Research On Indoor Positioning Technology Based On Wi Fi And Inertial Sensor

Because of the coverage of GPS signals,outdoor positioning problems can be solved by GPS.However,with the increasing range of indoor activities,GPS signals can not cover the indoor environment,which brings great inconvenience to people's daily life.Such as rescue scene to find trapped people,large shopping malls to find shops and large parking lots to find parking spaces.Therefore,the research of indoor positioning technology is very meaningful and valuable.At present,with the increasing coverage of public WiFi and the popularity of smartphones with inertial sensors.WiFi technology and inertial sensor technology have become a hot research direction in the field of indoor positioning.Single positioning technology has limitations in indoor positioning.No positioning technology can meet the needs of indoor positioning in terms of positioning accuracy,stability and cost.WiFi positioning is absolute positioning without cumulative errors,but with low accuracy and unstable positioning results.PDR(Pedestrian Dead Reckoning,PDR)has high positioning accuracy,but it has the characteristics of cumulative errors.In this paper,the integration of two positioning technologies is taken as a breakthrough to solve the positioning problem for in-depth study.The main contents of this paper are as follows:(1)To solve the problem of WiFi fingerprint location,it is divided into two stages: offline and online.In the offline stage,fingerprint database is constructed and positioning model is trained.The data in fingerprint database is preprocessed to improve the accuracy of fingerprint database construction.In the online positioning stage,a fingerprint database positioning region partition method based on support vector machine algorithm is proposed,which reduces the range of the positioning model.The parameter optimization of SVR location algorithm based on imperial competition algorithm improves the accuracy of WiFi location.(2)Aiming at the heading angle problem of PDR positioning based on smart phone inertial sensor,Kalman filter is used to process the obtained heading angle,which effectively reduces the impact of indoor environment and human body shaking on heading angle measurement.Aiming at the problem of step size estimation,a step size model based on the relationship between known step frequency and step size is proposed.A small amount of experimental data can accurately estimate the step size and reduce the amount of data to establish the step size model.In step detection,dynamic threshold algorithm is used to calculate step number.(3)Fusion positioning system based on WiFi and PDR is realized.In planar positioning,WiFi positioning fluctuates greatly,but has no cumulative error and PDR short-term positioning accuracy,but has the characteristics of cumulative error.Combining the advantages of the two positioning methods,a displacement-based fusion algorithm is proposed for positioning.The experimental results show that the proposed fusion positioning method can reduce positioning fluctuation,eliminate cumulative errors and improve positioning accuracy.In stereo positioning,aiming at the problem that the floor judgement method based on mobile phone pneumatic sensor needs additional ground reference pneumatic hardware equipment,this paper proposes a floor judgement method based on dynamic reference without adding hardware equipment.The experimental results show that the proposed floor judgement method can accurately judge the floor.