Multi-target Localization Technology Research Based On MDS For UHF RFID

With the rapid development of Internet of Things and mobile communication technologies,the demand for wireless localization technology is becoming more and more eager.Research statistics show that human beings live indoors for about 70% to 90%of their lifetime,and therefore the localization technology for indoor environments has attracted the attention for researchers.Radio Frequency Identification(RFID)plays an important role in indoor localization due to its advantages of strong penetration,large data capacity and rapid recognition.This paper studied the channel model of passive Ultra High Frequency(UHF)RFID localization system in the influence of multipath effect and Gaussian white noise.On this basis,the phase information obtained in the channel was used to construct the ranging model between the reader and the tag,which was Phase Difference of Arrival(PDOA)method.The results of simulation and experiment showed that the ranging error obeyed the Gaussian distribution.Based on the ranging model of PDOA method,the paper proposed the multi-target cooperative localization algorithm based on Nonmetric Multidimensional Scaling(NMDS)and the multi-target localization algorithm based on cluster-NMDS.The multi-target cooperative localization algorithm based on NMDS used the ranging error of PDOA method to establish a ranging model in the multipath environment,and then obtained the distance information between the RFID reader and the passive UHF tag with NMDS algorithm to achieve simultaneous localization of the multiple targets.However,this algorithm took too long time to locate a lot of target,which made it difficult to achieve real-time positioning effect.The multi-target localization algorithm based on cluster-NMDS can solve this problem.Firstly,lots of target tags were clustered by the K-means algorithm.And then the tags in each category were located by the multi-target cooperative localization algorithm based on NMDS.Simulation experiments showed that the multi-target localization algorithm based on cluster-NMDS had certain advantages in localization time and localization accuracy.