A5103064152- Blind Source Separation Techniques
- Advanced Adaptive Filtering Techniques
- Speech and Audio Processing
- Advanced Wireless Communication Techniques
- Wireless Communication Networks Research
- Direction-of-Arrival Estimation Techniques
- Target Tracking and Data Fusion in Sensor Networks
- Distributed Sensor Networks and Detection Algorithms
- Control Systems and Identification
- Indoor and Outdoor Localization Technologies
- Fault Detection and Control Systems
- Microwave Imaging and Scattering Analysis
- Optical Network Technologies
- Image and Signal Denoising Methods
- PAPR reduction in OFDM
- Sparse and Compressive Sensing Techniques
- GNSS positioning and interference
- Advanced MIMO Systems Optimization
- Advanced Data Compression Techniques
- Energy Efficient Wireless Sensor Networks
- Advanced Statistical Process Monitoring
- Neural Networks and Applications
- Statistical and numerical algorithms
- Radiation Detection and Scintillator Technologies
- Neural Networks and Reservoir Computing
University of Connecticut
2025
Hainan Meteorological Service
2025
Guangxi University
2017-2023
University of Cincinnati
2006-2019
Guangxi University of Chinese Medicine
2017
University of Science and Technology of China
2014
GIRD Systems (United States)
2002-2005
Mississippi State University
2003
Uppsala University
2002
When modeling a system from discrete-time data, continuous-time parameterization is desirable in some situations, In direct estimation approach, the derivatives are approximated by appropriate differences. For an ARX model this lead to linear regression. The well-known least squares method would then be very since it can have good numerical properties and low computational burden, particular for fast or nonuniform sampling. It examined under what conditions fit regression will give adequate...
It is shown how linear time-varying systems can be modeled in several different ways by discrete-time wavelets or, more generally, some set of functions. Interpretation physical meanings, possible efficiency, and other characteristics the modeling are considered. System identification minimizing mean square output error studied. Optimal coefficients corresponding minimum found, they are, general, time varying. Least-mean-square adaptive filtering algorithms derived for on-line system...
Evaluations of andexanet alfa for the reversal factor Xa inhibitor-associated bleeding have been small, with cohorts drawn from single/limited sites. Delays in providing anticoagulation due to hospital transfer may result poorer outcomes. To describe characteristics and outcomes users evaluate association between delay administration a different acute care incidence all-cause inpatient mortality. This was retrospective study using National Inpatient Sample data. Hospitalizations procedural...
China is facing a severe groundwater quality crisis amid economic development and climate change, yet the extent trajectory of this remain largely unknown. Here we developed machine-learning model, incorporating natural social-economic factors, to construct annual probabilistic maps poor (PGQ, i.e., Class V based on Chinese standard) across from 1980 2100. Alarmingly, our findings indicate concerning escalation in PGQ area ratio, rising 17.3% 30.1% 2000, surging 40.8% by 2020, adversely...
This study explores how Taiwanese SMEs strategically transform resource scarcity into digital innovation advantages, particularly through commerce capabilities enhanced by governmental subsidies. Drawing on Resource-Based View, Open Innovation, and Dynamic Capabilities theories, a longitudinal quasi-experimental design was employed, assessing SME performance pre- post-intervention. Results from repeated measures ANOVA regression analyses indicate significant improvement in technology...
Although previous research examine the impact of insufficient resources on innovation, few have investigated how it alters outcomes. Government assistance is crucial for this role, particularly with new technological advancements. Taiwan's economy predominantly consists small and medium enterprises, which many challenges such as limited The study examines governmental facilitates acquisition resources. A to medium-sized enterprise in Taiwan depleting its develop innovative digital concepts....
Charged hadron identification in the Compressed Baryonic Matter experiment (CBM) is realized via Time-of-Flight method [1]. For this purpose CBM-ToF collaboration designed a wall composed of Multi-gap Resistive Plate Chambers (MRPCs). Due to high interaction rate CBM 10 MHz key challenge development MRPCs above 25 kHz/cm2 which become possible after low resistive glass with extremely good quality. In article we present actual conceptual design ToF-wall subdivided three parts namely outer...
The problem of estimating continuous-time autoregressive process parameters from discrete-time data is considered. basic approach used here based on replacing the derivatives in model by differences, forming a linear regression, and using least squares method. Such procedure simple to apply, computationally flexible efficient, may have good numerical properties. It known, however, that all standard approximations highest order derivative, such as repeated use delta operator, gives biased...
Although the Newton algorithm has been extended to complex domain in different forms, none of them seems be directly applicable blind equalization. Therefore, objective this correspondence is develop an for equalization domain. We propose a Newton-like based on Taylor series. Stochastic algorithms (SNLA) two cost functions are developed. Simulations show that new perform slightly better than self-orthogonalizing algorithm.
We consider the blind multiuser detection problem for asynchronous DS-CDMA systems operating in a multipath environment. Using only spreading code of desired user, we first estimate column vector subspace channel matrix by multiple linear prediction. Then, zero-forcing detectors and MMSE with arbitrary delay can be obtained without explicit estimation. This avoids any estimation error, resulting methods are therefore more robust accurate. Corresponding batch algorithms adaptive developed....
An ordinary differential equation (ODE) approach is used to study the focal convergence speed of adaptive IIR filtering and system identification algorithms various structures: direct form, transform domain, lattice. The eigenvalue spreads associated information matrices for are calculated compared. Their limits as unknown poles unit circle obtained. For each three basic structures, two types have been studied: simple constant gain (SCG) type Newton type. It found that Newton-type form...
The feedback lattice filter forms, including the two-multiplier form and normalized form, are examined with respect to their relationships direct filter. Specifically, transformation matrix between forms is derived; parameter state therefore obtained. An IIR structure-the cascade structure-is constructed. Based on this structure, three adaptive filtering algorithms in can then be developed following gradient approach, Steiglitz-McBride approach hyperstability approach. Convergence of these...
Most methods for sensor self-localization require either synchronization or round trip communications among the networked sensors. A differential time-difference-of-arrival (dTDOA) method is proposed in this paper that does not such time nor communications, thus greatly reduces demand on resources and makes implementation much easier. Theoretical results are presented principle of new simulations also given confirm effectiveness method.
Non-GPS navigation techniques that use signals of opportunity (e.g. TV/AM broadcast with known source locations) without aid satellite-based location systems GPS) are promising since they readily available strong not susceptible to blockage or jamming. However, problems such as large asynchronous clock offsets among receivers and signal frequency arise because dedicated for ranging purposes. In this paper, we propose an differential time-difference-of-arrival (TDOA) positioning method...
In this paper, a new signal-selective direction-of-arrival (DOA) tracking algorithm is presented for moving sources emitting narrowband or wideband cyclostationary signals. algorithm, DOAs of the are updated recursively based on most current array output in way that no data association needed. Furthermore, by exploiting cyclostationarity, interference and noise suppressed. Thus, only interest tracked. It also shown via Kalman filter, performance could be further improved. The effectiveness...
Most eigenstructure-based blind channel identification and equalization algorithms with second-order statistics need SVD or EVD of the correlation matrix received signal. In this paper, we address new based on QR factorization signal directly without calculating matrix. This renders factorization-based more robust against ill-conditioned channels, i.e., those channels almost common zeros among subchannels. First, present a block algorithm that performs data as whole. Then, recursive is...
In this paper, we develop delta operator based Levinson and Schur type on-line RLS algorithms. Such algorithms have the potential of improved numerical behavior for ill-conditioned input data. These new are obtained by a unified transformation on existing q ones. We first show that conventional lattice structure can be naturally derived when backward is used. With operator, stationary stochastic model in q-domain easily transformed into domain. Then, same transformation, will applied to...
One main drawback of using time difference arrival (TDOA) methods for source localization and navigation is that they require centralization multiple copies a signal. This paper considers blindly estimating the location cyclic prefix (CP) in an orthogonal frequency division multiplexing (OFDM) signal, enabling distributed TDOA computation up to integer ambiguity. ambiguity can be resolved least-squares methods, if enough TDOAs are available, requiring only minimal cooperation between...
We describe adaptive methods for estimating FIR zero-forcing blind equalizers with arbitrary delay directly from the linear predictions of observations. While most current require inversion or singular value decomposition (SVD) correlation matrix, our need only to solve two prediction problems. They can be implemented as RLS LMS algorithms recursively update equalizer estimation. they are computationally efficient. The computational complexity in each recursion less than 15(LN)/sup 2/ case,...
The evolution of data driven optimization has been shown advantageous in many applications. In this paper, we propose a deep learning architecture for the wireless communications receiver to optimize detection terms Bit Error Rate (BER), which enables reliable communication over AWGN and multipath Rayleigh fading channels with bandwidth constraints. Our approach uses autoencoder estimate received signal along an additional layer symbol detection, are jointly trained fine-tuned. Significant...
Most existing zero-forcing equalization algorithms rely either on higher than second-order statistics or partial complete channel identification. We describe methods for computing fractionally spaced blind equalizers with arbitrary delay directly from of the observations without first develop a batch-type algorithm; then, adaptive are obtained by linear prediction and gradient descent optimization. Our do not require order estimation, nor rank estimation. Compared other statistics-based...
For direction of arrival (DOA) estimation under a multipath environment, algorithms exist to estimate the DOAs all paths associated with different sources, by employing various schemes cope coherent signals induced multipath. Most these methods do not provide any information on which source detected path belongs to, i.e., association. However, association is useful for identifying, locating, and tracking sources. In this letter, we propose novel method associate each based assumption that...
The use of gradient descent recursive algorithms in blind adaptive equalization requires a cost function with unique minimum such that the FIR equalizer setup removes sufficient intersymbol interference (ISI). A class functions is proposed. newly proposed based on minimizing difference between any two norms joint channel-equalizer impulse response, each raised to same power, i.e., /spl par//spl middot//spl par//sub p//sup zeta//-/spl par//sup zeta///sub q/, where p<q. set implementable...