Research And Implementation Of BLE/MEMS Fusion Positioning Technology

With the development of science and technology,positioning technology is applied to people’s daily life to meet various needs.The Global Positioning System(GPS)has been able to locate meters or even centimeter-level accuracy on the earth.Applications such as car navigation,car taxi software,group purchase software,and Wechat all require position positioning as the basic support technology.However,due to the construction of a modern city,skyscrapers and large buildings block the satellite signals,causing multipath effects and even signal loss,seriously affecting satellite positioning accuracy and cannot provide positioning services in the interior.Nowadays,a large number of large-scale indoor public places such as shopping malls,airports,railway stations and underground garages have emerged,leading to more and more people’s attention to the positioning technology applied indoors.In the paper,we design a indoor positioning scheme based on Bluetooth Low Energy(BLE)and Micro-Electro-Mechanical System(MEMS)data fusion.Through intelligent terminal application development,hardware design and software program to achieve the indoor positioning system.In this paper,indoor positioning system based on BLE and MEMS data fusion is mainly composed of four parts: fingerprint positioning module,sensor positioning module,height solution module and fusion positioning module.In the fingerprint location module,the fingerprint location subsystem is optimized mainly from the establishment of the fingerprint database dynamically,the data preprocessing and the updating of the fingerprint database.Establish the initial fingerprint database by dynamic walking to reduce the workload.The large noise found by the outliers detection and then removed.Update fingerprint database with crowdsourcing data.In the sensor positioning module,this paper mainly studies pedestrian gait information to realize the estimation of walking speed,and combines the quaternion and Extended Kalman filter(EKF)to obtain the more optimized heading angle information.Combined with the speed and heading angle information to calculate the pedestrian trajectory.In the height estimation module,the height estimation is mainly conducted by studying the barometer.Because of local air pressure by the airflow,temperature,season and other factors,it cannot be calculated directly by the altitude formula.This article combines barometric pressure trends,map information and step size to complete cross-floor height estimation.In the fusion positioning module,in order to solve the problems of the cumulative error of the MEMS sensor positioning and the jitter error of the BLE fingerprint positioning.Taking the result of fingerprint positioning as the observation,the recursive trajectory of the sensor is used as the predictor and the result is obtained by inserting the extended Kalman filter.Finally,this thesis uses hardware and software to build an experimental platform and implements the BLE / MEMS fusion positioning system.Through the actual scene positioning test results show that the average location error of fingerprint positioning subsystem can reach 2 meters after updating and optimizing.The average error of fusion positioning can reach within 1 meter.The average error of height estimation across the floor is 0.35 meters within.The system in the long time positioning test process remained stable,is conducive to the actual project promotion and application.