Effective Audio Signal Arrival Time Detection Algorithm for Realization of Robust Acoustic Indoor Positioning

Abstract

To achieve high precision and robust positioning on mobile devices, we propose a novel audio signal arrival time detection algorithm consisting of both coarse and fine searches. A coarse search uses the cosine theorem to extract the audio data segment (ADS) from the received signal of a mobile device, and a fine search then uses the waveform characteristics of the autocorrelation function of the source signal to estimate the audio arrival time (AAT) from the ADS. An indoor threshold determination experiment, a static positioning experiment, and a moving target positioning experiment were carried out on the developed acoustic indoor positioning system (AIPS) in a hall. The performance of the proposed audio detection algorithm was compared to two other typical detection algorithms. The results of the static positioning experiment show that, in a nonline of sight (NLOS) environment, 90% of the positioning results estimated by the proposed algorithm have an error of less than 0.50 m, whereas only 56% and 6% of the positioning results estimated by the other two classical algorithms have errors of less than 0.50 m. In the target-moving positioning experiment, the positions estimated by the proposed algorithm are closer to the true trajectories and have fewer (less than 0.5%) abnormal positioning errors than the other two classical algorithms. The experimental results also show that the reliability judgment of the positioning results can further improve the robustness of the system. The proposed method has an important application value for the implementation of a highly precise and reliable AIPS.