Device-free Target Tracking System Using Fine-grained RFID Signatures

Target tracking acts as a primary element of monitoring malicious or non-cooperative targets including intruders and one of the most fundamental type of sensing data for IoT,and thus plays an important part in everyday life.According to figures,people spend more than 80%of their daily time indoors,leading to increasing demand and flourishing markets of target tracking based services.During the past decades,various device-free tracking systems have been proposed,thanks to the wide availability of RF signals.However,as a result of current device-free tracking systems based on cheap commercial devices performing well in the trained environment but poor in other environments because of different multipath reflections,the device-free tracking system which can provide round-the-clock services still lacks.This paper introduces RDTrack that leverages doppler frequency shift,which is not sensitive to multipath,to sense moving targets and track them.Compared to the related work,RDTrack is no training for tracking,broadening the communication distance,using standard RFID reader and tag,and with no amendments.In summary,the core contributions as follows:1)Motion detection and segmentation.Based on the DWT theory,we can summarize the uncontinuous examination as a pet phrase which means that the gap is high frequency message,much higher than the frequency of general signal.Thus,due to the detected gap,RDTrack can segment the beginning and the ending of affecting time of the tag.And the experimental results show that motion detection accuracy is 99.3%.2)Trajectory estimating model.To achieve fine-grained device-free tracking system,this paper build trajectory estimating model using HMM which alters its own visible collection O and transition probability matrix B,to improve the matching accuracy and reduce the time complexity.We implement RDTrack with cheap commercial off-the-shelf RFID hardware and extensively evaluate RDTrack in laboratory,lobby,staircase and library.Meanwhile,we invite lots of volunteers whose body type is different from each other.Our results show that RDTrack is effective in tracking the moving target,with low tracking error of 32cm using a single RDTrack only.Amazingly,the tracking error will decrease to 27cm if we use two RDTracks.This accuracy is robust for different environments,highlighting RDTrack's ability to enable future essential device-free moving-based interaction with RFID devices.