Research On Method And Application Of RFID-based Location Perception

In the research and practice of smart buildings and building security,the location information of people or objects is an important attribute of building space.The location perception method based on radio frequency identification（RFID）technology are concerned by industrial applications and academic research for its advantages of low cost,non-contact and short delay.The traditional RFID-based location perception methods have many disadvantages,such as vulnerability of radio signals to environmental interference,multiple RFID tag interferences with each other,and low recognition accuracy.In this paper,oriented to the space of buildings,the research on the location perception method based on RFID is carried out.Firstly,in order to overcome the problem of low recognition accuracy caused by the discontinuity of RFID signal reception time,the idea of sequence analysis was introduced,based on a single RIFD beacon,a location perception model based on RFID and Long ShortTerm Memory（LSTM）network was constructed through machine learning technology,to realize the two-class identification of RFID location.Results show that the accuracy rate of the location perception model identification using sequence analysis is not less than 98%,and the identification accuracy rate of multi-tags can reach more than 93%.Further,in view of the limited sensing range of a single RFID beacon,a location perception model based on RFID and ensemble learning is constructed using the data of multiple RFID beacons for the needs of large-space location sensing.Results show that the proposed method can effectively improve RFID sensing scope.Secondly,in order to meet the more precise requirements for location perception in some application scenarios,this paper designs a localization method based on Back Propagation（BP）neural network.In order to overcome the problems of BP neural network easily falling into local optimum and slow convergence speed,this paper uses Particle Swarm Optimization（PSO）optimized BP neural network to build an RFID positioning model to estimate the specific position of people or objects.Results show that the proposed method can effectively solve the problems of long positioning time,large positioning error and high cost of traditional positioning methods.Finally,for the application requirements of smart buildings and building security,an RFID-based asset equipment intelligent management system is designed and implemented.The system consists of a front-end RFID signal acquisition device and a local information processing server.The RFID signal acquisition device is responsible for collecting RFID data,and the local information processing server is responsible for analyzing and processing the RFID data through the algorithm in this paper.The location information of the asset equipment is uploaded to the local database server,and finally the monitoring and management of the indoor asset equipment is realized through the Web platform.Results show that the system can meet the practical application requirements of smart building security system for asset equipment monitoring.Relying on a set of RFID information sensing device,the location perception and indoor positioning methods are studied in this thesis can quickly and accurately estimate the location information of people or objects through the analysis of RFID signals,and provides a new support for the intelligent operation of smart buildings and building security platforms,so that whether it is the operation of building equipment or the decision of building security,the location information of people or objects in the area can be used as a new constraint.Figure [48] Table [25] Reference [68]...