Research On The Indoor Pedestrian Positioning Based On Inertial Measurement/WiFi/Building MAP

With the advent of the intelligence age,indoor positioning technology is becoming a research hotspot.In addition to providing positioning services in private space such as the family,indoor pedestrian positioning has a wide application prospect in high speed railway station,library and public indoor environment of sudden natural disasters.At present,integrated positioning by making full use of indoor information sources has becoming a focus of research.Global positioning system(GPS)is widely used in outdoor positioning because of its high coverage and high precision.With Beidou navigation system and Galileo navigation system are in succession ripe,the selection and accuracy of outdoor positioning is increasing.However,the signal attenuation of this kind of positioning system is serious in the indoor environment,which leads to a poor precision.Inertial sensor is a fully autonomous positioning sensor with the character of high integration,small volume,which make the indoor positioning technology based on inertial measurement is constantly developing.However,due to the accumulative error increased with time,the positioning system based on inertial sensors cannot provide accurate position information alone during a long time.Therefore,it is of great significance to study the integrated positioning fusion algorithm to combine other information sources.As a kind of infrastructure,WiFi has a natural advantage as a positioning information source because of its characteristic of signal transmission.The accuracy of indoor positioning based on WiFi is mainly affected by random errors.Therefore,indoor positioning based on WiFi can be regards as observation to limit the acumulative error of indoor positioning inertial measurement.WiFi fingerprint positioning is adopted in this paper.With intelligent routing is more and more widely used(intelligent routing according to the number of users to dynamically adjust the transmission power,thereby reducing power consumption).At the same time,the transmission power of ordinary routing is also floating in different time periods.Therefore,the posiitoning error of traditional WiFi fingerprint Library based on absolute strength will increase dramatically.The cost is greatly increased.if the fingerprint library is adjusted frequently.In view of the above situation,this paper proposes a WiFi fingerprint location algorithm based on double database.Based on the absolute strength of fingerprint library,the gradient strength fingerprint database is presented.Stability by using signal strength difference between adjacent regions with the adjacent area,and linked to the high dimension information,in improving the stability and positioning system,the adaptive online matching algorithm is proposed based on t-test to further improve the positioning accuracy,realize the effective switching between the two libraries.Based on the traditional weighted K nearest neighbor algorithm(WKNN)matching algorithm,the online matching algorithm based on GS-WKNN(Gradient Smoothness-WKNN)is designed by taking into account the smoothness of intensity gradient fingerprints to makes the algorithm more reasonable.These studies have effectively improved the accuracy of indoor positioningIn order to achieve the effective fusion of multi information sources,we choose different filters for the characteristics of different models.Based on continuous integration inertial location algorithm,Zero Velocity Update.(ZUPT)based on Kalman Filter(KF)is adopted to suppress cumulative inertial positioning error.However,the ZUPT-based pedestrian inertial position correction can only correct the pedestrian's velocity error and part of the attitude angle error,and has poor correction effect on the heading error during the pedestrian movement.Therefore,it is necessary to introduce other observational measurement assistant suppression.Simultaneously,WiFi,inertial positioning and architectural structures information are in different frequencies,Kalman filter only for information fusion requires changing the measurement matrix,this will not only affect the efficiency of information fusion and stability,but also reduce the positioning system therefore.In order to effectively integrate the position information of the information source provides,Particle Filter(PF)is adopted as the main second order filter.The Building Map can be well constrained on pedestrian positioning trajectory,and prevent the pedestrian path into the wall and the wall or too close.However,it brings negative effects,when a particle at all into the wall,the particle weight were assigned to zero,thus interrupting pedestrian positioning system,we improved the particle filter,and the information fusion algorithm based on Bayesian estimation of class two are presented to increase the accuracy of positioning.Finally,in a general indoor environment covered by WiFi,indoor pedestrian positioning experiments using MTi-G710 inertial measurement unit and wireless signal acquisition function are carried out to verify the correctness and effectiveness of the proposed algorithm.The experimental results show that the proposed algorithm can effectively reduce the location error and improve the accuracy of indoor pedestrian location.