Research On Indoor Positioning Based On Inertial Navigation

In the 21st century,with the continuous progress of science and technology,people's demand for accurate location information is increasing.At present,the positioning method can be roughly divided into outdoor positioning and indoor positioning.At present,outdoor positioning has developed very mature.GNSS(global navigation satellite system)such as GPS(global positioning system)and BDS(beidou positioning system)are widely used in daily life,and they can roughly meet people's outdoor positioning needs.However,in the indoor space with narrower space and more complicated situation,GPS signal is very weak due to the block of the building and cannot realize the positioning function.However,in most cases,people move indoors,such as in large shopping square,airport,hospital,parking lot and other places.Such lifestyle inevitably increases people's demands for indoor navigation and positioning.Nowadays,indoor positioning methods mostly adopt some new technologies,such as bluetooth,radio frequency identification,and pedestrian track estimation,etc.However,different technologies are more or less restricted in different environments.In this paper,an improved inertial positioning algorithm is proposed.Based on the pedestrian dead reckoning algorithm,a series of improvements are made based on the disadvantages of low positioning accuracy,poor performance and low stability of the mobile phone sensor.First of all,reliable data can be obtained through the establishment of kalman filter model to process the built-in multi-source sensor data of the mobile terminal.However,this method depends on the sensor performance,and phones with poor performance have poor results.Therefore,how to use local technology to determine the location of people or objects becomes a challenge.Then,the walking path is predicted by the a-star algorithm assisted by map matching,so as to solve the problems such as multiple inflection points of the path and excessive searching time,so as to obtain the optimal route.Secondly,the traditional PDR algorithm is based on the solution of three parts(step number detection,step length detection and direction judgment).First,the walking step frequency was determinedby the combination of kalman filter algorithm and dynamic threshold judgment method.Then,a method was proposed to measure the step size based on photography and pass the step size value into the navigation interface to dynamically judge the walking step size.In addition,it was proposed to conduct gradient processing on the direction data to solve the problem of direction instability.Then,the region division method is used to adjust the trajectory error,solve the problem of particle passing through the wall,and make the track more smooth and accurate.Finally,the mobile phone APP interface was written to make it possible for the indoor inertial navigation function to run in real time on the mobile phone platform.The experimental results show that the improved algorithm eliminates the positioning inaccuracy caused by too much error,enhances the positioning stability and accuracy,and makes the inertial positioning a feasible positioning method.