TrackT: Accurate tracking of RFID tags with mm-level accuracy using first-order taylor series approximation☆
Introduction
RADIO Frequency Identification (RFID) technology is increasingly used in various applications such as assisted tracking of robots [1], product identification [2], asset assessment, indoor positioning and so on. Currently, the common method to locate passive tags in practice is described as follows. People often deploy many RFID readers in different monitoring areas to continuously read RFID tags. The tags are assigned with the unique Electronic Product Code (EPC) and given position information beforehand in the database. Once a reader captures a new tag's EPC, people consider that the tag has been moved to the reader's surveillance area. However, there are many disadvantages in this coarse method: (i) Low positioning resolution. The reading range of an RFID reader antenna is generally about 3∼10 meters, and RFID readers can only provide “absence and presence” results, so the system is far from meeting the high accuracy requirements. (ii) False negative reads [3,4] and false positive reads [5,6]. The former means that a reader fails to read a tag in the reading zone and the latter means that a tag in some other areas outside the intended read zone is read. These two problems can also affect the positioning accuracy.
Many applications will benefit from millimeter level (mm-level) localization accuracy. For example, false positive reads will be avoided by setting the intended reading zone beforehand. If the RFID tag is within the area, the RFID reader will record and report events related to this tracked tag. If not, the reader will ignore it. As another example, in a large-scale clothes shop a retailer could use the RFID location system to visually track clothes with RFID tags, making sure that sales representatives could easily find matching clothes which may have gone astray and customers could easily know the locations of wanted clothes.
At present, the two key approaches for RFID localization are based on received signal strength indication (RSSI) and radio frequency (RF) phase.
(i) RSSI. The RSSI methods [7], [8], [9], [10], [11] need to deploy many reference tags since the absolute calibration for RSSI measurements is rather difficult, and positioning accuracy is greatly affected by antenna design, impedance matching, and the changes in reflection coefficient [12]. The distance error is about 60 cm. At present, many commercial-off-the-shelf (COTS) RFID readers can report RF phase once an RFID tag is successfully read. Phase resolution can reach 1.5 × 10 − 3 radians, offering (1.5 × 10 − 3 × 32.587)/4π = 3.89 × 10 − 3cm ranging resolution for an RF carrier wave at the frequency of 920.625 MHz. In addition, the RSSI-based methods are not a good choice because the propagation environment will more easily affect the RSSI measurements than phase [13]. As a result, the method to track tags’ movement trajectories based on RF phase with higher resolution and better noise-tolerant ability than RSSI has received many researchers’ attention.
(ii) RF Phase. Angle of arrival (AoA) [14], [15], [16] uses multiple antennas to receive the tag's phase and then computes their angles based on phase difference. But these methods need to put a strict constraint on the antennas’ spacing. Backpos [17] deploys more than three antennas to read tags’ phase, where two adjacent antennas should be within a spacing of half a wavelength. After that, they use the hyperbolic positioning method to locate tags. However, if adjacent antennas’ spacing is greater than half a wavelength, the method will fail to deal with the phase periodicity, leading to numerous possible candidates. PinIt [18] captures and extracts multipath profiles of the target tag and reference tags at known positions via an antenna motion, and then adopts dynamic time warping (DTW) techniques to pinpoint a tag's location. However, it needs to deploy dense reference tags in advance and can't track mobile tags. DAH (Differential Augmented Hologram) [19] can track known and unknown trajectories of mobile tags with a mean error distance of 0.6 cm and 12 cm respectively using COTS RFID readers in a lab environment. The main idea is to construct a virtual antenna array in tag movement and leverage the statistical method to find the optimal trajectory. DAH is effective way to track and locate moving tags than previous work. However, thermal noise and external inference, perhaps from human moving besides RFID tags in actual circumstance will both affect the reported RF phase. Also, as the surveillance area expands, the computations will jeopardize the real-time performance. In addition, the accuracy of tracking unpredictable movements should be improved in some extent.
In this paper, we propose a novel positioning and tracking method based on measured phase using COTS RFID devices, called TrackT. The phase periodicity is mainly used in our approach to improve the ability of noise tolerance and real-time performance in tracking mobile tags. At first glance, the phase periodicity is a negative factor and many researchers try to eliminate the unknown parameter in previous work. However, the differences of the phase periodicity between two continuous reads are an effective indicator introduced in the paper, with stronger noise-tolerance than phase-difference method like DAH. The basic idea of TrackT is described as follows. We first divide the surveillance region into mm-level square grids and assume that there is a visual tag as same as the tracked tag on the center of every grid. (i) Known Movement. We compare the sequence of the difference of phase periodicity between each virtual tag and the physical one to acquire several candidate points and then separately compute the corresponding double difference of true phase to find the minimum value. So the grid with the minimum value is chosen as the optimal position of the tracked tag. Once the initial position has been determined, the predefined track will also be fitted. (ii) Unknown Movement. In this case, both the initial position and the tag's displacements should be estimated. We suppose that all of antennas read the tag at the same time, and then the tag's displacement every read round could be calculated using first-order Taylor series expansion. For the unknown initial position, we leverage the same method of tracking known movement trajectory to select the optimal one. For achieving real-time performance, a rule to reduce the computations is also proposed in the paper.
Summary of Results: We build the system using Impinj R420 RFID reader, four 8dBi antennas and RFID passive tags. Our main results are described as follows:
(i) When four RFID antennas are deployed around the monitoring area, TrackT can achieve mm-level accuracy with mean error distances of 0.26 cm and 0.55 cm for controllable and uncontrollable movements.
(ii) TrackT is a real-time localization system so the computation time for producing an intermediate result should be less than the time interval of about 33 ms between successive inventories of the same tag. After optimization, TrackT can achieve a mean of about 11.21 ms computation time, which can meet the real-time requirement.
Contributions: TrackT can capture the known and unknown movement trajectories of mobile RFID tags with mm-level accuracy. As far as we know, TrackT is the first to leverage the phase periodicity with high noise tolerance to track mobile tags. Besides, we also exploit the localization theory of carrier phase measurement in global positioning system (GPS) to track unknown trajectory of the moving RFID tag. As a result, TrackT with less positioning error and higher real-time performance can be easily implemented in real applications based on COTS RFID devices.
The remainder of the paper is described as follows. The background and empirical studies are introduced in Section II. The main design of the TrackT approach is presented in Section III. We discuss additional details in Section IV. The implementation and evaluation are given in Section V. Finally, Section VI introduces the limitations and concludes the paper.
Section snippets
Preliminaries
In this section, we mainly introduce RF phase, Doppler frequency shift, session persistence and phase periodicity.
TrackT overview
In this section, we give details on our tracking approach in two cases. In 2D scenario, a tracked RFID tag is aligned with the center of antennas. We first introduce how to track a mobile RFID tag's trajectory in the case of conveyor with known track that the trajectory function and the tag's speed are known for us before tracking. Then we describe how to track the unpredictable movement that both trajectory function and speed are unknown in prior.
Discussion
In this section, we attempt to answer some practical issues.
Implementation and evaluation
In this section, we present the implementation and conduct performance evaluation on the prototype.
Conclusion
In this paper we mainly achieve to track mobile tags based on phase collected by the COTS RFID reader. The difference of phase periodicity deduced by reported phase has strong tolerance to multipath effect, which is employed to acquire all of candidates in the surveillance area. On this basis, we calculate the double-difference values corresponding to these candidates and select the minimum value as the final position. In addition, since the time between successive inventories of the same tag
Zihongqin Wang (STM’16) received the Master degree in Computer Software and Theory from Nanjing University of Posts and Telecommunications, Nanjing, China in 2014. He is currently pursuing the Ph.D. degree at Nanjing University of Posts and Telecommunications. In 2016, he became a Student Member of IEEE. His research interest includes Radio Frequency Identification (RFID), sensor network and fancy Human Interaction.
References (33)
- et al.
Primate-inspired communication methods for mobile and static sensors and RFID tags
ACM Trans. Auton. Adapt. Syst. (TAAS)
(2011) - et al.
Radio frequency identification: technologies, applications, and research issues
Wireless Commun. Mobile Comput.
(Jul. 2007) - et al.
Issues with RFID usage in ubiquitous computing applications
- et al.
Reasoning about uncertainty in location identification with RFID
- et al.
Reasoning about uncertainty in location identification with RFID
- et al.
Fully autonomous trajectory estimation with long-range passive RFID
- et al.
LANDMARC: indoor location sensing using active RFID
Wireless Netw.
(Nov. 2004) - et al.
RFID-based 3-D positioning schemes
- et al.
RFID indoor positioning based on probabilistic RFID map and Kalman filtering
- et al.
Otrack: order tracking for luggage in mobile RFID systems
Complete link budgets for backscatter-radio and RFID systems
IEEE Antennas Propag. Mag.
Speedway Revolution Reader Application Note: Low Level User Data Support
Vire: active rfid-based localization using virtual reference elimination
Accurate localization of RFID tags using phase difference
New measurement results for the localization of UHF RFID transponders using an Angle of Arrival (AoA) approach
Phase based spatial identification of UHF RFID tags
Cited by (63)
Dense and long-term monitoring of earth surface processes with passive RFID — a review
2022, Earth-Science ReviewsCitation Excerpt :The localization of tags at longer distances uses UHF tags (Miesen et al., 2011), and exploits the received signal strength (RSSI) or the phase of arrival. Signal-strength-based methods have been initially introduced for tag localization (Griffin and Durgin, 2009; Ni et al., 2003), followed by phase-based methods which usually offer the best accuracy, reaching one centimeter or less (Scherhäufl et al., 2015; Wang et al., 2016; Zhou and Griffin, 2012). In practice the phase-based methods show the best potential for observing ground displacements (Le Breton et al., 2019).
Occupancy detection and localization strategies for demand modulated appliance control in Internet of Things enabled home energy management system
2022, Renewable and Sustainable Energy ReviewsCitation Excerpt :Generally, RSSI is affected by multipath, reader antenna gain and orientation of the tag. Hence, recent researches have embraced phase-based methods to mitigate the effects of environment dynamics and achieve fine-grained resolution for stationary [64] and moving tags [65]. Centimetre level accuracy could be achieved by detecting the phase difference between a pair of tags placed on the target and analysing the impact of reader antenna position on accuracy [66].
Automated book location and classification method using RFID tags for smart libraries
2021, Microprocessors and MicrosystemsForeground-aware Siamese tracker with dynamic template in wireless sensor networks
2021, Ad Hoc NetworksCitation Excerpt :In recent years, with the development of the embedded processor and the technology of image and video processing [39, 40] requirements, target tracking based on wireless sensor network has wide application in many fields, such as robotics [1, 51], battlefield detection [52],autonomous driving [2], suspect tracking [52], navigation [53] [54] and indoor positioning [55] [56].
Passive radio-frequency identification ranging, a dense and weather-robust technique for landslide displacement monitoring
2019, Engineering GeologyCitation Excerpt :Currently, phase-based methods are compatible with commercial readers and offer the best accuracy. Several studies used the phase to localize tags with an accuracy of one centimeter or less (Scherhäufl et al., 2015; Wang et al., 2016; Zhou and Griffin, 2012). However, these experiments were of short duration (<1 h), over a short range (<10 m) and were performed indoors.
Zihongqin Wang (STM’16) received the Master degree in Computer Software and Theory from Nanjing University of Posts and Telecommunications, Nanjing, China in 2014. He is currently pursuing the Ph.D. degree at Nanjing University of Posts and Telecommunications. In 2016, he became a Student Member of IEEE. His research interest includes Radio Frequency Identification (RFID), sensor network and fancy Human Interaction.
Ning Ye received the B.S. degree in Computer Science from Nanjing University in 1994, the M.S. degree in School of Computer & Engineering from Southeast University in 2004, and the Ph.D. degree in Institute of Computer Science from Nanjing University of Posts and Telecommunications in 2009. She is currently a Professor there. In 2010, Ning Ye worked as a Visiting Scholar and Research Assistant in the Department of Computer Science, University of Victoria, Canada. Her research interests include information processing in wireless networks and Internet of Things. She is a senior member of Chinese Computer Federation (CCF).
Reza Malekian received the B.Eng. degree in computer engineering, the M.Eng. degree in telecommunications engineering, and the Ph.D. degree in computer science. He is currently a Senior Lecturer with the Department of Electrical, Electronic and Computer Engineering, University of Pretoria, South Africa. His research lies in the area of advanced sensor networks, Internet of Things, and mobile communications. He is also a Professional Member of the British Computer Society and Chartered Engineer.
Fu Xiao (M’12) received the Ph.D Degree in Computer Science and Technology from Nanjing University of Science and Technology, Nanjing, China, in 2007. He is currently a Professor and PhD supervisor with the School of Computer, Nanjing University of Posts and Telecommunications, Nanjing, China. His main research interest is Wireless Sensor Networks. Dr. Xiao is a member of the IEEE Computer Society and the Association for Computing Machinery.
Ruchuan Wang was born in Anhui Province, China. He researched on graphic processing at University of Bremen and program design theory Ludwig Maximilian Muenchen Unitversitaet from 1984 to 1992. He is a professor and tutor of Ph.D. candidate in Nanjing University of Posts and Telecommunications since 1992. Major research interests include wireless sensor networks and information security.
- ☆
The research is support by National Natural Science Foundation of P. R. China (No. 61572260), Major Program of Jiangsu Higher Education Institutions (No. 14KJA520002), the Key Research and Development Program of Jiangsu Province (Social Development Program) (No. BE2015702), Jiangsu Planned Projects for Postdoctoral Research Funds (No. 1302055C), China Postdoctoral Science Foundation (No. 2014M560440), and National Research Foundation, South Africa (AOX220).