ASE: Practical Acoustic Speed Estimation Beyond Doppler via Sound Diffusion Field

TL;DR

ASE addresses the insufficiency of Doppler-based acoustic speed estimation by introducing a high-rate OTDM scheme and a diffusion-field model that enables full-speed estimation from a single microphone. By treating indoor sound propagation as a diffuse field and aggregating multipath information, ASE derives speed from the spatial correlation of the sound field rather than radial Doppler shifts. The system combines Kasami PN probing in an inaudible band, OTDM-based CSI-rate boosting, motion detection via Zero Crossing Count, and frequency-aligned, prominence-weighted subcarrier fusion to achieve robust, real-time speed estimation. Evaluations on commodity hardware in diverse indoor settings show mean errors around 0.13 m/s with high detection rates, outperforming DFS-based approaches and enabling practical applications in HAR, fall detection, and gait analysis.

Abstract

Passive human speed estimation plays a critical role in acoustic sensing. Despite extensive study, existing systems, however, suffer from various limitations: First, the channel measurement rate proves inadequate to estimate high moving speeds. Second, previous acoustic speed estimation exploits Doppler Frequency Shifts (DFS) created by moving targets and relies on microphone arrays, making them only capable of sensing the radial speed within a constrained distance. To overcome these issues, we present ASE, an accurate and robust Acoustic Speed Estimation system on a single commodity microphone. We propose a novel Orthogonal Time-Delayed Multiplexing (OTDM) scheme for acoustic channel estimation at a high rate that was previously infeasible, making it possible to estimate high speeds. We then model the sound propagation from a unique perspective of the acoustic diffusion field, and infer the speed from the acoustic spatial distribution, a completely different way of thinking about speed estimation beyond prior DFS-based approaches. We further develop novel techniques for motion detection and signal enhancement to deliver a robust and practical system. We implement and evaluate ASE through extensive real-world experiments. Our results show that ASE reliably tracks walking speed, independently of target location and direction, with a mean error of 0.13 m/s, a reduction of 2.5x from DFS, and a detection rate of 97.4% for large coverage, e.g., free walking in a 4m x 4m room. We believe ASE pushes acoustic speed estimation beyond the conventional DFS-based paradigm and inspires exciting research in acoustic sensing. Code is available at https://github.com/aiot-lab/ASE.

Figures (28)

• Figure 1: ASE vs. DFS. Walking speed $\vec{v}$ can be decomposed to radial speed $v_r$ and tangent speed $v_t$. DFS can only capture radial speed $v_r$, but fails to capture $v_t$. Conversely, ASE can capture both $v_r$ and $v_t$, a complete estimation of $v$.
• Figure 2: OTDM Scheme. (a) Given $s_1(t)$, the channel is estimated by frame. (b) Given orthogonal sequences $s_1(t)$ and $s_2(t)$, we can get two channel estimations at the same time. (c) Orthogonal sequences $s_1(t)$ and $s_2(t)$, with $s_2(t)$ delayed by $\Delta d$, are modulated into one sequence $s(t)$. $H_1(t)$ and $H_2(t)$ estimated from the two sequences are concatenated into $H(t)$ alternately.
• Figure 3: Sound Diffusion Model. (a) The sound wave traverses while an object is moving $a \rightarrow b$. (b) The sound wave is diffusive in all directions by irregular reflectors.
• Figure 4: Diffusive Scattering in the room. (a) Illustration of diffusion scattering and specular reflections. The acoustic waves not only experience specular reflections, but also diffusive scatterings on any non-flat planes. (b) A simulation result of how an acoustic wave is reflected and scattered in a common room. The black point emits the acoustic rays with a frequency range of 17 kHz- 20 kHz. By employing the rich multi-path semantics, these scatterers are creating "virtual speakers", thus creating different views of speed estimation.
• Figure 5: ACF curves for two different speeds and the theoretical function for $v_0 = 1m/s$.