graziella, Feo di Vito, 89100 Reggio Calabria, Italy;University Mediterranea of Reggio Calabria, DIMET, Via Graziella, Feo di Vito, 89100 Reggio Calabria, Italy;University Mediterranea of Reggio Calabria, DIMET, Via Graziella, Feo di Vito, 89100 Reggio Calabria, Italy;University Mediterranea of Reggio Calabria, DIMET, Via Graziella, Feo di Vito, 89100 Reggio Calabria, Italy;University Mediterranea of Reggio Calabria, DIMET, Via Graziella, Feo di Vito, 89100 Reggio Calabria, Italy;University Mediterranea of Reggio Calabria, DIMET, Via Graziella, Feo di Vito, 89100 Reggio Calabria, Italy

  • Venue:
  • Proceedings of the 2009 conference on Neural Nets WIRN09: Proceedings of the 19th Italian Workshop on Neural Nets, Vietri sul Mare, Salerno, Italy, May 28--30 2009
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

Many studies have been carried out by the scientists so far, and now we have many propagation models of electromagnetic waves for various kind of building structures. Position of the buildings in the streets in urban areas, or the corridors in the office buildings, can be thought as waveguides for electromagnetic waves. For each kind of building structure, different mathematical models have been proposed and good approximations have been done by successful studies. In this context, the path loss estimation on urban environment is presented. Particularly, an urban street of Reggio Calabria, Italy, has been considered. In order to proceed for the estimation of path loss, we firstly exploited the most applied numerical methods for generating training and testing data, and subsequently we evaluated the performances of suitable Support Vector Machines in approximating the path loss values. Precisely, used numerical method are the Okumura-Hata model and the Ray-Tracing method, carried out with Wireless InSite® software. Obtained results showed that Support Vector Regression Machines (SVRMs) provide more accurate prediction of path loss in urban area than Neural Networks. Final results pointed out the possible use of Support Vector Machines in this kind of application, with interesting applications since the lower computational cost than classical numerical method.