Research On Indoor Positioning Algorithm Based On Bluetooth And Inertial Navigation Fusion

With the development of the Internet of Things and the iteration of smartphones,mobile phone-based location navigation and peripheral services have received more and more attention.However,the complex motion poses of pedestrians and the limitations of a single indoor positioning technology have brought great challenges to high-precision indoor navigation.To this end,this thesis uses smartphones as a tool to conduct research on the pedestrian multi-motion posture positioning and indoor fusion positioning technology.The work carried out includes but is not limited to the following:1)Read and analyze the literature related to indoor positioning,and summarize the development status of wireless positioning technology as well as autonomous positioning technology.In view of the advantages of Bluetooth such as low power consumption and long standby time,this thesis first studies and summarizes the indoor characteristics and layout methods of IBeacon,and then improves the geometric indoor positioning method based on range measurement to improve the Bluetooth positioning accuracy.2)To address the problem of easy misjudgment and inaccurate step length estimation for step frequency detection in the current pedestrian inertial navigation in practical environment,an adaptive step length estimation algorithm based on motion state recognition is proposed.Firstly,a single-peak step detection algorithm based on dynamic threshold constraints is used to complete accurate step detection during the multi-state motion of a localised person holding a smartphone,on the basis of which an adaptive step length estimation algorithm is proposed.The algorithm collects the ensemble acceleration matrix for different motion states in the offline phase,and uses an improved dynamic time planning(SDTW)algorithm in the online phase to identify motion states based on the ensemble acceleration,and thereby selects the corresponding step length estimation model for PDR localisation.The experimental results show that the proposed adaptive step estimation algorithm based on motion state recognition has an average positioning error of 1.41% under various motion states,which significantly improves the positioning performance compared with the traditional pedestrian inertial navigation positioning method.3)A particle filtering(PF)fusion positioning algorithm based on location particle weight approximation is proposed to address the problem of excessive computational effort and difficulty of effective combination of Bluetooth and PDR fusion positioning algorithms on the smartphone side.The algorithm is suitable for indoor positioning environments with few base stations.The process of updating the location particle weights includes the determination of the location validity of IBeacon and the handling of special cases,and the particle weights are used to correct the accumulated errors.The experimental results show that the maximum positioning error of the fusion positioning algorithm proposed in this thesis is 1.48 m and the average positioning error is less than 1m,which better meet the needs of indoor precise positioning.4)In view of the limitations of the current indoor positioning 2D map,this thesis designs a software system and builds an indoor 3D map to import the front-end software for display.The positioning software system includes registration and login,3D map,compass and floor switching functions,which can realize indoor navigation display under the multi-motion state of pedestrians.Compared with the traditional indoor positioning map,it is more intuitive and clear.