An RFID Indoor Positioning Algorithm Based on Support Vector Regression

TLDR

Indoor positioning is essential for location-based services, but GPS fails indoors, prompting research into WiFi, Bluetooth, UWB, and RFID; RFID is favored for its low cost and accuracy, yet the LANDMARC algorithm’s precision depends on tag density and reader performance. The paper seeks to improve LANDMARC positioning precision by incorporating weighted path length and support vector regression. The authors augment LANDMARC with a weighted path length metric and train a support vector regression model to predict positions from RFID signal data. Experimental results demonstrate that the proposed method improves positioning accuracy.

Abstract

Nowadays, location-based services, which include services to identify the location of a person or an object, have many uses in social life. Though traditional GPS positioning can provide high quality positioning services in outdoor environments, due to the shielding of buildings and the interference of indoor environments, researchers and enterprises have paid more attention to how to perform high precision indoor positioning. There are many indoor positioning technologies, such as WiFi, Bluetooth, UWB and RFID. RFID positioning technology is favored by researchers because of its lower cost and higher accuracy. One of the methods that is applied to indoor positioning is the LANDMARC algorithm, which uses RFID tags and readers to implement an Indoor Positioning System (IPS). However, the accuracy of the LANDMARC positioning algorithm relies on the density of reference tags and the performance of RFID readers. In this paper, we introduce the weighted path length and support vector regression algorithm to improve the positioning precision of LANDMARC. The results show that the proposed algorithm is effective.