دانلود رایگان مقاله روش با استفاده از پشتیبانی رگرسیون برداری برای موقعیت مکانی تلفن همراه

Abstract

Wireless positioning systems have become very popular in recent years. One of the reasons is the fact that the use of a new paradigm named Internet of Things has been increasing in the scenario of wireless communications. Since a high demand for accurate positioning in wireless networks has become more intensive, especially for location-based services, the investigation of mobile positioning using radiolocalization techniques is an open research problem. Based on this context, we propose a fingerprinting approach using support vector regression to estimate the position of a mobile terminal in cellular networks. Simulation results indicate the proposed technique has a lower error distance prediction and is less sensitive to a Rayleigh distributed noise than the fingerprinting techniques based on COST-231 and ECC-33 propagation models.

5. Conclusions

In this study, a fingerprinting location algorithm using support vector regression was proposed to estimate the position of mobile terminals in a cellular network. In addition, fingerprinting location methods with COST-231 and ECC-33 propagation models were used as reference for comparison. In all methods, mobile radio wave propagation measurements at a carrier frequency of 1.8 GHz GSM were obtained in an urban environment in the city of Recife-PE, Brazil. Some field data, like antenna-separation distance, terrain elevation, and the theoretical path loss of the Okumura-Hata model were used as input of the SVR training algorithm, while the Laplacian kernel was adopted. In spite of the increased computational cost in the model training procedure, numerical results, represented by statistical analysis, prediction maps and histograms, showed that the fingerprinting SVR-based approach had a lower error distance prediction and was less sensitive to a Rayleigh distributed noise than the other fingerprinting techniques. A work is in progress to investigate whether a combination of SVRs can improve the target positioning, not only in cellular networks, but also in vehicular networks.