The unobtrusive collection of ballistocardiogram (BCG) data makes it a promising option for continuous, low-effort cardiovascular monitoring. Key heart rhythm metrics such as heart rate (HR), inter-beat interval (IBI), and heart rate variability (HRV) are widely recognized indicators of cardiovascular health. The primary clinical goal of BCG applications is to accurately estimate these metrics. Most previous studies have primarily examined healthy young adults, which contrasts significantly with high-risk populations, such as older adults or cardiac patients, who would benefit the most from long-term monitoring in everyday settings. To bridge this gap, we collected nighttime BCG recordings from 46 middle-aged and elderly participants during natural sleep. The subjects included individuals with atrial arrhythmias, ventricular arrhythmias, myocardial ischemia, and nonclinical sinus arrhythmias. This dataset also contains synchronized Holter monitor recordings, which may enhance the use of BCG in assessing cardiac rhythms and detecting early cardiovascular diseases, while also testing the generalizability of BCG algorithms in real-world scenarios involving various health conditions.

Fingerprinting is a practical technology for improving Wi-Fi-based positioning in complex indoor environments. However, the laborious and costly nature of site surveys hinders the creation of accurate fingerprints. In this paper, we propose Virtual Feature maps and Contrastive Learning-enhanced Indoor Positioning (VF-CLIP), a novel method for indoor positioning based on received signal strength (RSS), aiming at reducing the repetitive site survey overhead caused by the dynamic provisioning of access points (AP). VF-CLIP uses a deep neural network fine-tuning technique to reconstruct the fingerprints by incorporating newly detected APs. The proposed method converts raw received signal strength indicator (RSSI) queries into multiple virtual feature maps, which capture the differential similarities between the query vector and the fingerprints from virtual observational reference points. A depth-wise Transformer is then employed to learn the directional spatial relations of these virtual features. Subsequently, contrastive learning is applied to compress the features into a latent space, where the feature distribution at a reference point becomes compacted. We evaluated VF-CLIP on four public datasets and a fingerprint dataset collected within Huazhong University of Science and Technology, comparing its performance with other state-of-the-art methods. The experimental results demonstrated the effectiveness of VF-CLIP in terms of positioning accuracy and its adaptability to varying AP configurations, suggesting its potential applicability in real-world environments.

The emerging location-based applications depend on the fast and accurate positioning of mobile targets.Wi-Fi received signal strength (RSS) fingerprinting provides a promising solution to localize an object in indoor environments. Among the factors challenging the RSS fingerprinting based algorithms are the site survey cost and the time-varying environment, given the unreliable signal qualities. Here we present a novel approach to indoor object positioning using the manifold assumption on the radio map in RSS-location space. Thinking the measured RSS from one access point (AP) in different locations are randomly drawn from a nonlinear manifold (ground truth), we propose an expectation-maximization (EM) style algorithm to reconstruct the sparse representation of this manifold from the noisy RSS observations. Motivated by the observation that the radio map has a strong local correlation in the RSS-location space, we introduce a multi-scale constrained quadratic programming to approximate the manifold.Within limited iterations, we can estimate the ground truth RSS values and parameters simultaneously. As a result, the learned manifold is exploited to predict the object’s position: we develop a positioning algorithm by minimizing the manifold distortion effort which integrates both measurement error and manifold shape preservation. We conducted extensive simulations and experiments in different settings, testing the datasets collected in a building in the last 8 months. The results showed that the proposed approach was adaptive to the varying environmental noise levels, presenting robust positioning performance.

Wi-Fi received signal strength (RSS) fingerprinting is a cost-effective indoor positioning method that has attracted broad attention because it only requires the existing Wi-Fi infrastructure and ubiquitous smart devices. However, conducting a site survey to establish a fingerprint database is time-consuming and labor-intensive, hindering the fast deployment of fingerprint-based indoor positioning systems. This paper presents a novel fingerprint reconstruction framework to reduce the site survey cost. The proposed MoDeFA is a semi-supervised deep learning framework that uses only a fraction of RSS data tagged with known locations to reconstruct the radio map. The resulting augmented fingerprints ensure the implementation of online fingerprint matching. First, we design a deep regression model to predict the RSS distribution from the perspectives of the multiple observational spots at surveyed locations. Next, we apply denoising to the initial radio map, a noisy representation of the ground truth, using the synthetic noise independent of the map errors without the need for ground-truth labels. Finally, we develop a noise-reduction model to compress the feature representation using unlabeled samples. Building on this, a model that combines multi-scale self-attention and gated scale selection can enhance the prediction of a target’s position. We tested MoDeFA on four public and a private dataset collected from the School of Software Engineering Building at Huazhong University of Science and Technology. The results indicate that the semi-supervised MoDeFA outperforms other state-of-the-art methods in various settings, even with only a fraction (30%) of the labeled RSS samples.