Design And Implementation Of LORA-based Indoor Positioning System

With the increasing demand for indoor positioning in the society,indoor positioning technology has been increasingly concerned by domestic and foreign scholars.However,the hardware platforms used today are mostly short-range wireless communication technologies.The positioning systems using these platforms have the disadvantages of small scope,high cost and large node density.Therefore,LORA with long-distance transmission characteristics is used as the hardware platform in this thesis.For the positioning method,this thesis uses the time difference of arrival(TDOA).In practical applications,the effect of NLOS on accuracy cannot be ignored.Therefore,optimizing the NLOS effect is the difficulty in improving the accuracy of the positioning algorithm and is also the focus of this study.This thesis focuses on the LORA platform and analyzes the methods and error sources of TDOA implementation on this platform.According to the analysis of LORA,it is a narrowband signal.Therefore,TDOA data cannot be measured by using cross-correlation function.Instead,TDOA values must be collected by direct measurement.The error is mainly due to the influence of NLOS.First of all,in this thesis,in order to reduce the precision of the large NLOS value caused by the memory characteristics possessed by the Kalman filter in the ranging stage,the Kalman gain formula is increased by the threshold factor,so that the modified Kalman can effectively filter out the measured values of the large NLOS effect.The TDOA value converges to the true value better,and the ranging accuracy is 8% better than that of the standard Kalman.Secondly,the comparison shows that the Tayler series algorithm performs better than the Chan algorithm.Therefore,the Tayler series expansion method is used to optimize the positioning process.Multiple iterations is used to reduce positioning errors.Overall,the accuracy of Kalman+Tayler series fusion method was improved by 39% compared with Kalman+Chan fusion method.In this thesis,STM32F103 and SX1278 are used to implement the positioning system in the range of 1km×1km.The number of nodes used is seven,which is used to verify the hardware platform and the improved algorithm.Tests are performed in a well-built system.The experimental data shows that the positioning error within the positioning area is estimated to be 83% below the position of 20 m.