| Title | RAMTEL: Robust Acoustic Motion Tracking Using Extreme Learning Machine for Smart Cities |
|---|---|
| ID_Doc | 41876 |
| Authors | Liu, Y; Zhang, WX; Yang, Y; Fang, WD; Qin, F; Dai, XW |
| Title | RAMTEL: Robust Acoustic Motion Tracking Using Extreme Learning Machine for Smart Cities |
| Year | 2019 |
| Published | Ieee Internet Of Things Journal, 6, 5 |
| Abstract | Motion tracking is attractive in what concerns a smart city environment, where citizens have to interact with Internet of Things (IoT) infrastructures spread all around one particular city. Motion tracking is important for smart services and location-based services in smart cities, since it provides natural ways for users to interact with the IoT infrastructures, such as the ability to recognize of a wide range of hand motion in real-time. Compared with dedicated hardware devices, ubiquitous devices with reliable speakers and microphones can be developed to achieve cheap acoustic-based motion tracking, which is appropriate for low-power and low-cost IoT applications. However, for complex urban environments, it is very difficult for acoustic-based methods to achieve accurate motion tracking due to multipath fading and limited sampling rate at mobile devices. In this paper, a new parameter called multipath dispersion vector (MDV) is proposed to estimate and mitigate the impact of multipath fading on received signals using extreme learning machine. Based on MDV, a robust acoustic motion tracking (RAMTEL) method is proposed to calculate the moving distance based on the phase change of acoustic signals, and track the corresponding motion in 2-D plane by using multiple speakers. The method is then proposed and implemented on standard Android smartphones. Experiment results show, without any specialized hardware, RAMTEL can achieve an impressive millimeter-level accuracy for localization and motion tracking applications in multipath fading environments. Specifically, the measurement errors are less than 2 and 4 mm in 1-D and 2-D scenarios, respectively. |
No similar articles found.