- Non-Invasive Vital Sign Monitoring
- ECG Monitoring and Analysis
- Human Pose and Action Recognition
- Hemodynamic Monitoring and Therapy
- Multimodal Machine Learning Applications
- Anomaly Detection Techniques and Applications
- Microwave Imaging and Scattering Analysis
- Electrical and Bioimpedance Tomography
- Antenna Design and Analysis
- Healthcare Technology and Patient Monitoring
- Indoor and Outdoor Localization Technologies
- Domain Adaptation and Few-Shot Learning
- Advanced SAR Imaging Techniques
- Microwave Engineering and Waveguides
- Target Tracking and Data Fusion in Sensor Networks
- Gait Recognition and Analysis
- Advanced Neural Network Applications
- Biomedical Text Mining and Ontologies
- Hand Gesture Recognition Systems
- Antenna Design and Optimization
- Electronic Health Records Systems
- Advanced Wireless Communication Technologies
- Satellite Communication Systems
- Data Quality and Management
- UAV Applications and Optimization
Northwestern Polytechnical University
2019-2025
Zhejiang University
2013-2020
Short-range continuous-wave Doppler radar sensors have been mainly used for noncontact detection of various motions. In this paper, we investigate the feasibility to implement function a remote mouse, an input device computer, by recognizing human gestures based on dual-channel sensor. Direct conversion architecture, symmetric subcarrier modulation, and bandpass sampling techniques are obtain cost-effective solution. An arcsine algorithm motion imaging proposed linearly reconstruct hand...
Noncontact detection of human vital signs based on miniaturized Doppler radar systems (DRSs) can be widely used in healthcare and biomedical applications. Although significant progresses have been achieved, a reliable wireless the presence large-scale random body movements remains technical challenge. In this paper, comprehensive use high-dynamic-range architecture linearized phase demodulation algorithms, we further introduce matched filters to retrieve respiration heartbeat spectra...
Object detection for remote sensing is a fundamental task in image processing of sensing; as one the core components, small or tiny object plays an important role. Despite considerable advancements achieved with integration CNN and transformer networks, there remains untapped potential enhancing extraction utilization information associated objects. Particularly within structures, this arises from disregard complex intertwined interplay between spatial context channel during global modeling...
Miniaturized Doppler radar sensor (DRS) for noncontact motion detection is a hot topic in the microwave community. Previously, small-scale physiological signals such as human respiration and heartbeat rates are primary interest of study. In this paper, we propose comprehensive approach that can be used to improve demodulation linearity DRSs, detailed time-domain information ranging from micro-scale large scale accurately reconstructed. Experiments show based on digital-IF receiver...
Most medical instruments invented to measure human heart activities, such as electrocardiograms (ECGs), rely on contact electrodes. This causes discomfort and limits application scenarios. Remote acquisition of μV-level bioelectrical cardiac signals through ECG measurement is theoretically challenging. Based the analysis magnetic resonance imaging volume change hearts, we found that a single radar sensor can be used remotely detect Doppler cardiogram (DCG) at distance up 1 m, by retrieving...
Doppler radar sensor has been widely used in non-contact bio-signal monitoring. This paper aims at recovering bio-signals from body movement. To solve the severe phase wrapping and saturation problems large-scale movement, as well unwanted DC offsets gradual changes of received microwave power problems, curve fitting technology is employed to compensate for movement recover small-scale based on a digital-IF structured, high-dynamic-range linearized demodulated algorithms. Experimental...
In the past decade, many efforts have been devoted to indoor localization solutions. While significant progresses achieved, short-range of passive moving objects in environments remains a technical challenge, especially when available radio spectrum is limited. this paper, we resolve challenge by proposing single-frequency continuous-wave Doppler radar sensor implemented with redundant single-input multiple-output architecture. Since only phase shift information used proposed...
Capable of remote presentations human heart activities, the Doppler cardiogram (DCG) can be detected by an agile radar sensor. However, as a recently developed technology, its performance needs sustained improvements for future practical applications. In this work, continuous-wave, <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">${V}$ </tex-math></inline-formula> -band sensor was implemented to improve...
It is a challenging task for continuous-wave Doppler radar sensor (DRS) to linearly separate independent motions without any prior information. In this paper, low-IF single-input multiple-output (SIMO) DRS system designed and implemented motion separation. The SIMO detects the signals scattered off multiple objects, two-step blind separation approach proposed from output. Experiments show that can successfully combinations of triangular, sinusoidal, random motions, as long velocities or...
Calculating semantic similarity is paramount in medical information processing, and it aims to assess the of professional terminologies within databases. Natural language models based on Bidirectional Encoder Representations from Transformers(BERT) offer a novel approach representation for calculations. However, due specificity terminologies, these often struggle with accurately representing semantically similar terms, leading inaccuracies term consequently affecting accuracy To address this...
In this paper, we present a Doppler radar sensor based approach capable of detecting both small- and large-scale biological motions. To fully recover the time domain motion information from backscattered microwave signals, gradient descent extended DACM algorithms are utilized to solve problems caused by DC offset phase wrapping issues. Experimental results show that bio-signals small-scale human cardiac walk can be accurately detected. The proposed provides generalized method for different...
In the past decades, continuous Doppler radar sensor-based bio-signal detections have attracted many research interests. A typical example is heartbeat detection. While significant progresses been achieved, reliable, time-domain accurate demodulation of bio-signals in presence unavoidable DC offsets remains a technical challenge. Aiming to overcome this difficulty, we propose paper novel algorithm that does not need trace and eliminate dynamic based on approximating segmented arcs quadrature...
In this work, we proposed and verified that a noncontact measurement to the human cardiogram could be implemented based on inner connection between bioelectrical signals control diastole systole of hearts volume change atriums ventricles taking place at surface skin. A K-band, board-integrated, highly linear sensitive Doppler radar sensor was designed detected such weak motions. Experimental results showed by linearly retrieving information from scattered skin, can measured. While...
Indoor localization is a hot topic in microwave engineering. However, short-range of passive moving objects indoor environments remains technical challenge. In this paper, we propose single-frequency continuous-wave Doppler radar sensor implemented form single-input multiple-output (SIMO) architecture with redundant receiving channel. It only uses the information localization. case, naturally immune to stationary clutters such as walls and furniture, making it responsive objects. The...
Practical applications of the quantitative inversion are restricted by limited availability measurement data, which is denoted as limited-aperture inverse scattering problem (ISPs). To circumstance this challenge, a novel limitedaperture data complementation method utilizing generative adversarial network (GAN) for imaging presented first time. Based on powerful nonlinear fitting capabilities and low computational complexity, pix2pix GAN utilized to complement measured into comprehensive...
Existing works in few-shot action recognition mostly fine-tune a pre-trained image model and design sophisticated temporal alignment modules at feature level. However, simply fully fine-tuning the could cause overfitting due to scarcity of video samples. Additionally, we argue that exploration task-specific information is insufficient when relying solely on well extracted abstract features. In this work, propose simple but effective adaptation method (Task-Adapter) for recognition. By...