Multimodal sensors data fusion for improving indoor pedestrian localization

A large-scale hospital with several floor levels and confusing corridors is too complex for visitors to find their way to the destinations. The indication signs are set to guide people to their destinations. However, the names of units are too confused to put into effects. A navigation system decomposed to localization system and route planning system can assist people find their way. The localization of moving subject is an important information for a navigation system. In this work, an indoor localization system based on inertial sensors and light sensor in smartphones is proposed for accurate localization information. The results of proposed system demonstrate that the positioning accuracy is improved with the combination of inertial sensor and light sensor information. The step length modification and sub-edge heading reset can reduce the accumulative error generated during walking.