Research On Indoor Positioning And Tracking Algorithm Based On UWB

With the development of positioning technology and the increasing demand for indoor positioning services,it is increasingly important to securely obtain high-precision indoor location information.Ultra-Wide-Band(UWB)technology is widely used to study and achieve precise positioning in complex indoor environments because of its high positioning accuracy,strong anti-interference capability,and good safety performance.Therefore,this paper focuses on UWB-based indoor positioning and tracking algorithms,with the goal of improving the algorithm performance and increasing the positioning and tracking accuracy.(1)To address the problems of non-line-of-sight(NLOS)errors and multipath interference in static target localization,which lead to the degradation of localization performance,an ultrawideband localization algorithm with gravitational field optimized BP neural network is proposed.The strong nonlinear approximation capability of BP neural network is utilized to solve the problems that the localization accuracy of time difference of arrival(TDOA)algorithm decreases with the increase of measurement error and the high time complexity.The gravitational field algorithm is further combined with its movement factor to make the weights and thresholds quickly and centrally distributed near the extremes,which improves the convergence speed of the network;at the same time,its rotation factor is used to make the weights and thresholds near the extremes randomly move away from the extremes,which effectively allev iates the problems of slow convergence and easy to fall into local extremes of the BP neural network-based ultra-wideband localization algorithm.The experimental results show that the root mean square error of the proposed localization algorithm is 7.95 cm,which basically meets the demand of indoor high-precision localization.(2)A sparrow search optimized particle filter tracking algorithm based on UWB is proposed to address the problem of degradation of particle weights when tracking moving targets using particle filter(PF)algorithm,which causes degradation of tracking performance.In order to alleviate the problem that the late iterations of the sparrow search algorithm are prone to fall into local extremes,a Gaussian variational operator with a greedy criterion is introduced to enhance the ability of the algorithm to leap out of the local space,and an improved sparrow search algorithm is obtained.The improved sparrow search algorithm optimizes the particles after sequential importance sampling,so that most of the particles are concentrated in the highlikelihood region and a few particles are retained in the low-likelihood region to solve the problems of degradation of particle weights and loss of diversity.In order to verify the tracking performance of the algorithm,tracking simulation experiments are simulated for two cases:straight-line path and back-shaped path.The experimental results show that the mean square error of the proposed tracking algorithm is 0.5876m and 1.5556m in the straight-line path and back-type path,respectively,which effectively alleviates the problem of particle weight degradation and achieves the goal of improving the tracking accuracy of the node to be measured,and has certain reference value for accurately realizing dynamic target tracking in practical environments.