Magnetic-beacon Localization Technology Based On Feature Vector

With the increasing popularity of satellite positioning technology,the accuracy of outdoor positioning services has basically met people's daily needs.However,due to the weak penetration of satellite signals and multipath effects,the positioning result is prone to large errors in the complex environment where there are a lot of obstacles hidden indoors,underground,so it is difficult to meet the task requirements.Therefore,the complex environment positioning technology is regarded as the key research direction and has great research potential.The magnetic positioning technology has always been a research hotspot.Compared with the traditional positioning method,the low-frequency magnetic field positioning has the characteristics of strong signal penetration,no multipath effect in the propagation process,and positioning error does not accumulate over time,so it is suitable for providing the positioning service in the complex environment.In this paper,the low frequency magnetic field positioning technology is deeply studied,and a positioning technology based on magnetic feature vector is proposed.Firstly,in order to describe the relationship between the magnetic field generated by magnetic beacons at any point in space and position,the magnetic field distribution model is studied.Aiming to solve the problem that the single magnetic dipole model has large error in the near-field region,a magnetic field distribution model of double magnetic dipoles is proposed refer to distributed magnetic field multipole model.Because the raw magnetic field data collected by the sensor contains the interference magnetic field in the environment,in order to accurately extract the sinusoidal magnetic field generated by the magnetic beacon,a sinusoidal magnetic field signal parameter identification algorithm is designed.Based on the trigonometric function fitting method,the spectrum analysis is introduced to improve the frequency identification accuracy and the sliding window structure is added to store the intermediate data which improves the computational efficiency of the signal identification algorithm.Then,based on the rotating magnetic field's feature vector immutability,a single magnetic beacon positioning algorithm is designed.Aiming at solving the problem that the same frequency magnetic field generated by the orthogonal solenoid is difficult to extract separately during the feature vector calculation process,a regional asynchronous cross-multiplication algorithm is designed which can directly calculate the feature vector by using the synthetic magnetic field generated by the magnetic beacon.Meanwhile,the algorithm restrains the fluctuation of the orientation caused by the sensor measurement noise and improves the stability of the positioning output.Finally,the error analysis of single magnetic beacon positioning algorithm is carried out to study the influence of various error factors on the positioning solution.According to the characteristics of relative pitch angle being insensitive to common error factors,a multi-magnetic beacon localization algorithm based on feature vector is proposed.However,the position constraint in the multi-magnetic beacon positioning solution is a nonlinear equation,so it is transformed into a nonlinear least square fitting.The Levenberg-Marquardt method is used for position optimization to achieve high robust positioning output.