A5101795039- Neural Networks and Applications
- Blind Source Separation Techniques
- Natural Language Processing Techniques
- Higher Education and Teaching Methods
- Advanced Algorithms and Applications
- Advanced Sensor and Control Systems
- Lexicography and Language Studies
- Image and Signal Denoising Methods
- Ideological and Political Education
- Language, Discourse, Communication Strategies
- Advanced Wireless Network Optimization
- Power Systems and Renewable Energy
- Multilingual Education and Policy
- Advanced Measurement and Detection Methods
- Advanced Wireless Communication Techniques
- Syntax, Semantics, Linguistic Variation
- Education and Work Dynamics
- Advanced Power Amplifier Design
- Swearing, Euphemism, Multilingualism
- PAPR reduction in OFDM
- Advanced Computational Techniques and Applications
- Civil and Geotechnical Engineering Research
- Advanced Adaptive Filtering Techniques
- Fault Detection and Control Systems
- Advanced Control Systems Optimization
China Southern Power Grid (China)
2024
Wuxi Vocational College of Science and Technology
2021
Guangzhou Vocational College of Science and Technology
2021
Institute of Physics
2013
Gannan Normal University
2012
Peking University
2004-2011
Center for Applied Linguistics
2007-2011
North China Electric Power University
2008-2009
Chiba University
1993-2007
Beijing University of Chinese Medicine
2007
This paper presents a new method for self-tuning control of nonminimum phase discrete-time stochastic systems using approximate inverse obtained from the least-squares approximation. We show how unstable pole-zero cancellations can be avoided, and that this has advantage being able to determine an system independently plant zeros. The proposed scheme uses only available input output data stability is analyzed. Finally, results computer simulation are presented effectiveness method.
It was assumed that output Q and dc-feed inductance are high for the design of class E amplifier with nonlinear shunt capacitance in previous papers. This paper presents a method amplifiers any finite capacitance. And values them clarified. By carrying out SPICE simulations, we show simulated results agree desired ones quantitatively denote validity our procedure.
This paper presents a method of MRAC (model reference adaptive control) for multi-input multi-output (MIMO) nonlinear systems using NNs (neural networks). The control input is given by the sum output NN network). used to compensate nonlinearity plant dynamics that not taken into consideration in usual MRAC. role construct linearized model minimizing error caused nonlinearities systems.
In signal processing applications and image applications, some problems are difficult to process based on the linear theory. these problems, nonlinear will be more suitable for obtaining better results. On other hand, neural networks with a input-output relation applied various such as pattern recognition estimation of systems. This paper presents method noise removal degraded images using networks. Furthermore, in order effectively apply different features from teacher images, is proposed...
Recurrent neural network (RNN) is a kind of with one or more feedback loops. In this paper, complex-valued fully connected RNN real-time recurrent learning presented for the equalization systems, such as quadrature amplitude modulation (QAM), in presence intersymbol interference and nonlinear distortions. Simulation results show that proposed scheme quite effective channel when facing
The idea of combining both wavelets and neural networks has resulted in the formulation wavelet network, whose basic functions are drawn from a family orthonormal wavelets(1), which absorbs advantage high resolution advantages learning feedforward networks. usual method to train is backpropagation (BP) algorithm described by Rumelhart et al. However, this converges slowly for large or complex problems. In paper, we propose network nonlinear time series prediction using Unscented Kalman...
In this paper, we propose a novel method for multiuser detection (MUD) based on the branch-and-bound (BBD) using user grouping. The proposed partitions K users into M disjoint groups and applies BBD to each group. Due grouping, number of applied group is less than K. Therefore, worst-case computational cost can be reduced dramatically compared with that method. For an improvement probability error (GDE), signals other are removed in parallel by solution decorrelating decision-feedback...
Based on the 47 U.S. BITs,and according to six investment restriction domains in negative lists,the items lists of BIT counter-parties could be sorted into four categories:items are exactly same as items;items different from but have implications with implication within domain lists;items exceed lists. This paper gives a comprehensive study sector and country distribution characteristics types counter-parties,and then analyzes negotiation difficulty listing industries by China China-U.S....