Research On Indoor Positioning Of Smart Phone Based On PDR And Geomagnetism

With the universal use of intelligent smartphones,obtaining accurate location information is the key to solve the problem of increasing demand for indoor services,such as intelligent parking in large parking lots,accurate marketing,emergency rescue,health care and path navigation.In outdoor environment,Global Navigation Satellite System(GNSS)can easily obtain accurate position information,such as Global Positioning System(GPS)and Beidou Navigation Satellite System.However,due to the existence of roofs,walls and other buildings,the GNSS signal will be seriously weakened,so that it is not feasible to use GNSS positioning indoors.Besides,there is no indoor GNSS signal or unified standard for indoor maps.At present,the inertial measurement unit with built-in MEMS,that is,the mobile device composed of gyroscope,magnetometer and accelerometer,can provide independent solutions for pedestrian path planning in different environments.In addition,the geomagnetic field is suitable to be used as a kind of position fingerprint due to its rich information,convenient real-time matching,high precision,good stability,low cost and so on.In this paper,smart phones are used to carry out the following research.The main contents include three parts.Firstly,this paper will introduce many indoor location methods based on geomagnetic field,such as the method based on geomagnetic fingerprint,the method based on particle filter(PF)and the method based on machine learning.Meanwhile,the performance of different indoor positioning methods is compared and analyzed by using the experimental data.From the mean value,variance and root mean square error of the coordinates estimated by geomagnetic fingerprint algorithm and machine learning algorithm,it can be seen that indoor positioning using land magnetic field is feasible.Secondly,this paper aims at the problem that the fingerprint acquisition method cannot build the database quickly in the indoor positioning in the geomagnetic field which lead to few matching features,unable to match quickly and low accuracy of heading angle estimation.Therefore,this paper adopts the Gauss-Newton iterative method combined with the trajectory from the PDR measurement to match and locate the magnetic field contour is aim at improve the distinguishability of the single point magnetic field fingerprint based on the fast approximation of the magnetic field gradient.MFS is used as the reference position,and EKF is used to control the position drift error of inertial navigation,so as to further increase the accuracy of trajectory profile.Thirdly,this paper proposes a pedestrian dead reckoning algorithm on the basic of strapdown inertial navigation system according to the updating of quasi-static attitude,gait model and motion constraints as pseudo-measurement information.A walking test experiment is designed to compare and analyze the performance of C-INS and the existing E-PDR in position estimation and heading estimation.The experimental results display that the RMS position estimation error of geomagnetic matching location algorithm is 1.27 m.The statistical positioning errors of the C-INS method at the three positions are 1.15 m,1.84 m and 1.73 m respectively,and the RMS heading error keeping positions are 3.74,6.52 °and 16.86 °respectively.The accuracy of the C-INS method is better than that of the E-PDR method.It is proved that the robustness of the C-INS method is better.There are in total 9 tables,42 pictures and 70 references in this paper.