Indoor pedestrian positioning algorithm with feature point detection and virtual path matching

Compared with the outdoor vehicle navigation system, the indoor environment is complex, the difficulty of positioning is greatly increased. It is effective to improve the positioning accuracy by combining the feature points of the indoor building plane information and the feature points in pedestrian walk. This paper chooses the intersection of the paths and the turning points of the pedestrian movement. The intersection of the paths can be given by the building information map. The turning points of pedestrian can be detected by the difference of the mean value of the azimuth (DMA). If the turning points are in the corridor, they can be matched to feature points and a virtual path algorithm is designed to match the trajectories. If the turning points are in the room, nothing is done for them. Thus a new indoor pedestrian location algorithm is obtained. This algorithm can effectively distinguish whether pedestrian is in the corridor or in the room. Combined with the virtual path, pedestrians in the indoor walking trajectory can be clearly demonstrated, which decreases the positioning error of strapdown inertial navigation and improves the user experience of indoor positioning navigation.