Using Virtual Reference Tags to Improve Accuracy of Active RFID-Based Positioning Systems Saber Seyyedi Valieasr Hospital Rasht, Iran ss.seyyedi@gmail.com Behzad Akbari Tarbiat Modares University Tehran, Iran b.akbari@modares.ac.ir Esmaeil Arab Valieasr Hospital Rasht, Iran dr.e_arab@hotmail.com Iraj Ramezani Valieasr Hospital Rasht, Iran irajramezani83534@yahoo.com Milad Mahdavi Valieasr Hospital Rasht, Iran sir.milad.mahdavi@gmail.com Abstract— Up to now, different indoor positioning algorithms based on Active RFID technology have been presented. These algorithms use the concept of reference tags to locate tracking tags that are attached to objects. Experimental results have shown that the increase of reference tags caused the increase of positioning accuracy. However if the number of these reference tags exceeds a limit two problems can arise. The first problem is the increase of multipath propagation and signal interference. The second problem is the increase of cost. In order to improve the performance of indoor environment this paper presents a new method to estimate the tracking tags’ location based on virtual reference tags instead of increasing the number of real reference tags. In this method after calculating RSSI of each virtual reference tag by its surrounding real reference tags RSSI, sensing area is divided into several sections and tracking tags location is calculated by real or virtual reference tags which are in section that the objects has the highest probability to be enclosed within. Experimental results show that our purposed algorithm considerably reduced error estimation of previous algorithms. In comparison with LANDMARC and VIRE approach, we reduced average estimation error by about 70 percent and 10 percent respectively. Keywords- Active RFID, positioning algorithm, virtual reference tag, tracking tag, RSSI, Euclidian distance. I. INTRODUCTION Active RFID systems have been able to make a good place in industry and commerce in the past decade [1]. One of the most important applications of these systems is positioning of patients, personals and medical equipments in hospitals and healthcare systems [2,3]. For this purpose, each tracking object is equipped with an active RFID tag and its location is calculated by RSS method. However, because of various environmental factors such as signal interference and multipath propagation, positioning accuracy is changed in different times. So that different places may be calculated even for a fixed object in different times [4]. A way to conquest this problem is using reference tags with fixed place that are compatible with dynamism of indoor environment and increasing localization accuracy [5]. For the first time, LANDMARC algorithm used this method to determine tracking tags’ position with an acceptable accuracy without using extra readers, as k reference tags which have least difference of RSSI with tracking tag [5,6] chosen to calculate its coordination. Experimental results have shown that the increase of the number of reference tags caused the increase of positioning accuracy [5]. But if the number of these reference tags exceeds a limit, due to various problems such as increase of multipath propagation and signal interference, accuracy will decrease [5]. Therefore, the VIRE algorithm was presented to abate the multipath propagation and signal interference problem caused by the increase of reference tags. Accordingly, a proximity map was maintained by each reader and an elimination algorithm was used to eliminate those unlikely locations to reduce the estimation error [6]. The VIRE uses the concept of virtual reference tags which are locations with known distance between real reference tags [6]. The RSSI of Each virtual reference tag is calculated by its surrounding real reference tags RSSI [6]. But, how to select the correct real or reference tags to calculate the tracking tag coordination is one of the most important problems of VIRE algorithm. In this paper, we propose a new method to estimate the tracking tags’ location based on virtual reference tags instead of increasing the number of real reference tags. In this method after calculating the RSSI of each virtual reference tag by its surrounding real reference tags RSSI, sensing area is divided into several sections and tracking tags’ location is calculated by real or virtual reference tags which are in the section that the objects has the highest probability to be enclosed within. II. METHODOLOGY In the deployment of our RFID indoor localization system, we need m=I×J real reference tags. These m reference tags have to be distributed in a matrix pattern around the sensing area as shown in fig. 1. The purpose is to find tracking tags location, which are implicated as objects or people in an indoor environment. To improve positioning accuracy we propose using virtual reference tags. So, we determine these virtual reference tags locations with specific coordinates between real reference tags and calculate their RSSI using the RSSI of real reference tags. According to the method presented in VIRE algorithm, n-1 virtual reference tags are assumed between each two real reference tags [6]. Therefore, there are (n+1) 2 -4 virtual reference tags, 4 real reference tags and totally (n+1) 2 reference tags in each section [6]. Therefore, the RSSI values of each virtual reference tag will be calculated according to what was said in VIRE approach. Since n-1 virtual reference tags are assumed Don’t write anything in this box. This is meant for the WAIRCO content.