NA-AEKF: A NLOS-Aware Adaptive Extended Kalman Filter for Robust Indoor Localization

Accurate indoor localization remains challenging due to frequent Non-Line-of-Sight (NLOS) conditions, which degrade the performance of Ultra-Wideband (UWB)-based positioning systems. Traditional filters like the Extended Kalman Filter (EKF) and its variants assume ideal Line-of-Sight (LOS) scenarios and struggle in complex indoor environments. To address this, we propose the NLOS-Aware Adaptive Extended Kalman Filter (NA-AEKF), designed to improve robustness and accuracy under dynamic conditions. NA-AEKF uses machine learning to classify real-time measurements as LOS or NLOS, adjusting filter parameters accordingly. This integration improves adaptability to changing conditions and mitigates NLOS-related errors. Simulations show that NA-AEKF outperforms EKF, the Unscented Kalman Filter (UKF), the Adaptive EKF (AEKF) and the Improved adaptive EKF (I-AEKF) [1], achieving higher precision in complex indoor settings.