While earables present a promising avenue for cardiac sensing, whether they may replace the stethoscope to perform heart sound (a.k.a. PCG) monitoring remains questionable. The latest effort attempts to generate PCG-like waveform out of in-ear audio collected via earphones, yet its data-driven approach does not seem to be grounded in the underlying physics. To this end, this paper introduces EarPCG, a system for continuous PCG monitoring leveraging physics-informed neural models. As opposed to the debatable belief that bone-conducted PCG appears within ear canal, EarPCG generates PCG waveforms from the (actually existing) photoplethysmography (PPG) waveforms conveyed via blood vessels. Arising from pressure variations induced by heartbeats, PPG can be mathematically described by a Partial Differential Equation (PDE). Therefore, solving this PDE inversely may reconstruct cardiac dynamics and in turn enable the generation of PCG waveforms with another PDE characterizing the pressure oscillations propagating through soft tissues. Pipelining the two PDE-solving neural models, EarPCG achieves accurate PCG monitoring from in-ear audio, while requiring minimal training. Our extensive experiments leveraging a custom-built prototype demonstrate the efficacy of our proposed system. Furthermore, we have conducted clinical trials, with clinicians reporting no perceptible difference between authentic PCG and the sounds reconstructed by EarPCG.

Millimeter-wave radar is a useful tool for Internet-of-Things (IoT) enabled indoor subject tracking, gesture recognition, or other human-computer interaction. However, existing proposals are post-processing techniques, depending on unreliable processed outcomes, often resulting in ghost points. In addition, current solutions mostly exploit wide-band (up to 4 GHz) mmWave radar. However, for indoor IoT applications, regulations limit the available bandwidth to a maximum of 250 MHz, rendering current solutions infeasible for practical deployment. In light of this, we propose a hybrid approach in a pre-processing manner. Our proposal differs from existing solutions that leverage posterior point clouds, but apply a processing algorithm to original range-Doppler profiles. Meanwhile, we leverage physical constraints to further smooth tracking trajectories. To enhance the network generalizability and tracking performance, we incorporate a curriculum learning strategy. We have implemented our solution on a commercial mmWave radar with 250 MHz bandwidth. Experimental results show that, in complex indoor multi-person scenarios, the average estimation errors of AoA, range, and speed are 0.1 rad, 0.4 m, and 0.17 m/s, respectively, with a tracking RMSE of 0.77m for multiple targets.

Many reconnaissance missions require a group of mobile terminals (such as soldiers, mobile robots, and unmanned boats) to jointly operate within a region which is far away from the command centre. When a critical event occurs and is detected by a terminal, it is often required for the terminal to upload some critical data to the command centre (or via the satellite). As the bandwidth of the upload link is usually low due to the long distance, uploading the critical data often has long latency. To reduce the latency, a feasible way is to utilize the nearby terminals’ idle uplinks to help with the upload process, which requires the terminal’s data to be disseminated to other terminals as soon as possible. This is a new dissemination problem because the data being disseminated is also partially being uploaded, which seems as a noveldata-leakingdissemination problem. To solve it, we propose LHDS (Low-latency Heterogeneous Dissemination Scheduling) scheme by transforming the problem into two special sub-problems, i.e., constructing a special degree-decreasing tree with maximum multichild nodes, and designing a leaking-sustained dissemination schedule for each subtree. Extensive simulation experiments have been conducted on LHDS as well as two heuristic algorithms (i.e.,DBOandS-GA) designed for baselines. The results show that LHDS scheme significantly outperforms theDBOandS-GAalgorithms in terms of total collaborative data uploading latency, with saving 41% and 42% latency on average, respectively.

Thick metal barriers pose a fundamental challenge to internal wireless monitoring due to their electromagnetic shielding properties, rendering traditional radio frequency (RF) communications ineffective. Acoustic communication offers a promising alternative that is also inherently secure against remote eavesdropping. However, the solid medium itself, when acting as a communication channel, exhibits complex characteristics of high loss and severe dispersion, which severely constrains reliability. This paper aims to solve this core problem by designing and implementing a novel intra-solid acoustic communication system. The core of our approach lies in proposing and validating a robust physical layer communication paradigm for the reliable propagation of acoustic signals in such an extreme channel environment. This paradigm, by encoding and spreading signal energy across a wide frequency band, is designed to fundamentally combat the severe attenuation and frequency-selective fading introduced by the channel. We build a complete system prototype and validate the effectiveness of this paradigm through a series of experiments in a real-world solid metal environment. The experimental results demonstrate that our system can successfully establish reliable communication links through thick metal barriers, providing a validated and feasible solution for the deployment of wireless sensor networks in extremely shielded environments.

Acoustic-based respiration sensing is promising due to its ubiquitous device support and great freedom in signal design. However, existing proposals often either fail to function properly when a target is non-static or is under multipath interference, or address it in an algorithmic manner. To this end, in this paper, we propose CORA, a COntinuous RespirAtion monitoring system using purely analytical signal processing methods. CORA is the first approach that achieves physical separation between motion artifacts and respiration, other than existing algorithmic solutions, and hence can obtain results that are closer to ground truth. CORA leverages the edges of Orthogonal Time Frequency Space signals in monitoring motion states and addressing multipath interference. The ability to tackle these challenges can help to compensate motion-induced artifacts for FMCW-based sensing techniques, enabling continuous respiration monitoring even in non-static scenarios. To achieve high-quality compensation, a pipeline of signal processing techniques is proposed, including robust moving target tracking, accurate frequency bin selection, and effective phase denoising. Unlike existing deep learning-based approaches, CORA is explainable and is readily deployable, without sophisticated adaptation or exhausted training processes. We have implemented a system prototype and evaluated its performance. Experiment results demonstrate a median error of 0.86 respiration per minute.

With the wide adoption of versatile IoT devices, device providers may desire to locate users around those devices to plan context-aware intelligence, which may improve the quality of daily life. As most IoT devices are WiFi enabled, the WiFi-based indoor positioning system is supposed to achieve this future scene. However, the state-of-the-art WiFi indoor positioning systems face challenges when being practically deployed as they may have to tradeoff between, say accuracy and computational overhead. In light of this, this article mainly introduces PLP-Track, a practical lightweight passive indoor tracking system based on channel state information (CSI) fingerprints. To settle the low granularity of fingerprints in distinguishing different positions, we propose a fingerprint preprocessing algorithm based on unsupervised learning and incorporate this algorithm with a state-space model to enable lightweight real-time tracking. Our implementation and evaluation of commodity WiFi devices demonstrate that PLP-Track can achieve indoor localization with high accuracy and low-computation cost.