Dynamic Radio Map Based Particle Filter for Indoor Wireless Localization

Indoor wireless localization is the basis of various location-based applications. So far, the majority of Radio Map based wireless indoor localization algorithms adopted the static radio map in which signal environment was regarded as stable over time, and didn’t make good use of continuous motion information of the target. In this paper a dynamic Radio Map based particle Filter for wireless indoor localization (DRMPF) algorithm is proposed, which combines the particle filter with Radio Map based positioning technology. By constructing spatial correlation model (SCM) based dynamic Radio Map, DRMPF uses some reference nodes to capture the real-time signal changes in the environment. Contiguous spatial correlation model breaks the limitation of traditional grid-shaped Radio Map, and converts the wireless indoor localization from classification problem into regression problem. It also reduces the training cost and algorithm complexity of the online localization stage. Extensive experiments demonstrates that the proposed SCM based Radio Map model has good time generalization ability. Compared with the static Radio Map, the DRMPF algorithm improves the positioning accuracy by about 20%, which demonstrates a good ability to adapt to the environment changes.