Simultaneous localization and mapping in indoor positioning systems based on round trip time-of-flight measurements and inertial navigation

This paper describes the algorithm for positioning of both indoor base stations and pedestrians in nanoLOC (IEEE 802.15.4a) wireless network. The algorithm uses three sources of information: time-of-flight ranging results between a mobile tag and base stations, the trajectory obtained from an embedded inertial measurement unit, and the map of the building. The iterative algorithm consists of two consequent interconnected stages. The particle filter is used to localize the mobile object on the map, whereas the discrete Bayes filter (grid filter) is applied for estimating the positions of base stations. Time-of-flight non-line-of-sight error distribution function was used as a likelihood in the discrete Bayes filter. Simulation shows that the described SLAM algorithm can be effectively used for setting the initial positions of base stations on the predefined map to save the installation time without significant decrease of localization accuracy. The proposed technique was applied in RealTrac local positioning technology.