Research On Positioning Method And Positioning System Of High Precision RSSI

With the rapid development of indoor localization technology,RFID technology becomes an excellent candidate for indoor localization by its advantages such as low-cost,non-contact,non-line-of-sight and high location accuracy.However,the accuracy of RFID localization algorithm is easy to be affected by the label density and the efficiency of the algorithm,so it is very important to study a reasonable localization algorithm to solve the above problems effectively.This paper focuses on the analysis of the advantages and defects of the LANDMARC algorithm,and analyzes the feasibility of the ELM training algorithm in the localization field.We have done the following work:A segmentation algorithm based on LANDMARC and ELM method is proposed in this paper,which can be divided into two stages.In the off-line phase,the system is used to collect the RSSI of all the entity reference tags,and then calculates the RSSI of the virtual tags using the propagation loss model.ELM algorithm establishes a model through the training of the database.In the on-line phase,first of all,LANDMARC location algorithm is used in regional location,to lock the target area quickly;then in the locked area,the selected area is introduced into the ELM training model,which is used to determine the exact location of the target by ELM training.The applicability of the algorithm in indoor positioning is verified by simulation.The indoor localization system by using the radio frequency development platform is built to do the actual test.The influence of different parameters of ELM model on the localization accuracy is discussed,which provides a reference for the practical application of the algorithm.It was showed by the experimental results that the average location error of the proposed algorithm is efficient and the average positioning accuracy is 0.30 meters in interior space and the same density of reference tag.The reduction is 38%over LANDMARC in estimation error and is 76%over LANDMARC in localization time.It is showed by the experimental results that the average location error of the proposed algorithm is efficient and good enough to meet the requirements of indoor localization.