Research On Indoor Positioning And Tracking System Based On Fusion Of Acoustic Signal And IMU

With the popularization of mobile devices and the ever-increasing interior activities,the demand of location-based services(LBS)for indoor positioning has grown significantly in recent years.On the one hand,satellite positioning systems represented by the Global Positioning System(GPS)are not available in indoors because of the blockage of buildings.On the other hand,although the indoor positioning technology has been extensively studied and developed in the past two decades,there is still no widely applicable high-precision indoor positioning solution due to the influence of unfavorable factors such as the complex and varied indoor environment.Due to numerous limitations of electromagnetic wave,new localization approaches based on acoustic signal as well Inertial Measurement Unit(IMU)have drawn more attention recently for the advantages of low-cost,high-accuracy and the compatibility with mobile devices.This thesis focuses on the research of the indoor positioning and tracking system combining TOA(Time of Arrival)localization and IMU dead reckoning for pedestrian based on the mobile platform of smartphones.Hierarchical architecture is employed to fuse the TOA and IMU measurements and the algorithms have been realized and verified by experiments.The main work of the thesis is as follows:A two-step TOA estimation method using adaptive thresholds has been proposed on the basis of the indoor acoustic channel model and the study of common TOA estimation methods.Simulation results shows a better performance of the proposed method than previous algorithms.Localization algorithm for one-time observation and Bayesian filtering algorithms for continuous observations are developed for TOA pseudo-range measurements.To avoid fluctuations of positioning results,the localization algorithm is utilized for initialization and a Cubature Kalman Filter(CKF)is implemented to estimate the target position recursively and obtain positioning results between large time intervals.Step-and-Heading System(SHS)is implemented to generate step length and heading information by utilizing IMUs embedded in smartphones and thus the detailed continuous track between TOA localization can be obtained.To handle the influence of casual posture of smartphones,rotation matrix is used to optimize the raw data of acceleration and angular velocity.TOA localization,SHS subsystems and map data are fused by particle filter in order to improve the localization and tracking accuracy.Taking the nonlinear characterization of pedestrian motion into account,map(floor plans)data is utilized to develop the pedestrian movement model including nonlinear effects.By using cascaded estimation architecture,TOA filtering output is used as the input of the fusion filter,lower level CKF and upper level particle filter cooperate to reduce the computation complexity effectively.To deal with the problem of asynchronous measurements caused by lower update rate of TOA localization compared to walking frequency,an asynchronous weight update method has been proposed and simulation results has proved the validity of the method and the fusion algorithm.Based on the study of various algorithms,an indoor location and tracking system has been designed and realized utilizing Field-Programmable Gate Array(FPGA),smartphone based clients and remote server.The base station is implemented with FPGA,a smartphone APP and a server software have been developed for data collection and following processing.Experiments have been conducted in indoor scenarios and the proposed system yields about 0.5m accuracy of continuous tracking under low update rate of TOA localization,which further proves the validity of the proposed location and tracking system.