Research Of Indoor Localization System Based On Radio Frequency Identification

Radio Frequency Identification(RFID), is a generic term for the technology using radio waves to automatically identify individual items and collect associated information. RFID systems with the advantages such as large quantity of data, confidentiality, anti-interference capability, short recognition time and lower cost, are widely used in various industry sectors. With the popularity of mobile computing devices, the development of wireless location technique, especially indoor wireless location technique is getting more and more concern, such as wireless navigation, warehouse management, logistics tracking, personal entertainment, mobile office units, staff management and other areas. However, ordinary location technologies can’t satisfy the needs of indoor environment and location accuracy, RFID technology becomes an excellent candidate for indoor location technique because of its advantages of non-contact and non-line-of-sight nature.Growing convergence among mobile computing devices and embedded technology sparks the development and deployment of location—aware applications. Therefore, location—aware systems are being paid more and more attention. Global Positioning System(GPS)is the most well—known location service in use today. However, GPS is invalid or having a poor positioning accuracy in the place where users cannot maintain Line of Sight(LOS)with the orbit satellite. In order to improve the location estimation accuracy, many different approaches such as Infrared, Ultrasonic, Wireless LANs and RFID have been proposed in recent years. Compared with other techniques, RFID technologies have several advantages such as no contact, non-line—of-sight nature, promising transmission range and cost-effectiveness. The LANDMARC approach, which presented the concept of reference tags, is the most typical indoor positioning system based on RFID. It was also the first attempt using active RFID for indoor localization with gratifying results. Moreover, with reference tags adopted, the LANDMARC system does not need a large number of RFID readers which are expensive and makes RFID technology cost-effective for indoor location sensing.This paper made a careful analysis of LANDMARC and VIRE system,and researched the layout of the system and the algorithm flow. In LANDMARC system, the main causes of the system positioning precision are hardware equipment, the choice of the nearest neighbors, the numbers and location of the reference tag. In LANDMARC system, using the reference tag number to increase precision has limit, and too much tags even decrease accuracy. At the same time, the edge tags’positioning accuracy cannot achieve satisfactory accuracies. In VIRE system, the main factors influencing the system performance lies in the selection of the virtual reference tags RSSI value and the threshold value of the nearest neighbors. VIRE and LANDMARC systems face a same problem-the edge tags’ positioning accuracys are significantly less than the positioning accuracys of center position tags.In some cases, accuracys even can’t meet the needs of the location.For fixed number of nearest neighbors caused losing of the good nearest neighbors and introducing of the bad nearest neighbors, an improved KNN(k-Nearest Neighbor algorithm) algorithm is proposed. The algorithm determines the selection area of the nearest neighbors by the estimated position of the undetermined tag and selects the nearest neighbors according to the RSSI difference. Under the same conditions, simulation results using MATLAB show that the new algorithm can improve the positioning accuracy by9.8%and reduce the maximum positioning error by21.8%.