• Transactions on Electrical and Electronic Materials
• /
• v.6 no.6
• /
• pp.262-271
• /
• 2005

In order to achieve timely and accurate fault detection of plasma etching process, neural network based time series modeling has been applied to reactive ion etching (RIE) using two different in-situ plasma-monitoring sensors called optical emission spectroscopy (OES) and residual gas analyzer (RGA). Four different subsystems of RIE (such as RF power, chamber pressure, and two gas flows) were considered as potential sources of fault, and multiple degrees of faults were tested. OES and RGA data were simultaneously collected while the etching of benzocyclobutene (BCB) in a $SF_6/O_2$ plasma was taking place. To simulate established TSNNs as incipient fault detectors, each TSNN was trained to learn the parameters at t, t+T, ... , and t+4T. This prediction scheme could effectively compensate run-time-delay (RTD) caused by data preprocessing and computation. Satisfying results are presented in this paper, and it turned out that OES is more sensitive to RF power and RGA is to chamber pressure and gas flows. Therefore, the combination of these two sensors is recommended for better fault detection, and they show a potential to the applications of not only incipient fault detection but also incipient real-time diagnosis.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2011.10a
• /
• pp.616-618
• /
• 2011

본 논문에서는 음성의 특징벡터를 추출하여 음성인식을 위한 인식 알고리즘을 제안한다. 본 논문에서 제안하는 방법은 사람의 음성을 정규화하여 시간지연신경회로망을 사용하여 음성인식을 하는 인식 알고리즘이다. 본 논문에서는 시간지연신경회로망을 이용하여 입력되는 음성정보를 일정시간 동안 학습시킨 후에 새로이 입력되는 정보를 인식하는 수법이다. 본 실험에서는 음성인식률에 의하여 본 알고리즘의 유효성을 확인한다.

• The Journal of the Acoustical Society of Korea
• /
• v.25 no.6
• /
• pp.277-281
• /
• 2006

This paper presents a method for the automatic recognition of pitch accents over syllables. The method that we propose is based on the time-delay recursive neural network (TDRNN). which is a neural network classifier with two different representation of dynamic context: the delayed input nodes allow the representation of an explicit trajectory F0(t) along time. while the recursive nodes provide long-term context information that reflects the characteristics of pitch accentuation in spoken English. We apply the TDRNN to pitch accent recognition in two forms: in the normal TDRNN. all of the prosodic features (pitch. energy, duration) are used as an entire set in a single TDRNN. while in the distributed TDRNN. the network consists of several TDRNNs each taking a single prosodic feature as the input. The final output of the distributed TDRNN is weighted sum of the output of individual TDRNN. We used the Boston Radio News Corpus (BRNC) for the experiments on the speaker-independent pitch accent recognition. π 1e experimental results show that the distributed TDRNN exhibits an average recognition accuracy of 83.64% over both pitch events and non-events.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.18 no.5
• /
• pp.126-141
• /
• 2019

One of the Cooperative Intelligent Transportation System(C-ITS) priority services is collision prevention support service. Several studies have considered V2I2V communication-based collision prevention support services using Artificial Neural Networks(ANN). However, such services still show some issues due to a low penetration of C-ITS devices and large delay, particularly when loading massive traffic data into the server in the C-ITS center. This study proposes the Artificial Neural Network-based Collision Warning Service(ACWS), which allows upstream vehicle to update pre-determined weights involved in the ANN by using real-time sectional traffic information. This research evaluates the proposed service with respect to various penetration rates and delays. The evaluation result shows the performance of the ACWS increases as the penetration rate of the C-ITS devices in the vehicles increases or the delay decreases. Furthermore, it reveals a better performance is observed in more advanced ANN model-based ACWS for any given set of conditions.

• Journal of Electrical Engineering and Technology
• /
• v.11 no.6
• /
• pp.1812-1824
• /
• 2016

The neural dynamics underlying the causal network during motor planning or imagery in the human brain are not well understood. The lack of signal processing tools suitable for the analysis of nonlinear and nonstationary electroencephalographic (EEG) hinders such analyses. In this study, noise-assisted multivariate empirical mode decomposition (NA-MEMD) is used to estimate the causal inference in the frequency domain, i.e., partial directed coherence (PDC). Natural and intrinsic oscillations corresponding to the motor imagery tasks can be extracted due to the data-driven approach of NA-MEMD, which does not employ predefined basis functions. Simulations based on synthetic data with a time delay between two signals demonstrated that NA-MEMD was the optimal method for estimating the delay between two signals. Furthermore, classification analysis of the motor imagery responses of 29 subjects revealed that NA-MEMD is a prerequisite process for estimating the causal network across multichannel EEG data during mental tasks.

• Smart Media Journal
• /
• v.11 no.2
• /
• pp.70-76
• /
• 2022

IoT application development using a cloud server causes problems such as data transmission and reception delay, network traffic, and cost for real-time processing support in network connected hardware. To solve this problem, edge cloud-based platforms can use neuromorphic hardware to enable fast data transfer. In this paper, we propose a model optimization method for supporting spiking neural networks on FPGA hardware. We focused on auto-adjusting network model parameters optimized for neuromorphic hardware. The proposed method performs optimization to show higher performance based on user requirements for accuracy. As a result of performance analysis, it satisfies all requirements of accuracy and showed higher performance in terms of expected execution time, unlike the naive method supported by the existing open source framework.

• The Journal of Information Technology
• /
• v.2 no.2
• /
• pp.145-161
• /
• 1999

This paper proposes a high speed back propagation neural networks which uses the residue number system. making the high speed operation possible without carry propagation Consisting of MAC(Multiplication and Accumulation) operator unit using Residue number system and sigmoid function operator unit using Mixed Residue Conversion is designed, The Designed circuits are descripted by VHDL and synthesized by Compass tools. Result of simulations shows that critical path delay time is about 19nsec and the size can be reduced to 40% compared to the neural networks implemented by the real number operation unit. The proposed design circuits can be implemented in parallel distributed processing system with desired real time processing.

• Journal of the Institute of Convergence Signal Processing
• /
• v.3 no.2
• /
• pp.30-36
• /
• 2002

This paper describes a nonlinear adaptive noise canceler(ANC) using neural networks(NN) based on filter to make up for the drawback of the conventional ANC with the linear adaptive filter. The proposed ANC was tested its noise rejection performance using broadband time-varying noise signal and compared with the ANC of TDL linear filter. Experimental results show that in cases of nonlinear correlations between the noise of primary input and reference input, the neural network based ANC outperforms the linear ANC with respect to mean square error It is also verified that the recurrent NN adaptive filter is superior to the feedforward NN filter. Thus, we identify that the NN adaptive filter is more effective than the linear adaptive filter for rejection of broadband time-varying noise in the ANC.

• Journal of Information Processing Systems
• /
• v.15 no.5
• /
• pp.1029-1035
• /
• 2019

In recent years, artificial intelligence (AI) services have become one of the most essential parts to extend human capabilities in various fields such as face recognition for security, weather prediction, and so on. Various learning algorithms for existing AI services are utilized, such as classification, regression, and deep learning, to increase accuracy and efficiency for humans. Nonetheless, these services face many challenges such as fake news spread on social media, stock selection, and volatility delay in stock prediction systems and inaccurate movie-based recommendation systems. In this paper, various algorithms are presented to mitigate these issues in different systems and services. Convolutional neural network algorithms are used for detecting fake news in Korean language with a Word-Embedded model. It is based on k-clique and data mining and increased accuracy in personalized recommendation-based services stock selection and volatility delay in stock prediction. Other algorithms like multi-level fusion processing address problems of lack of real-time database.

• 제어로봇시스템학회:학술대회논문집
• /
• 2005.06a
• /
• pp.842-847
• /
• 2005

This paper proposes a novel real-time EMG pattern recognition for the control of a multifunction myoelectric hand from four channel EMG signals. To cope with the nonstationary signal property of the EMG, features are extracted by wavelet packet transform. For dimensionality reduction and nonlinear mapping of the features, we also propose a linear-nonlinear feature projection composed of PCA and SOFM. The dimensionality reduction by PCA simplifies the structure of the classifier, and reduces processing time for the pattern recognition. The nonlinear mapping by SOFM transforms the PCA-reduced features to a new feature space with high class separability. Finally a multilayer neural network is employed as the pattern classifier. We implement a real-time control system for a multifunction virtual hand. From experimental results, we show that all processes, including virtual hand control, are completed within 125 msec, and the proposed method is applicable to real-time myoelectric hand control without an operation time delay.