Landmark Detection Based On Behavioral Perception Assists WiFi-PDR Indoor Positioning

With the continuous popularity of smart phones and the progress of positioning technology,as well as the growing demand for location services,location-based services have become one of the essential functions of smart phones.Products based on location services provide convenient for our lives.However,in order to provide accurate location services,we must continue to improve smartphone positioning technology.The current outdoor positioning technology has gradually matured and been used in our daily life.In the outdoor environment,it mainly relies on the global navigation satellite system(GNSS)to provide location services for people's travel,such as car navigation,recommendation of nearby people and friends.Compared with outdoor positioning technology,the progress of indoor positioning technology is relatively slow.Since people spend most of their time indoors,location-based services in indoor environments have a wider range of application scenarios.This paper aims at the problems that WiFi positioning in current WiFi-PDR indoor positioning is susceptible to environmental interference and pedestrian dead reckoning(PDR)accumulated errors,etc.,and proposes a behavior-aware landmark detection assisted by WiFi-PDR indoor positioning method.This method reduces the positioning error of indoor positioning to a certain extent,and improves the accuracy of indoor positioning.The specific research content is as follows:(1)The research and experiment of a coarse-grained location detection method based on WiFi-PDR.Through comparative analysis of various indoor positioning technologies,the indoor positioning method based on WiFi-PDR has received extensive attention because of its low cost and strong universality.In this paper,the research and experiment of WiFi fingerprint positioning technology and PDR are carried out to obtain the coarse-grained location information of the user.The maximum positioning error is 1.8m.(2)Conducted research analysis and experimental verification of behavior perception technology.Aiming at the problem of behavior recognition based on smart phones,the behavior recognition technology based on convolutional neural network is used to recognize the four behaviors of pedestrian,walking,stationary,left-turning and right-turning,with an average recognition rate of 86.7%.(3)Focus on research on landmark detection methods based on behavioral perception.This paper proposes a landmark detection method based on behavior perception,which combines user behavior information and coarse-grained location information.When it detects that the user is stationary,turning left,turning right,etc.,the user's coarse-grained position coordinates are recorded and stored in the landmark information database in the form of <behavior status,coarse-grained position coordinates>,and then the clustering of the landmark information database is further processed,clustering coarse-grained location coordinates of the same behavior,the center of the class is the new landmark point,and recorded the landmark library.(4)Aiming at the problem of increasing of accumulated errors in the process of pedestrian track estimation,a pedestrian track correction method based on landmark detection is proposed,which corrects the indoor pedestrian track estimation based on the detected landmarks.In addition,according to the different behaviors of users,the step size estimation method is adjusted to reduce the error as much as possible and improve the accuracy of indoor positioning.Finally,the experimental results show that the behavior-aware landmark detectionassisted indoor positioning method proposed in this paper can greatly reduce the dependence of indoor positioning methods on the indoor environment,effectively reduce indoor positioning errors,and improve indoor positioning accuracy.