Research On High-precision Relative Positioning Technology Based On RFID Phase Information

Since the beginning of the new century,wireless positioning technologies have attracted more and more attention due to the rise of IoTs(the Internet of Things)and LBS(Location Based Services).Among those technologies,RFID(Radio Frequency Identification)technology has been drawing a lot of interest for its large-scale deployment in industrial environments and commercial systems and its excellent features such as no-battery tag,low cost and remarkable scalability.In the field of RFID positioning technology,a lot of research work has been proposed and different practices have been carried out in recent years.However,the performance of RFID-based localization can easily be affected by signal fluctuations and multipath effects,which can lead to a decrease in positioning accuracy.In many target positioning applications such as the baggage carousel of the airport and the book sorting of the library,compared to absolute positions,it is relative positions between the targets that are more concerned and needed to be known.Although some scholars have already proposed many RFID based approaches for object localization,they mainly focus on absolute object localization.Furthermore,the large cost requirements and error margins of the hardware required in those approaches make them unsuitable for relative object localization.A few researches that do involve relative localization have the problems of low accuracy,bad robustness,and poor scalability.To address those problems,a relative positioning scheme based on RFID phase information is proposed in this paper.The main work and innovations of this paper are as follows:Firstly,A dual-window moving average filtering method is designed to preprocess the original RFID phase data to reduce the noise signal interference which can lead to an increase in sorting error.Secondly,an improved dynamic time warping algorithm is proposed to find the characteristic phase-time zone in the entire time series of RFID phase,and multiple strategies are applied to calculate the timestamp of the peak point in the fund phase-time zone at which the tag is closest to the reader.And the timestamps of all tags are what the tags are sorted by.Thirdly,a vertical distance adaptive strategy and a correlation coefficient based method are introduced to ensure the stability of the sorting performance when the vertical distance between the reader and the tag changes or the distances between tags are too short.A Speedway R420 reader and multiple UHF passive tags,both produced by a company named ImpinJ,were used in the experiments which proves that the scheme of this paper do words and has considerable accuracy,good stability and scalability and has a good future of development in indoor wireless relative positioning applications.