Research On Optimal Distribution Of RFID Multi-tag Based On Photoelectric Sensor And Neural Network

As the core technology of the Internet of Things,photoelectric sensing and radio frequency identification have been widely used.Because of its advantages of non-contact,fast response and high accuracy,photoelectric sensing can play an important role in the application of RFID multi-target automatic.Therefore,the research of RFID multi-target identification combined with photoelectric sensor is now a hotspot and frontier topic in the field of RFID technology.In practical applications,RFID multi-target identification system will be influenced by various factors,such as temperature,media,multi-tag distribution,which will have a significant impact on the RFID system.The current solutions are relatively simple.Therefore,the research of relationship between the multi-tag distribution and the reading should be carried out,which will provide a new direction for improving the reading performance and popularizing the application of RFID technology.In this paper,photoelectric sensing and neural network are combined to study the relationship between the multi-tag distribution and the reading performance of RFID.The RFID multi-tag distribution detection and verification system has been designed and built based on photoelectric sensor.Based on the combination of photoelectric sensing and neural network,a series of simulation analysis and experimental verification for the optimization of RFID multi-tag distribution are carried out,which improve the dynamic reading performance of the RFID system and provide an important reference for the research of the RFID system.The main contents of this paper are as follows:(1)An RFID multi-tag distribution detection and verification system has been designed and built based on photoelectric sensor.Firstly,the design and construction of pallet-level RFID dynamic detection system were carried out.Then,the location systems are set up respectively for the different RFID multi-tag distribution.The projects can be tested and verified by the RFID multi-tag distribution detection and verification system include identification range and recognition rate test,multi-label anti-collision performance test and multi-tag distribution optimization test.(2)An RFID multi-tag positioning method based on image processing is proposed.According to the different distribution of RFID tags,this method uses threshold segmentation algorithm,template matching algorithm and edge recognition algorithm respectively.Compared with the positioning function of RFID reader,this method can accurately identify the RFID multi-tag distribution in a variety of interference environments.(3)An RFID multi-tag distribution optimization method based on neural network is proposed.This method simulates the relationship between RFID multi-tag distribution and its reading distance by BP,GA-BP and PSO neural network algorithm respectively.On this basis,the prediction and experimental verification has been carried out to find the optimal distribution of RFID multi-tag,which effectively improves the reading performance of RFID system.In addition,the three kinds of neural networks are compared in many aspects,and explain their respective environments.