EHTrack: Earphone-Based Head Tracking via Only Acoustic Signals
Linfei Ge,Qian Zhang,Jin Zhang,Huangxun Chen
DOI: https://doi.org/10.1109/jiot.2023.3298412
IF: 10.6
2023-01-01
IEEE Internet of Things Journal
Abstract:Head tracking is a technique that allows for the measurement and analysis of human focus and attention, thus enhancing the experience of human-computer interaction (HCI). Nevertheless, current solutions relying on vision and motion sensors exhibit limitations in accuracy, user-friendliness, and compatibility with the majority of commercial off-the-shelf (COTS) devices. To overcome these limitations, we present, an earphone-based system that achieves head tracking exclusively through acoustic signals. employs acoustic sensing to measure the movement of a pair of earphones, subsequently enabling precise head tracking. In particular, a pair of speakers generates a periodically fluctuating sound field, which the user’s two earphones detect. By assessing the distance and angle alterations between the earphones and speakers, we propose a model to determine the user’s head movement and orientation. Our evaluation results indicate a high degree of accuracy in both head movement tracking, with an average tracking error of 2.98 cm, and head orientation tracking, with an average error of 1.83 degrees. Furthermore, in a deployed exhibition scenario, we attained an accuracy of 89.2% in estimating the user’s focus direction.
computer science, information systems,telecommunications,engineering, electrical & electronic