An Empirical Study Of Pedestrian Localization Method Based On Smartphone MEMS

In indoor environment, the satellite signal is blocked by the wall, metal objects, and so on. Therefore, it leads to pedestrian localization failure. The dead reckoning algorithm is used for localization, because of no satellite signals and low-cost localization. However, traditional dead reckoning can’t eliminate accumulated error and no consideration of the pedestrian elevation.Aimed at above issues, a 3D localization algorithm for indoor pedestrian based on smartphone sensors (magnetometer, accelerometer, gyroscope and barometer) was proposed. On the basis of two-dimensional localization information obtain by pedestrian dead reckoning, pedestrian azimuth is calculated by fusing multi-sensor output and weighted azimuth combined with electronic compass is proposed to improve the calculation accuracy. Combined with the virtual landmarks layer and its matching algorithm, pedestrian behavior state detected by sensor output is used to match present location to the special locations. Thus, the localization error is eliminated with the way of resetting pedestrian initial location. Using the pedestrian elevation calculated by differential barometric altimetry method, three-dimensional localization information is obtained. The experiment is carried out to verify the effectiveness of the method on the smartphone platform. With the experimental results, the algorithm can not only provide accurate, continuous two-dimensional positioning information but also improve the resolving power of the pedestrian to the floor. Therefore, the algorithm can offer effective localization information for Location Based Services, such as pedestrian navigation, tracking.