Research On Indoor Localization Based On Intelligent Terminal Integrated Inertial Navigation System

In recent years, mobile devices have saw a significant development and the number of people’s daily lives surfing the Internet with their mobile terminals is growing. Location based service has changed lives unconsciously causing interior space positioning technology become a research hotspot. Currently, outdoor positioning technology like GPS (Global Positioning System) has been relatively mature, however in the interior satellite signal can’t penetrate due to the walls interference which makes it impossible for GPS localization in these areas. After did lots of research on indoor positioning technology, this paper focused on the magnetic field proposing a precise positioning algorithm and realized the system.Firstly, in order to deeply understand indoor magnetic field property, this paper conducted a variety of experiments in real scenarios to deeply analyze the geomagnetic of indoor space-varying, time-varying, device-varying and direction-varying. Studied the feasibility of indoor localization using geomagnetic and raised the challenges and explored a new implementation.Secondly, this paper proposed an indoor localization based on optimized particle filter with turning model. The algorithm, using traditional indoor localization with particle filter to make a coarse-grained localization and turning model to make a fine-grained and precise positioning, effectively solved the problem of positional deviation, giving85percentile accuracy was about2m. And, applying a turning detection system based on hidden Markov Chain, which is independent with the gesture, as a driver. In addition, to solve the slow convergence as the whole space match, this paper given an initial localization algorithm that can make a quick convergence and to speed up the path match proposed a search algorithm based on multiple index balancing the search speed and accuracy.Finally, according to the proposed algorithm, we have realized the localization system called MagLoc and the sampler system named MagLocSampler. Deployed and tested in real environment, they given satisfied results.