A Novel Method Locating Pedestrian With Smartphone Indoors Using Acoustic Fingerprints

Acoustic-based positioning method have high accuracy and good compatibility with commercial-off-the-shelf consumer devices. However, the traditional acoustic-based method, which utilizes the time-of-arrival (TOA) and time difference of arrival (TDOA), is highly affected by occlusion and non-line-of-sight (NLOS); limited signal transmitting stations can only support a small range of positioning; the first path of the acoustic signal is difficult to accurately detect in corridors and remote locations from the stations. To solve these situations, in this paper, a novel indoor localization method for smartphones based on acoustic fingerprinting is presented, which utilizes the difference of characteristics in the frequency domain of a predefined acoustic chirp signal at different positions. The innovations of this method include the design of a signal transmission scheme, detection of the first path according to the time-division-multiple-address (TDMA), and use of the PSD of signals in the frequency domain as the fingerprint feature. To verify the effectiveness, universality and performance of this method, three experiments were implemented in different environments, including a comparison experiment with Wi-Fi-based fingerprinting. The experimental results demonstrate that the proposed method can achieve accurate positioning performance, and the mean error is within 1 m in the mix of LOS and NLOS indoor environments.