Track: Machine Learning

Abstract

At present, there is a growing interest in the possibility of obtaining information about the location of an object in indoor environments. Various algorithms have been developed for indoor local position systems. One of the most cost-effective choices is using the WLAN infrastructure of Wi-Fi technology based on the received signal strength indicator (RSSI). In this paper, the estimation of the indoor location of an object in a multi-floor environment using a Random forest classifier is proposed, in which RSSI is used as a metric to solve the position problem. Experimental measurements were carried out on the first floor and the second floor of the main building of Technological University (Hmawbi), Myanmar. RSSI values from both 2.4 GHz and 5 GHz frequency bands are measured and used as features to build the RSS Fingerprint database. Performance analyses are carried out and results present that using combined features from both 2.4 GHz and 5 GHz frequency bands improve the localization performance as trained random forest classifier model could increase the performance with extra features. In addition, the proposed research takes into consideration the effect of antenna orientation effect and performance improvements are highlighted with results.