Acoustic Indoor Localization based on Range and Relative Velocity in Non-Line-of-Sight Environment

This paper investigates a problem of acoustic indoor localization in dense Non-Line-of-Sight (NLOS) environments by using range and relative velocity measurements obtained from received acoustic signals, without the need of additional sensors. The relationship between adjacent positions to be estimated is established by introducing velocity measurements and assuming a first-order motion model. This reduces the number of required Line-of-Sight (LOS) anchors. The principle of the range and relative velocity-based localization method is systematically studied and a basic solution based on a Least Square estimator (LSE) and its closed-form has been proposed. In order to achieve precise localization in dense NLOS environments, a multi-position joint estimation (MPJE) method is proposed. This method involves jointly estimating multiple positions in a short-time position sequence and is solved using a Levenberg-Marquardt (LM) algorithm. A LOS measurements redundancy metric is proposed to balance the localization accuracy and algorithm time cost by adjusting the sequence length. The results obtained from numerical simulations and experimental investigations demonstrate that the proposed range and relative velocity-based localization method outperforms the conventional range-based methods. The number of required LOS anchors for accurate localization is reduced from 3 to 2 for 2D positioning. Additionally, the proposed multi-position joint estimation method requires only 2 LOS anchors and permits the use of only 1 LOS anchor for a short period. This enables its application in dense NLOS environments with a low anchor deployment density, thereby increasing its scene adaptive capability and promotional value.