Research On Indoor Positioning Technology Based On Ultra WideBand

With the development of the Internet of Things technology,positioning technologies are becoming more and more diversified,and location services have become an indispensable basic service in people's daily lives.Among them,the global positioning system and the Beidou satellite navigation system provide people with accurate location information in an outdoor environment,but in an indoor environment,the satellite signal attenuation is severe and cannot meet people's needs for indoor highprecision positioning.Ultra Wide Band(UWB)technology has many advantages such as high accuracy,low power consumption,fast speed,large capacity,and strong security,and is very suitable for indoor precise positioning systems.However,the UWB signal has a short duration,non-line-of-sight and synchronization errors have a greater impact on UWB ranging and positioning,which limits the development of UWB indoor systems.This thesis analyzes the characteristics and technical advantages of UWB signals,summarizes the characteristics of common indoor positioning algorithms,and studies the channel quality characteristics of UWB signals.Aiming at the synchronization error in the UWB ranging process,this thesis proposes an N-element symmetrical two-way ranging(N-SDS-TWR)method,which combines the characteristics of high accuracy of SDS-TWR and high efficiency of N-TWR.First,let the node under test use the NTWR method to broadcast a signal to the surroundings and get a response,and then add a broadcast response of the node under test based on the principle of the SDS-TWR method,and finally calculate the location of the target node based on the signal roundtrip time.An optimized calculation method is proposed in the ranging process to reduce the impact of clock drift on positioning accuracy.Through experiments,the N-SDSTWR method is compared with the N-TWR method,and it is proved that the N-SDSTWR has a small error and good stability.In addition,in view of the positioning error caused by the Not Line of Sight(NLOS)environment in UWB positioning,this thesis proposes a NLOS identification method based on an improved random forest algorithm.This method is based on an improved random forest model to model the NLOS positioning environment.This method first uses the Relief F feature selection algorithm to select the features that have a greater impact,and then determines the weight of each decision tree according to the classification accuracy of the decision tree in the random forest algorithm,and makes full use of the classification effect of the decision tree to realize the identification of the NLOS path.Experiments show that the accuracy of the NLOS identification method based on the improved random forest algorithm reaches89.84%,which lays the foundation for the improvement of subsequent positioning accuracy.Finally,this article uses the DWM1000 chip to build an indoor positioning system based on the N-SDS-TWR method and the NLOS identification method,develops the hardware and designs the software,and realizes the indoor target positioning and display.Through field positioning experiments,the feasibility of the system was tested,and the effectiveness of the N-SDS-TWR method and the NLOS identification method was verified.