Accurate tracking in NLOS environments using integrated IMU and fixed lag smoother

Multi-sensor data fusion using Inertial Measurement Units (IMUs) is a promising technique for improving the performance of positioning systems. However, the performance of conventional sensor fusion algorithms based on the Kalman Filter (KF) is compromised in indoor environments due to non-line-of-sight (NLOS) propagation. In this paper, we propose a semi-real time tracking algorithm which uses a fixed lag smoother for sensor fusion and achieves high accuracy in NLOS environments. The computational complexity of the algorithm is taken into consideration and is reduced by decreasing the operating rate of the smoother. The performance of the proposed algorithm is validated experimentally using a real indoor positioning platform. It is shown that the 90th percentile positioning error for a pedestrian is reduced by 42% using the proposed semi-real time tracking algorithm with 10 s lag, compared with using a KF-based real time tracking algorithm.