Answers are grounded in this paper and cite their sources.
Most human tracking techniques based on channel state information (CSI) require array antennas or high-frequency CSI acquisitions. Particularly, in a method having an arbitrary antenna arrangement, it is necessary to extract information about human gait that appears at 20–60 Hz. Therefore, these techniques require super-Nyquist sampling frequencies, above 120 Hz, for CSI acquisition. These restrictions make it challenging to achieve device-free human tracking using a wide variety of commodity WiFi devices within a competitive environment with other transmitters. To address this problem, a device-free human tracking method is proposed, using sub-Nyquist-sampled CSI, for example CSI acquired at approximately 10 Hz, with an arbitrary antenna arrangement. The key idea is to correspond the features extracted from CSI by utilizing statistics to the relative shift of the target position, enabling aliasing-free estimation of the target position even from sub-Nyquist-sampled CSI. Assuming that the CSI consists of two-ray of static and dynamic reflected waves, we used two features that corresponded to the relative shift of the target position. These features were inversely proportional to the reflected path length. Experiments showed that it is possible to estimate the walking direction with approximately 80% accuracy and tracking with decimeter-level RMSE, thereby confirming the effectiveness of the proposed method.
No takes yet. Share an insight, caveat, or question.
Tamai et al. (2024) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: