One-to-all regularized logistic regression-based classification for WiFi indoor localization

Abstract

Wi-Fi based indoor localization is gaining popularity because of the wide adoption of WiFi technologies in existing infrastructure. In order to increase the accuracy of Wi-Fi localization, we develop a novel localization method using One-to-all Regularized Logistic Regression-based Classification (ORLRC). This method is based on logistic regression. The proposed ORLRC is compared with the k-means clustering approach and achieves a location estimation accuracy of 95.8% comparing to an accuracy of 80% by the k-means clustering approach.