Enhanced Radio-SLAM Algorithm Using Building Geometry Constraints

Wi-Fi-based radio-SLAM (simultaneous localization and mapping) estimates the positions of users and access points (APs) simultaneously in GPS-denied indoor environments. To address the observability degradation in radio-SLAM caused by using only relative distance measurements between APs and the user, this paper proposes a positioning method based on the extended Kalman filter (EKF) and global geometry constraints. This method first uses the relationship between signal strength and ranging in the free space path loss (FSPL) model to determine the visibility of the AP. Then, it establishes a globally constrained positioning framework by integrating AP’s observed visibility and estimated visibility (from the estimated results of the geometric collision detection of walls). Simulation results show the improvement of our positioning methods, compared with the traditional EKF, the proposed method improves user positioning accuracy (RMSE) by 34%, and AP location estimation accuracy by 37%, enhancing users’ perception capability in unknown environments.