Research On RFID Indoor Positioning Algorithm Based On CQPSO-BP Wireless Channel Modeling Method

RFID positioning technology has the advantages of low power consumption,long working distance,non-contact,small size,and adapt to complex environments,and stands out in a large number of indoor positioning technologies.At present,most researches on RFID positioning technology focus on two-dimensional planes,but in real life,the demand for precise positioning of three-dimensional space is increasing day by day.Therefore,from the two aspects of wireless signal propagation model and indoor positioning technology,two methods are proposed to improve the accuracy of indoor positioning system.In view of the fact that there are obstacles in the wireless signal during the propagation process,there are phenomena such as refraction,scattering and reflection of multipath propagation during the propagation of the signal,which makes the existing indoor wireless signal propagation model cannot be accurately described.A wireless channel modeling method based on CQPSO-BP network is proposed to change the signal indoors.Firstly,the BP neural network is optimized by the cultural quantum particle swarm optimization algorithm.Then the BP neural network is used to reasonably define and extract the BP neural network signal samples and train them to obtain the wireless signal propagation model.Using the obtained wireless signal propagation model and the RSSI value received by the reader to obtain the distance between the reader and the target tag,and use the distance for the LANDMARC indoor positioning system,and finally build the model with matlab2014 a platform for simulation verification.The results show that the proposed indoor wireless channel modeling method is more suitable for the complex and variable indoor environment than the traditional model,which helps to improve the positioning accuracy of the LANDMARC system.Aiming at the positioning performance of LANDMARC system,the grid-based density peak clustering algorithm selects the nearest reference label from the target label by clustering,and applies the reference label to the LANDMARC algorithm.Improve the positioning performance of the LANDMARC system.Firstly,it can be seen from the analysis that different indoor environments correspond to different RSSI values for the same communication point,and there are different "estimation tags".When the number of obstacles in the room is small,the "estimated tag" distribution is relatively concentrated.In the case where there are fewer obstacles indoors,a concentrated distribution of "estimated labels" may occur.In response to this phenomenon,the "estimation label" of the target label is obtained by repeating the experiment several times,and then the "estimation label" space is meshed into units by using the mesh division rule in the algorithm,and then the grid unit object set is density-based.The peak value is classified.Finally,the cluster with the highest density value is used as the reference label.It is used in the LANDMARC positioning system to calculate the coordinates of the target label.Simulation results show that the optimized algorithm has better positioning accuracy than the LANDMARC algorithm.Finally,the CQPSO-BP wireless channel modeling method and the grid-based density peak clustering algorithm are combined with the optimized LANDMARC algorithm,and the simulation experiment is carried out to analyze the algorithm compared with the proposed two methods and the LANDMARC algorithm.Still have a certain improvement in stability.