Adaptive Sage-Husa Kalman filter for 3D indoor positioning with multi-sensor fusion

In complex indoor environments, standalone bluetooth positioning often suffers from significant performance degradation due to non-line-of-sight (NLOS) propagation and environmental disturbances. Meanwhile, position estimates derived from inertial measurement unit (IMU) data are prone to cumulative drift over time, while barometer-based altitude estimates are sensitive to atmospheric pressure variations. To address these inherent limitations, this work presents a three-dimensional indoor positioning algorithm based on an adaptive Sage-Husa Kalman filter (ASHKF). The proposed method integrates bluetooth positioning, IMU, and barometer data to achieve accurate and robust pedestrian localization. Specifically, an adaptive noise covariance estimation scheme is incorporated into the Sage-Husa Kalman filter to dynamically adjust the reliability of each sensor’s measurements during fusion, thereby mitigating the impact of NLOS-induced errors and sensor drift simultaneously. Experimental results verify that the proposed multi-sensor fusion algorithm outperforms single-sensor solutions in both positioning accuracy and stability.