An non-line-of-sight detection algorithm on the mobile phone with 3-axis accelerometer sensor

When the mobile device user meet the obstacles in the walking, non-line-of-sight (NLOS) errors happen in Time Difference of Arrival (TDOA)-based location system, which make the tracking accuracy low. Existing proposals failed when the gap between a NLOS condition and a line-of-sight (LOS) condition is not obvious. In this paper, we propose a new algorithm to solve this problem, which utilizes Chauvent´s criteria to detect a NLOS error, when the abnormal change of the difference between the user´s predicted and measured location is identified. The user´s predicted location information is estimated by the user´s step size information obtained from 3-axis accelerometer sensor in the mobile phone. Simulation results show that the proposed algorithm improved NLOS error detection about 17.3%.