Noninvasive blood glucose monitoring has long been a critical research focus in diabetes management. Among emerging technologies, photoacoustic sensing, combining the molecular specificity with deep penetration, has garnered significant attention. It offers rapid response and pain-free operation, making it a strong candidate for next-generation portable blood glucose monitoring devices. This review systematically traces the development and current state of photoacoustic glucose sensing, with a particular focus on the selection and optimization of core system components. It also summarizes common interference in glucose detection and outlines strategies for their mitigation, along with signal processing and signal-to-noise ratio enhancement techniques suitable for real-world applications. Addressing the growing demand for wearable continuous glucose monitors, this work analyzes the key challenges in system integration and outlines recent advances in enabling technologies. It proposes multi-technology integration approaches to bridge the gap between photoacoustic sensing and microsystem design, offering theoretical foundations and practical guidance for future research on wearable photoacoustic systems.

Recent advances in Graphical User Interface (GUI) and embodied navigation have driven progress, yet these domains have largely evolved in isolation, with disparate datasets and training paradigms. In this paper, we observe that both tasks can be formulated as Markov Decision Processes (MDP), suggesting a foundational principle for their unification. Hence, we present NaviMaster, the first unified agent capable of unifying GUI navigation and embodied navigation within a single framework. Specifically, NaviMaster (i) proposes a visual-target trajectory collection pipeline that generates trajectories for both GUI and embodied tasks using a single formulation. (ii) employs a unified reinforcement learning framework on the mix data to improve generalization. (iii) designs a novel distance-aware reward to ensure efficient learning from the trajectories. Through extensive experiments on out-of-domain benchmarks, NaviMaster is shown to outperform state-of-the-art agents in GUI navigation, spatial affordance prediction, and embodied navigation. Ablation studies further demonstrate the efficacy of our unified training strategy, data mixing strategy, and reward design. Resources will be released to the community.

Photoplethysmography (PPG) is a non-invasive optical sensing modality that captures peripheral blood-volume dynamics related to cardiac activity and vascular function, making it useful for frequent cardiovascular monitoring with wearable devices. However, existing public PPG datasets remain limited in their ability to represent bilateral and posture-dependent physiological variability in wearable settings. We present MAIBO, a multi-posture, asymmetry-aware bilateral PPG dataset for continuous cardiovascular monitoring. MAIBO comprises synchronized PPG signals acquired from both hands using wearable rings under three standardized postural conditions. The dataset includes 1810 participants and 7478 measurement sessions, with reference annotations of systolic blood pressure (SBP), diastolic blood pressure (DBP), heart rate (HR), and demographic attributes including age, gender, and body mass index (BMI). All data were collected under controlled multi-postural protocols and subjected to quality-control procedures, including the removal of incomplete samples and missing signal segments. By offering a well-curated and reproducible resource, MAIBO can support method development for wearable cardiovascular monitoring and studies of bilateral and posture-dependent digital phenotypes.

Wearable photoplethysmography (PPG) enables non-invasive cardiovascular monitoring, but most systems rely on single-site measurements and rarely test whether signals are spatially homogeneous across limbs. Here, we built a large-scale bilateral ring-PPG benchmark comprising 97,559 valid samples from 1810 participants with synchronized measurements from both hands. Bilateral averaged predictions showed target-dependent performance, with mean absolute errors of 3.21 bpm for heart rate (HR), 12.36 mmHg for systolic blood pressure (SBP), 8.21 mmHg for diastolic blood pressure (DBP), and 9.46 years for age. Relative to a cohort-mean baseline, HR showed the clearest improvement (10.65-3.21 bpm), whereas age improved more modestly (11.08-9.46 years). BP estimation performance remained limited, with MAE reductions from 13.81 to 12.36 mmHg for SBP and from 9.27 to 8.21 mmHg for DBP, while error dispersions remained close to target variability. Similar trends across 12 backbones suggested that the pattern was not architecture-specific. Samples with smaller bilateral differences further reduced HR MAE by 19.3%. Together, these findings establish a bilateral wearable PPG benchmark with strongest utility for HR, cautious utility for age, and limited, transparently quantified BP performance.

Continuous noninvasive blood pressure monitoring is important for neonatal hemodynamic management. However, cuffless photoplethysmography (PPG) is still difficult to use clinically because acquisition-related errors and their effects on blood pressure estimation are not well understood. Most studies report only overall error, which can hide error patterns under different signal-quality and motion conditions. Here, we developed a soft multi-wavelength wearable that combines reflective and transmissive optical paths to record complementary PPG signals. We also built NEO-BP, a prospective clinical dataset containing synchronized multi-wavelength PPG and invasive arterial blood pressure waveforms from 42 neonates. In a held-out segment-level test set without excluding samples based on signal quality (n = 9378), the green-red-infrared (G+R+IR) model achieved mean absolute errors of 9.67 mmHg for systolic blood pressure and 6.17 mmHg for diastolic blood pressure. Using this dataset, we established a retrospective descriptive subgroup error profile by stratifying estimation deviations across measurable acquisition and physiological conditions. Estimation errors increased under lower signal quality and stronger motion. These results identify error-prone acquisition states, support quality-aware interpretation of segment-level cuffless blood pressure estimates, and provide a basis for future models that estimate the error risk of individual readings.

Traditional cuff-based blood pressure (BP) monitoring methods suffer from inherent limitations in user discomfort and intermittent measurement. In contrast, photoplethysmography (PPG)-based BP estimation provides a promising alternative. However, several critical challenges remain in PPG-based BP estimation. First, prior evidence indicates that finger-worn devices can yield signals with higher feasibility and quality than wrist-worn devices. Second, long-term variations in blood pressure have not been sufficiently investigated in this field. To address these challenges, this study develops a novel smart ring platform and collects samples to train and evaluate the proposed model. A large-scale dataset comprising 116,995 samples from 96 subjects is constructed to evaluate long-term performance. The trained model achieves mean absolute errors (MAE) of 11.20 mmHg for systolic blood pressure (SBP) and 7.51 mmHg for diastolic blood pressure (DBP). These results demonstrate that a smart ring provides a comfortable and practical solution for continuous blood pressure management in personalized healthcare.

Early detection and treatment of cardiovascular diseases (CVDs) can be significantly enhanced through the use of flexible wearable electrocardiogram (ECG) sensors, potentially reducing CVD-related mortality. This paper introduces an ECG-on-Chip (EoC) solution tailored for flexible ECG sensors, incorporating novel features to address common challenges in wearable ECG technology. The EoC integrates a chopper-stabilized capacitively-coupled instrumentation amplifier (CS-CCIA), ensuring a high common-mode rejection ratio (CMRR) and low noise performance. Performance is further boosted via a positive feedback loop (PFL) and a programmable gain amplifier (PGA) with shared on-chip calibration logic, which enhances input impedance and minimizes gain variability to ensure consistent algorithm performance. Additionally, a secondary chopping technique is employed to further reduce noise, achieving an input-referred noise level of 454 nVrms. The embedded algorithm within the EoC is designed to extract clinically meaningful features, facilitating robust real-time arrhythmia analysis. Fabricated using a 0.18 µm CMOS process, the EoC consumes 14.9 µW with a supply voltage of 1.2 V. The algorithm’s efficacy has been validated over 0.4 million heartbeats, demonstrating a sensitivity of over 99.7

Face clustering, a critical task for annotating large-scale unlabeled face recognition datasets, aims to group facial images of the same identity while minimizing annotation costs. Recent approaches model face relationships as graphs and leverage graph neural networks (GNNs) to capture structural dependencies. However, the performance of these methods heavily relies on the quality of the input graph, where conventional kNN-based graph construction suffers from two key limitations: 1) sensitivity to hyperparameter k, leading to either excessive false-positive edges (large k) or fragmented true-positive connections (small k), and 2) propagation of noise through message passing in GNNs. To address these challenges, we propose AtC (Align then Clip), a novel framework that refines graph structures through dual-phase optimization. During training, we introduce a distribution alignment branch that aligns node representations between noisy and clean graphs, enhancing robustness to edge noise. For inference, we design Post-Clipping, an adaptive edge pruning strategy that eliminates spurious connections while preserving critical linkages. Furthermore, our Post-Linkage mechanism mitigates cluster fragmentation by reconnecting split components. Extensive experiments validate its scalability (e.g., 5.21M samples), robustness to 30% false-positive noise, and superior cross-domain generalization, demonstrating broad applicability in real-world scenarios. Extensive experiments demonstrate the method’s scalability (on datasets of up to 5.21 million samples), robustness (even with 30% false-positive noise), and superior cross-domain generalization. These results showcase its strong potential for real-world applications.

Cardiovascular disease (CVD) remains the leading cause of global mortality, highlighting the need for continuous vital sign monitoring. Photoplethysmography (PPG) is well suited for wearable devices. Smart rings, benefiting from dense capillary distribution and minimal tissue interference, can capture high-quality PPG signals with comfort, making them a promising next-generation wearable. However, ring rotation relative to the finger alters the optical path, especially for multi-wavelength light, thus reducing accuracy. This paper proposes a rotation-robust PPG sensor for smart rings. Monte Carlo simulations analyze photon transmission under different LED–photodiode (PD) angles, showing that at ±60°, green, red, and infrared light achieve optimal penetration into the microcirculation layer. Considering non-ideal conditions, the green-light angle is adjusted to ±30°, and a symmetrical sensor design is adopted. A prototype smart ring is developed, capable of recording 4-channel PPG, 3-axis acceleration, and 4-channel temperature signals at 100, 25, and 0.2 Hz, respectively. The system achieves reliable PPG acquisition with only 0.59 mA average current consumption. In continuous testing, heart rate estimation reached mean absolute errors of 0.82, 0.79, and 0.78 bpm for green, red, and IR light. The results provide a reference for future smart ring development.

Motion synthesis in real-world 3D scenes has recently attracted much attention. However, the static environment assumption made by most current methods usually cannot be satisfied especially for real-time motion synthesis in scanned point cloud scenes, if multiple dynamic objects exist, e.g., moving persons or vehicles. To handle this problem, we propose the first Dynamic Environment MOtion Synthesis framework (DEMOS) to predict future motion instantly according to the current scene, and use it to dynamically update the latent motion for final motion synthesis. Concretely, we propose a Spherical-BEV perception method to extract local scene features that are specifically designed for instant scene-aware motion prediction. Then, we design a time-variant motion blending to fuse the new predicted motions into the latent motion, and the final motion is derived from the updated latent motions, benefitting both from motion-prior and iterative methods. We unify the data format of two prevailing datasets, PROX and GTA-IM, and take them for motion synthesis evaluation in 3D scenes. We also assess the effectiveness of the proposed method in dynamic environments from GTA-IM and Semantic3D to check the responsiveness. The results show our method outperforms previous works significantly and has great performance in handling dynamic environments.

Wearables provide a promising solution for proactive health by monitoring real-time vital signs long-term. However, wrist-worn devices, like smartwatchcs, have faced issues with inaccuracy and information loss. This is primarily due to two factors: motion artifacts, which interfere with recorded signals due to relative movement, and the anatomical complexity of the wrist, complicating sensor placement, especially for optical sensors that require optimal positioning of light sources and receivers. Photoplethysmography (PPG) is a critical signal recorded by wearables. The Finger is a more suitable measurement site for PPG signals due to its high vascular density and few static tissues. Additionally, fingerworn des ices adhere more securely to the skin than wrist-worn devices, reducing sensitivity to external interference and improving PPG signal quality. Advances in electronic devices have led to significant progress in smart rings, making them a promising next-generation wearable. This paper discusses the development, potential, and challenges of smart ring technologies. With accelerating research, we believe smart rings are poised to become a main player in the wearable market.

We present SDTracker, a method that harnesses the potential of synthetic data for multi-object tracking of real-world scenes in a domain generalization and semi-supervised fashion. First, we use the ImageNet dataset as an auxiliary to randomize the style of synthetic data. With out-of-domain data, we further enforce pyramid consistency loss across different "stylized" images from the same sample to learn domain invariant features. Second, we adopt the pseudo-labeling method to effectively utilize the unlabeled MOT17 training data. To obtain high-quality pseudo-labels, we apply proximal policy optimization (PPO2) algorithm to search confidence thresholds for each sequence. When using the unlabeled MOT17 training set, combined with the pure-motion tracking strategy upgraded via developed post-processing, we finally reach 61.4 HOTA.

The utilization of sound waves for the precise arrangement of particles and cells represents a significant advancement in the field of micromanipulation. Acoustic tweezers, based on surface acoustic waves, offer a non-invasive and non-contact approach, showcasing immense potential in biology and medicine. In comparison to conventional micromanipulation tools like optical and magnetic tweezers, acoustic tweezers exhibit distinct advantages in simultaneously manipulating a large number of particles. In this study, we employ piezoelectric ceramics to convert electrical signals from a generator into sound waves, enabling the manipulation of particles. Through experimental investigations, we explore the influence of input voltage, and input frequency on particle cluster spacing and aggregation levels. The outcomes of this research provide a simple and convenient operational platform for further cell micromanipulation and subsequent high-throughput screening of particles or cells.

A multihole probe is a very effective sensor for measuring the velocity and direction of a flow field. Compared with other sensors, it has a highly reliable and wide range of application prospects. A comparative analysis of probes of different shapes is carried out, where it is found that a hemispherical seven-hole probe with a vertical surface structure opening is the upgrade structure. Based on the spherical structure, a theoretical calibration method for the probe is proposed, and the correlation between the pressure in the hole and the velocity of the flow field is established by a pressure-velocity parameter equation. By calibrating the hemispherical seven-hole probe, it is found that the speed error is less than 5%, and the angle error is less than 2°. Compared with other calibration methods, the proposed calibration process is greatly shortened. At the same time, the hemispherical seven-hole probe is used to measure the flow field around a cylinder, where frequency changes of 3.497, 2.083, and 1.657 Hz are observed in 3, 5, and 9 cm cylindrical vortexes, respectively. It is proved that the multihole probe also has the ability to detect complex flow fields.

This paper presents a study on a configurable multimodal therapeutic system designed to analyze the impact of various therapeutic environments on participants' health and determine more effective approaches. The system aims to relax individuals and gradually enhance their well-being by manipulating the surrounding lighting conditions. The experiment involved 22 adults participating in four rounds, each exploring the effects of colored light, flickering frequencies, feedback regulation, and multimodal stimuli. EEG parameters were measured, and pre- and post-assessment stress surveys were conducted to understand the effects of different therapeutic environments on participants' stress levels. Statistical analysis using one-way analysis of variance (ANOVA) and t-tests revealed significant results. Blue light, non-flickering light, stimuli with feedback regulation, and multimodal stimuli were found to be more effective in promoting emotional relaxation. The findings emphasize the importance of considering different therapeutic environments and highlight the potential benefits of a configurable multimodal therapeutic system for identifying effective approaches.

Assembly accuracy of aeroengines influences operation performance and service life. The coaxiality of the aeroengine is the main index of assembly accuracy and is also a core index to represent assembly quality. However, direct measurement of coaxiality is a difficult technical problem due to the sealed structure of the aeroengine casing system. A coaxiality prediction method is proposed to obtain coaxiality and assist assembly by geometric distribution error modeling and point cloud deep learning. The prediction process consists of three steps. In the beginning, the geometric distribution error model is established to construct the accurate dense point cloud of aeroengine part surfaces by the non-uniform rational B-splines (NURBS) method based on the coordinate measuring machine collecting information. Then, the mapping between the dense point cloud and coaxiality is established to obtain an assembly dataset by the virtual assembly. Finally, the dataset is fed to a new point cloud deep learning backbone, Self-channel cross attention point network, and realizes end-to-end coaxiality prediction based on the aeroengine surface point cloud. The geometric distribution error model is tested on the aeroengine simulated parts with 0.001 mm accuracy. The prediction method is verified on the aeroengine simulated parts and compared with other point cloud deep learning baselines. The method proposed in this paper realizes 93.17% prediction accuracy with 0.01 mm coaxiality precision which is a high performance and meets the requirements of industrial measurement. This paper provides an effective coaxiality prediction model for the aeroengine casing system, to improve the accuracy and efficiency of the aeroengine assembly.

In order to solve the problems of difficult convergence and local optimal solution of ant colony optimization (ACO) algorithm, and low convergence accuracy of particle swarm optimization (PSO) algorithm, a particle swarm optimization ant colony optimization (PSO-ACO) fusion algorithm is proposed to deal with the three-dimensional (3D) path planning problem of unmanned underwater vehicle (UUV). In this algorithm: based on the idea of spatial stratification, a 3D grid model is established to build underwater environment model; PSO algorithm is used to pre search the path and quickly obtain the solution, which is used as the initial pheromone increment of ACO algorithm; the pheromone global updating method of ACO algorithm is improved: an adjusting factor is added to the pheromone global update equation to accelerate the convergence speed of ACO algorithm; the state transition equation of ACO algorithm is also improved, so that the algorithm has a greater probability to select the point with the largest weighted product of pheromone and heuristic information as the next path point. Experimental results show that the fusion algorithm effectively improves the global search ability and shortens the search time.

Recently, studies considering domain gaps in shape completion attracted more attention, due to the undesirable performance of supervised methods on real scans. They only noticed the gap in input scans, but ignored the gap in output prediction, which is specific for completion. In this paper, we disentangle partial scans into three (domain, shape, and occlusion) factors to handle the output gap in cross-domain completion. For factor learning, we design view-point prediction and domain classification tasks in a self-supervised manner and bring a factor permutation consistency regularization to ensure factor independence. Thus, scans can be completed by simply manipulating occlusion factors while preserving domain and shape information. To further adapt to instances in the target domain, we introduce an optimization stage to maximize the consistency between completed shapes and input scans. Extensive experiments on real scans and synthetic datasets show that ours outperforms previous methods by a large margin and is encouraging for the following works. Code is available at https://github.com/azuki-miho/OptDE .

In order to meet the personalized charging needs of different users and avoid the negative impact of large number of electric vehicles charging in a disordered manner on the power grid, this paper proposes an ordered charging strategy for electric vehicles. In this strategy, users set subjective priority according to their different travel convenience and economy, charging station operators set objective priority according to users’ required power and latest departure time, and calculate comprehensive priority by combining subjective priority and objective priority, so as to set charging power for users at different times; according to the different subjective priority of each user, each user’s satisfaction is given different weights. According to the different subjective priority of each user, different weights are assigned to each user’s satisfaction, and the optimization model is established with the goal of maximizing the user’s comprehensive satisfaction, and the inertial weight particle swarm algorithm is used to solve the optimization model. The results show that the strategy proposed in this paper can meet the personalized charging demand of users, improve the satisfaction of users, and play the role of peak-shaving and valley-filling.