Font Size: a A A

Research On Shopping Behavior Recognition And Analysis Of Physical Store Based On RFID

Posted on:2021-04-26Degree:MasterType:Thesis
Country:ChinaCandidate:J WangFull Text:PDF
GTID:2568306104471344Subject:Computer Science and Technology
Abstract/Summary:PDF Full Text Request
Shopping behavior data is important for marketing understanding and sales strategy adjusting.Through detailed and reliable shopping behavior analysis,retailers can understand customers’ shopping preferences and grasp the market trend in time,so as to optimize commodity strategies to provide more favored products continuously.This paper introduces a novel shopping behavior identification and analysis solution,which is mainly applicable to the shopping behavior detection and analysis in physical stores,that is,to collect customers’ shopping behavior in physical stores and analyze the potential fine-grained behavior value.Firstly,for the problem of inaccurate phase measurement of current RF readers,this paper proposes to use trigonometric function to standardize the measurement of phase value,and apply the periodicity of trigonometric function to overcome the original phase skipping and confusion,so as to obtain a stable label phase stream output.Second,since there is inevitably behavior detection error under a single RF tag,this paper applies a new dual-tag model,and analyze the motion trajectory of typical shopping behavior under the mode of dual-tag,and then utilize the synthetic doppler spectrum fused by the two tags to enhance target behavior signal,this help us obtain specific doppler characteristics of shopping behavior to realize accurate behavior detection ultimately.Thirdly,a fine-grained behavior information extraction scheme based on peak detection is proposed,which attempts to extract the time and frequency characteristics of behaviors based on accurate behavior recognition.At the same time,this paper takes a full consideration of the importance of fine-grained behavior information,explores to establish a ranking list of commodity popularity from the perspectives of commodities and consumers using these fine-grained behavior characteristics.Finally,a prototype system is deployed in two real environments to analyze and evaluate its performance.The system takes advantage of commercial RFID devices,and does not require store upgrades or additional equipment or hardware modifications to the RFID devices during actual deployment.
Keywords/Search Tags:RFID, Shopping behavior, Doppler calibration, Information extraction, Commodity ranking list
PDF Full Text Request
Related items