Indoor Localization And Sensing Through Radio Frequency Identification

As a key enabler of Internet of Things,RFID has got broad attention from both Industrial and research area from the start.Compared against traditional barcode,RFID system has the advantage of non-contact identification,long communication range,long service life,environmental resistance,high security,capability of identifying fast moving or multiple objects and so on.RFID has already been used in a variety of applications,such as supply chain management,tracking of goods or animals,access control,airport baggage tracking logistics,transportation management,etc.With increasing number of objects carrying electronic tags,purely identification ability can not meet practical demands of many applications.This work explores how to extend the sensing capability(e.g.,location sensing and velocity sensing)of RFID in various scenarios.Our key innovations are the following three folds.First,we propose a light-weighted,inexpensive yet highly precise reader localization schema via a few spinning tags.To this end,we enable each tag to emulate a circular antenna array and design an SAR-based method to estimate the angle spectrum of the target.Compared against previous AoA-based techniques,the proposed method is more accurate and immune to measurement noise,phase shifts caused by hardware characteristics.We also quantify the correlation between tag's measurements and its orientation to increase the final accuracy to centimeter level.Second,we present a hybrid RFID and computer vision system for fine-grained tracking of tagged objects.We reinforce the deficient localization capabilities of current RFID deployments with the supplement of simple CV equipment(camera).CV manages to give accurate trajectory tracking results of all current motion blobs while the reader collects phase measurements of target tag.A fusion algorithm is then designed to organically combine the two data modalities and match the tag to the motion blob that is most likely to carry it.The mean tracking accuracy is 10.33 mm.Third,we propose a robust and accurate spinning sensing system that can work in noisy settings.To overcome the challenge of device shaking,we attach dual RFID tags on the spinning surface and theoretical demonstrate that their relative phase is a periodic signal,which maintains the spinning frequency and is shaking-resilient.Besides,we introduce compressive sensing technique to recover high-frequency rotation signal with relatively low reading rate of RFIDs.The target's spinning speed can be finally inferred by analyzing the frequency of the acquired signal.The proposed method achieves a mean sensing accuracy of 0.27 Hz.All of these research efforts are based on commercial RFID devices.We believe the sensing capability of RFID technology in location and velocity can be facilitated through this study.