| Driven by technological development and social needs,location-based services such as vehicle monitoring,disaster relief,military,and pedestrian heat mapping have emerged.In an outdoor environment,the Global Navigation Satellite System(GNSS),as the most competitive positioning technology,can provide accurate and reliable navigation services for outdoor users around the clock,all day long.However,in indoor environments,GNSS is not available due to the inaccessibility of satellite signals caused by building obstructions.Although the traditional single-mode positioning technology can be a complementary means of location-based services,it is inefficient in using information about the target’s environment,resulting in low positioning accuracy.Multi-source information fusion localization is a method based on multiple information fusion strategies,which increases data redundancy through multi-information acquisition and obtains a more accurate and reliable localization performance to make up for the shortcomings of single localization technology.In this paper,the multi-source information fusion localization technology is studied in terms of the acoustic signal and Inertial Measurement Unit(IMU).Aiming at the problems of refraction,blocking,reflection,interference of other frequency signals when the sound signal propagates indoors,incidental errors in the positioning process,and cumulative errors of PDR positioning technology over time,the research of multi-source information fusion indoor positioning method is carried out.The main contents are:An indoor localization method based on acoustic signals is studied.To address the problems that acoustic signal localization is susceptible to multi-path effects and non-visual propagation,and easily generates outliers,a distance threshold detection method is proposed.The design and implementation of an indoor positioning system based on CHAN position estimation is carried out.Firstly,several mainstream acoustic localization algorithms are described;secondly,the acoustic signal-based indoor localization model is established in terms of localization signal design,indoor environment construction,signal pre-processing,and time delay estimation;finally,indoor localization based on Time Difference of Arrival(TDOA)and CHAN estimation is implemented in a simulated environment.The experimental results show that the mean square error of the CHAN algorithm is improved by 24% compared with TDOA under the same conditions.The indoor localization method of pedestrian heading estimation is investigated.A dynamically improved PDR algorithm is proposed to address the problem of cumulative error of traditional PDR positioning over time.Firstly,the sensor data is collected using the terminal device and then filtered;secondly,the improved algorithm addresses the problem of low accuracy of the traditional step estimation model,combines acceleration and the first two steps to update the step length adaptively according to the characteristics of pedestrian movement,and corrects the heading angle drift and static drift of different mobile terminals according to the filtered inertial data for the heterogeneity of different devices to obtain the direction of the current step forward after correction;finally,the experimental verification in the real environment shows that the error of the step estimation model of the improved PDR algorithm is less than 1.5%,the average error is 0.5 m,and the RMSE of the improved PDR algorithm reaches 0.6 m.The improved PDR algorithm not only improves the positioning accuracy but also solves the cumulative error problem of PDR.The fusion localization method of CHAN estimation and improved PDR is investigated.A CHAN-IPDR-ILS algorithm is proposed for the problem that single localization techniques are difficult to be balanced in terms of compatibility,pervasiveness,cost,and accuracy.The algorithm uses CHAN estimation to initialize the user position,estimates the current step size based on the first two-step sizes,corrects the heading direction of different mobile terminals due to heterogeneity,and then adaptively fuses acoustic signals and PDR for localization through dynamic equations of motion.In addition,a distance threshold detection method is designed for the problem of outliers in acoustic signal and improved PDR localization.The experimental validation in two different real environments shows that the cumulative density function of the 90 th percentile of the CHAN-IPDR-ILS method is0.3 m,the average error is 0.2 m,and the RMSE reaches 0.2 m.The proposed CHAN-IPDRILS method makes full use of the information of the target’s environment,and the measured localization results provide a high degree of universality for different devices and scenarios,and are compatible with smartphones in terms of low cost and high accuracy. |
