Research On Key Technologies In Passive Indoor Localization Based On Interference From Target

Location information plays an important role in our daily life and industrial production,and indoor localization has gained much attention recently.Numerous applications including storage management and somatosensory systems would benefit from accurate passive indoor localization.At present,the new technologies,such as the smart home and Internet of Things(IoT),develop rapidly and put forward new demands for indoor localization.Passive indoor localization is of great significance in daily life and frontier technology.Radio Frequency Identification(RFID)and Ultra Wide Band(UWB)are two technologies that have great potential research of indoor localization.RFID is a typical wireless identification technology and it is the important technology of the Internet of Things(IoT).In particular,Ultra High Frequency(UHF)passive RFID technology has the advantages of low cost,low complexity,and time-efficient inventory verification.UWB is a technology that uses high-bandwidth to obtain high time resolution.It has high ranging accuracy and can acquire information at different distances.In this dissertation,we studied the three-dimensional passive indoor localization based on the target interference to UHF RFID reference tags,and proposed a localization algorithm,named Lobain,based on RFID channel parameter estimation.Lobain can estimate a target's coordinates without large number of reference tags.Then,the passive multiple target indoor localization was studied.The passive multiple target indoor localization algorithm,named Spinca,based on joint interference cancellation was proposed.Spinca can estimate the target number and positon from coherent signal superposition.Finally,the posture recognition of passive target was studied.The RFID tags can estimate the position of targets,but it is difficult to obtain more detail of the target.Therefore,the ultra-wide band was adopted in the posture recognition study.All the presented methods were evaluated in the real environment.The main contents are as follows:(1)The passive RIFD tag was simulated using CST to obtain the electromagnetic parameters of RFID tag.These parameters were used to explore the principle of interference from target to passive RFID tag.Then the actual interference of tag was tested.An object moved around a tag and the measured phase of tag was sampled.The simulation and the test verified the feasibility of indoor localization using interference from target to RFID tags.Finally,the planar array of reference tags was constructed,and the preliminary two-dimensional localization of the ground target was realized.(2)A three-dimensional passive indoor localization method based on target-to-tag interference is proposed.The traditional RFID reference tag methods need too many reference tags,so the corresponding three-dimensional localization is almost impossible.This problem was solved by constructing the spatial signal transmission model and making full use of the phase characteristics of adjacent reference tags.The target-related channel information is extracted from the complex received signals,and the relative phase wrapping number is used to solve the problem of tags' phase wrapping.The least squares method is used to estimate the channel parameter and to complete the target coordinate estimation.The algorithm reduces the dense reference tag array to the L-shaped array,achieving three-dimensional localization of passive targets.(3)A passive multiple-target indoor localization method based on joint interference cancellation is proposed.In the multiple-target application scenario,the performance of some single-target RFID localization methods will deteriorate.The main problem is to accurately estimate the target number and target position from the coherently superimposed signals.In order to solve this problem,the spatial spectrum of the observation area was constructed using the maximum likelihood.A uniform circular array was used instead of a linear array to avoid the influence of the aperture limitation.The relative position of the reference tags was designed according to the test results.Then the corresponding spatial signal transmission model was modified to avoid the influence of the mutual coupling of RFID tags.Finally,the movement of the reader antenna was used to eliminate false positives and false negative estimates.Multiple-target Indoor localization is achieved through this series of interference cancellation methods.(4)An indoor passive target localization and posture recognition method based on deep learning is proposed.The distance from the module to the target was obtained by analyzing the signal of UWB module.The target position estimation was calculated using the least squares method.Then the received signal of UWB module was pre-processed using the distance information,and then the target's posture classification and recognition was realized by convolutional neural network.The system can complete the target position and posture estimation.The target's posture can be different kinds of item placement or different postures of the human body.