A5101543507- Indoor and Outdoor Localization Technologies
- Advanced Image and Video Retrieval Techniques
- Robotics and Sensor-Based Localization
- Advanced Vision and Imaging
- Evaluation Methods in Various Fields
- ECG Monitoring and Analysis
- Advanced Decision-Making Techniques
- Advanced Memory and Neural Computing
- Target Tracking and Data Fusion in Sensor Networks
- Speech and Audio Processing
- EEG and Brain-Computer Interfaces
- Advanced Sensor and Control Systems
- Evaluation and Optimization Models
- Human Mobility and Location-Based Analysis
- Internet Traffic Analysis and Secure E-voting
- Video Surveillance and Tracking Methods
- Smart Grid Security and Resilience
- Underwater Vehicles and Communication Systems
- Anomaly Detection Techniques and Applications
- Image Processing Techniques and Applications
- Parallel Computing and Optimization Techniques
- Network Security and Intrusion Detection
- Advanced Neural Network Applications
- Inertial Sensor and Navigation
- Brain Tumor Detection and Classification
Jiading District Central Hospital
2022-2025
Shanghai University of Medicine and Health Sciences
2021-2025
Lanzhou Jiaotong University
2025
University of Electronic Science and Technology of China
2014-2024
Shanghai Jiao Tong University
2008-2024
PLA Army Engineering University
2021
First Affiliated Hospital of Hunan University of Traditional Chinese Medicine
2020-2021
China Electric Power Research Institute
2016-2020
North China Electric Power University
2007-2020
Sun Yat-sen University
2010-2015
Smart grid has emerged as the next generation of power grid, due to its reliability, flexibility, and efficiency. However, smart faces some critical security challenges such message injection attack replay attack. If these cannot be properly addressed, an adversary can maliciously launch injected or replayed attacks degrade performance grid. To cope with challenging issues, in this paper, we propose efficient authentication scheme that employs Merkle hash tree technique secure gird...
The ubiquity of smartphones and their rich set on-board sensors has created many exciting new opportunities, where are used as powerful computing platforms to sense analyze pervasive data. One important application mobile sensing is activity recognition based on smartphone inertial sensors, which a fundamental building block for variety scenarios, such indoor pedestrian tracking, health care, smart cities. Although approaches have been proposed address the human problem, several challenges...
Phishing detection in Semantic Web systems is crucial to safeguarding users from malicious attacks. In this context, work presents a deep learning-based phishing attack model using MobileBERT for feature extraction and hyperparameter optimization covariance matrix adaptation evolution strategy (CMA-ES). The obtained 95% classification accuracy. Important benchmarks like accuracy, recall, F1-score show good ability discriminate between legitimate emails. Applying CMA-ES, which improved helps...
ECG classification is a key technology in intelligent electrocardiogram (ECG) monitoring. In the past, traditional machine learning methods such as support vector (SVM) and K-nearest neighbor (KNN) have been used for classification, but with limited accuracy. Recently, end-to-end neural network has shows high However, large computational complexity including number of parameters operations. Although dedicated hardware field-programmable gate array (FPGA) application-specific integrated...
Smart grid1 is a modern power transmission network. With its development, the computing, communication and physical processes getting more connected. However, an adversary can destroy production by attacking secondary equipment. Accurate fast response to cyber-attacks prerequisite for stable grid operation. Therefore, it critical identify classify attacks in smart grid. In this paper, we propose novel approach that utilizes machine learning algorithms help cyber-attacks. We built deep neural...
The rapid development of smart cities and the integration IoT devices have significantly increased security vulnerabilities, especially within consumer electronics, exposing them to complex cyber-attacks. This paper aims develop a robust model enhance threat detection protection these in city environments. We propose novel approach utilizing Harris Hawks Optimizer (HHO) for feature selection Mountain Gazelle (MGO) hyperparameter tuning an XGBoost-based framework network traffic analysis....
Bus networks are a crucial support for urban commuting. By studying the evolutionary characteristics of bus networks, we can uncover their development patterns, coverage efficiency, and changes in regional balance, providing scientific basis sustainable optimization transportation resources. This study systematically analyzes spatiotemporal evolution network Beijing from 2006 to 2024 using specific spatial analysis tools analyze characteristics. analyzing rates transit stations road...
Phishing detection is a critical challenge in virtual realities, where malicious activities can compromise user security. This paper presents novel approach integrating AI and Semantic Web Technologies for robust phishing detection. The proposed model preprocesses text data leverages reduced six-layer BERT encoder to extract contextual embeddings. Outputs from BERT, including classifier, attention, layers, are combined with features derived custom deep learning layer form unified...
A new type of half-mode substrate integrated waveguide (SIW) periodically loaded with different lumped elements and structures is proposed in this paper. The propagation constants, Bloch impedances, voltage distributions the Floquet TE-modes are all characterized by solving eigenvalues generalized transmission matrix a single period. Several typical capacitor-loaded, inductor-loaded, corrugated, electromagnetic-bandgap (EBG)-loaded SIWs were fabricated to demonstrate their modal dispersion...
Seizure-detection processors using machine learning have been proposed to detect the seizure onset of patients for alert or stimulation purposes [1–4]. Existing designs can achieve high accuracy when large amounts data from a patient is available training. However, unlike collection non-seizure data, with low occurrence requires undergo time-consuming and costly hospitalization, which difficult in practice. To address this issue, [5] zero-shot-retraining seizure-detection processor achieving...
Zero Velocity Update (ZUPT) has played a key role in Pedestrian Dead Reckoning (PDR) with inertial measurement units (IMU). However, it is both crucial and difficult to determine ZUPT conditions given complex varying motion types such as walking, fast walking or running, different habits of distinct people, which have direct significant impact on the tracking accuracy. In this research we proposed model based deep neural networks moments when should be conducted. The ensures nearly identical...
map matching has played a crucial role in technologies related to indoor positioning. Conventional algorithms based on particle filter (PF) have some limitations, such as the limited use of information, poor generalization and low precision. To solve these problems, we propose an adaptable network (AdaPFnet), novel technique that integrates algorithm into neural network. AdaPFnet uses local views particles represent so information about location can be learned sufficiently through...
Several closed-form equations are obtained for characterizing the average power handling capability (APHC) of bandpass filters (BPFs) using planar half-wavelength open-circuited microstrip resonators. It is found that temperature rise (TR) BPF product two terms corresponding to single resonator and filter, respectively. The APHC an eighth-order open-loop determined numerically, with good agreements between its maximum TRs calculated our model 3-D-FEM software.
On-chip thermal sensors are employed by dynamic management (DTM) techniques to appropriately manage chip performance. However, the effectiveness of DTM mechanisms is directly dependent on number placed sensors, which should be minimised, while guaranteeing accurate tracking hot spots and full characterisation. In this study, authors propose a rigid sensor allocation placement technique for determining fewest optimal locations based dual clustering. Initially, utilise clustering algorithm...
The ubiquity of smartphones and their rich set onboard sensors have created many exciting new opportunities. One important application is activity recognition based on smartphone inertial sensors, which a fundamental building block for variety scenarios, such as indoor pedestrian tracking, mobile health care smart cities. Though approaches been proposed to address the human problem, number challenges still present: (i) people's motion modes are very different; (ii) there limited amount...
In this paper, an extension of spatial channel model (SCM) for vehicle-to-vehicle (V2V) communication in roadside scattering environment is investigated the first time theoretically and by simulations. Subsequently, to efficiently describe reflect nonstationary properties V2V channels, proposed SCM divides objects into three categories clusters according location effective scatterers introducing critical distance. We derive general expressions most important statistical such as impulse...
As one of the most significant technology in wireless sensor networks (WSN), localization has drawn much attention. In this paper, received signal strength (RSS) values are used as indicator distance between blind node and reference nodes. The position is calculated via multilateration algorithm (MA). order to improve accuracy, Kalman filter (KF) utilized estimate actual position. Due flaw model, divergence phenomenon occurs when moving direction changes. Therefore, performs badly location...
The location method based on TOA estimation is a research hotspot in the field of wireless location. In this paper, an new algorithm proposed for providing better performance. cross correlation matrix employed to process channel frequency domain, which benefit noise suppression. Channel optimal threshold algorithm, too. Compared with existing algorithms, adopted scenario low SNR.
In order to solve the problem of poor fusion between spots deformation camouflage and background, a small-spot design algorithm based on background texture matching is proposed in this research. The combination textures improved spot pattern background. An adversarial autoencoder convolutional network was designed extract features. image loss added reconstruction improve clarity generated generalization ability model. digital formed by obtaining mean value square area replacing main color....
The environmental sound classification (ESC) has attracted increasing attention as the contains a wealth of information that can be used to detect particular events. However, so far, most existing work in ESC still remains stage algorithm design and processor not been thoroughly investigated. designs have issues meeting low power consumption high accuracy simultaneously due lack joint-optimization between hardware, very few demonstrated complete system containing all necessary modules. In...
This paper describes the effects of RF power level on performance varactor-tuned resonator circuits. A variety topologies are considered, including series and parallel resonators operating in both unbalanced balanced modes. As these were designed to produce oscillators with minimum phase noise, initial small signal insertion loss was set 6 dB and, hence, QL/Q0 = 1/2. To enable accurate analysis simulation, S parameter PSPICE models for varactors optimized developed. It is shown that start...