Research On Indoor Location Algorithm Of Smartphone Based On Multivariate Information Fusion

Location information is playing an increasingly important role in People's Daily work and life.Location based services have wide application prospects in fire rescue,medical services,Internet of Things,intelligent perception and other aspects.In recent years,driven by the booming development of mobile Internet technology,smartphone products are constantly upgraded and the consumption scale is gradually expanding.The indoor technology based on smartphone is very suitable for large-scale promotion and application,which can benefit more related industries and has an unlimited development prospect.Currently,the most commonly used smartphone-based indoor positioning technologies are Wi-Fi,geomagnetism,Pedestrian Dead Reckoning and Bluetooth positioning.However,a single indoor positioning technology is often difficult to meet the positioning needs of most users presently,such as Wi-Fi indoor positioning.Due to existing Wi-Fi network is mainly used for communication rather than location,fingerprinting method is commonly used to realize location.Although this method has high positioning accuracy,it has the disadvantages of high system complexity and low scalability,so it is difficult to be popularized and applied.This topic focuses on the indoor positioning with low cost,low complexity and reliable positioning accuracy,and emphasis on Pedestrian Dead Reckoning and Bluetooth positioning technology.Aiming at traditional dynamic step length estimation model in PDR need offline data fitting model parameters and the problem of poor universality,this paper puts forward a kind of based on continuity and acceleration of the pace of online step length estimation method.It is estimated based on the absolute position online of the landmark step length parameters in the model to make the step model suitable for different users;Aiming at the problems of the traditional indoor location fusion algorithm,such as the inability to determine the initial position,the slow convergence speed and the lack of particle diversity,a positioning technology based on Bluetooth signal was proposed.The initial position was determined by the proximity method,and the Bluetooth ranging information was used as the particle filter observation value.Finally,the particles are reset according to the absolute position information obtained by Bluetooth peak landmarks and displacement compensation,so as to reduce the impact of particle degradation on positioning.The indoor positioning algorithm based on multi-source information fusion proposed in this paper uses Bluetooth and map information to assist PDR positioning on the basis of the principle of particle filter,making full use of the high short-term positioning accuracy of PDR and the non-parametric characteristics of particle filter algorithm.Through experiments,it is proved that the average positioning error of the proposed algorithm is within1 meter,and its implementation only requires the map information of the positioning area,smart phones and a small number of low-price and easy-to-deploy Bluetooth beacons,which provides an implementation algorithm for the positioning technology based on smart phones and has certain application value.