Researches On UWB Indoor Localization Algorithm And Anchor Node Deployment

With the rapid spread of smart devices,the acquisition of high-precision indoor location information has attracted widespread attention in the field of positioning.Ultra Wideband(UWB)positioning technology has become one of the commonly used indoor positioning methods due to its high temporal resolution,low power consumption and strong antiinterference capability.However,the accuracy of UWB position is easily affected by shielding and clock drift in the complex indoor environment,which makes it difficult to meet the requirements of high-precision indoor positioning.Therefore,this thesis addresses the problem of improving the accuracy of ultra-wideband positioning in complex indoor environments and conducts a detailed study of positioning algorithms.This thesis summarises the advantages of UWB positioning technology in indoor positioning scenarios based on the current research on indoor positioning and ultra-wideband positioning technology.Based on the advantages and disadvantages of existing UWB positioning technologies,this thesis selects the Time of Arrival(TOA)positioning algorithm and adopts bilateral bi-directional ranging to suppress clock drift and solve the problem of time synchronization,and also selects the evaluation index of positioning accuracy in this thesis.Firstly,the analysis is carried out for the distance measurement error of the UWB device indoors.To mitigate the distance measurement error between the anchor node and the tag,eight features are extracted based on the Channel Impulse Response(CIR)to improve the accuracy of distance measurement.An e Xtreme Gradient Boosting(XGboost)distance measurement accuracy enhancement algorithm is designed,which can be used to identify Line of Sight(LOS)and Non Line of Sight(NLOS),and the least squares method is used for coordinate estimation.The ranging dataset is collected in a real indoor positioning scenario and a data collection scheme is designed for use with the UWB device in this thesis.The CIR feature parameters of the three sides are extracted in a minimal positioning system with three anchor nodes and one tag node,and static fixed-point ranging experiments and static fixedpoint positioning experiments are conducted to verify the performance of the XGboost ranging accuracy boosting algorithm for indoor positioning accuracy improvement.Secondly,the optimization problem of anchor node deployment for indoor positioning is investigated in conjunction with Geometric Dilution Precision(GDOP).The relationship between positioning accuracy and GDOP is investigated,and the effects of the number of anchor nodes,anchor node rule deployment and anchor node height deployment on positioning accuracy are analysed in conjunction with the GDOP mean value,so as to obtain the optimal anchor node deployment model for indoor positioning systems.In order to solve the anchor node deployment problem for heterogeneous indoor structures,the variational operator,variational strategy and crossover operator of the differential evolution algorithm are improved.The improved differential evolution algorithm is used to solve for the optimal anchor node deployment,and simulation experiments are carried out under the coordinate constraints.The results show that the improved differential evolutionary algorithm is faster and more applicable for solving the optimal deployment of anchor nodes.Finally,based on the findings of the XGboost ranging accuracy-boosting localisation algorithm and the anchor node deployment simulation study,a single-tag dynamic localisation experiment is carried out in an actual indoor scenario.The particle filtering algorithm based on XGboost range accuracy enhancement is introduced to compare the performance of particle filtering and Kalman filtering on a non-linear system,and the influence of the number of particle filtering particles and resampling method on the experimental results is analysed.To address the problem that dynamic positioning trajectories tend to fluctuate,a two-stage filtering model is established according to the motion state of the trolley.Dynamic positioning experiments in line-of-sight and non-line-of-sight conditions,the experimental results show that the particle filter fusion localisation algorithm based on XGboost range accuracy enhancement has higher localisation accuracy.