Moving Trajectory Analysis Based On RFID Signal

Radio Frequence Identification(RFID)technology provides a new way to get information and is theoretically similar to bar code Identification.It uses radiofrequency electromagnetic fields to transmit data in a wireless,contactless way to automatically identify and track tags attached to objects.Due to the low price of RFID tags,simple deployment and passive RFID has the advantages of non-line-of-sight,non-contact,no blind spot and low cost,which can realize the rapid and accurate identification of items.More and more large shopping malls and well-known manufacturers use it for product tracking and antitheft.Following this trend,rfid-based offline shopping behavior analysis and mining has gradually become a new research hotspot.However,most of the existing schemes are only applicable to the identification of specific actions(pick up actions)in special scenes(clothing selection scenes).Due to the limitation of indoor RFID positioning and tracking accuracy,it is still unable to effectively distinguish shopping choices of different users at different times,and there are many key problems to be solved.This paper intends to study the moving track based on RFID signal.Specifically,it includes the following three processes :(1)Measure the similarity of tag movement track and analyze the correlation between goods;(2)Track the movement track of the tag in the local scope and extract the movement characteristics of the tag;(3)correlate the monitoring results at different times and locations,and track the selection process of user shopping.By analyzing the trajectory information of commodities at key points and user behavior,the main innovation points and specific work are as follows:1)Accurate tag track tracking.By deploying a small number of readers in key locations(for example,elevators or partition entrances,etc.),multiple monitoring points can be built to monitor the tag's local trajectory and global direction.This paper intends to realize high-precision(centimeter-level)whole-course tracking(from entry to exit)for multiple tags within the monitoring range of a single monitoring point.The trajectory is analyzed by collecting phase and time information within the detection range.2)Theoretical value and measured value match.This article first to match the theoretical value and measured value of phase to achieve goods path analysis,through under static conditions to measure phase value with the theoretical value of a series of comparison and deduction,at the same time will be collecting device(and antenna)own initial phase deflection factors into account,the range of the phase to the phase matching conditions to establish inequalities.Then the theory is applied to the processing of dynamic time series to obtain a series of matching points within the monitoring range,that is,the motion track.3)Accurate label group division.In the shopping process,the goods purchased by the user usually move in the same way as the user(for example,speed and direction,etc.).By analyzing the similarity of label track,the different attribution of goods can be determined.This paper intends to realize efficient and accurate classification of a large number of product labels given a small number of reference labels(labels attached to shopping cars).The similarity measurement of tag track is mainly studied.4)Application software development.The algorithm of each process is implemented with specific code,and visual JUI is written with Java FX components to form a complete trajectory analysis system.