• Journal of the Institute of Convergence Signal Processing
• /
• v.4 no.4
• /
• pp.30-39
• /
• 2003

In this paper, we carried out the performance evaluation of HM-Net(Hidden Markov Network) speech recognition system for Korean speech databases. We adopted to construct acoustic models using the HM-Nets modified by HMMs(Hidden Markov Models), which are widely used as the statistical modeling methods. HM-Nets are carried out the state splitting for contextual and temporal domain by PDT-SSS(Phonetic Decision Tree-based Successive State Splitting) algorithm, which is modified the original SSS algorithm. Especially it adopted the phonetic decision tree to effectively express the context information not appear in training speech data on contextual domain state splitting. In case of temporal domain state splitting, to effectively represent information of each phoneme maintenance in the state splitting is carried out, and then the optimal model network of triphone types are constructed by in the parameter. Speech recognition was performed using the one-pass Viterbi beam search algorithm with phone-pair/word-pair grammar for phoneme/word recognition, respectively and using the multi-pass search algorithm with n-gram language models for sentence recognition. The tree-structured lexicon was used in order to decrease the number of nodes by sharing the same prefixes among words. In this paper, the performance evaluation of HM-Net speech recognition system is carried out for various recognition conditions. Through the experiments, we verified that it has very superior recognition performance compared with the previous introduced recognition system.

• Journal of Advanced Navigation Technology
• /
• v.13 no.1
• /
• pp.41-47
• /
• 2009

In this paper, a band pass filter with a rejection band is proposed for UWB(Ultra Wide Band) communication system. First, low pass filter accessed cut off frequency of 10.2 GHz was designed using structure stepped impedance. And high pass filter accessed cut of frequency of 3.2 GHz was designed using parallel short-stub. There was implemented composite connection of designed low pass filter and high pass filter. The relative dielectric constant, the height, the loss tangent of the PCB substrate were ${\varepsilon}_r$=2.2, h=0.508 mm and loss tangent = 0.0009 respectively. The fabricated band pass filter shows a compact size of 3 cm. The fabricated band pass filter was characterized using 37169A VNA(Vector Network Analyzer). And measured result were obtained 7.5 GHz of bandwidth and -10 dB of return loss and -3 dB of insertion loss from pass band. The result of the research can be used for the UWB communications and MIC/MMIC, RFIC system.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.3 no.1
• /
• pp.105-111
• /
• 1999

This Paper is about how to deside the best parameter of PID controller, using Genetic Algorithms and Neural Networks. Control by Genetic Algorithms, which is off-line pass, has weakness for disturbance. So we want to improve like followings by adding Neural Network to controller and putting it on line. First we find PID parameter by Genetic Algorithms in forward pass of Neural Network and set the best output condition according to the increasing number of generation. Second, we explain the adaptability for disturbance with simulation by correcting parameter by backpropagation learning rule by using the learning ability of Neural Network.

• Proceedings of the Korean Institute of Communication Sciences Conference
• /
• 1987.04a
• /
• pp.219-222
• /
• 1987

A Semi-cusion VLSI Digital Filter Design for TDM/FDM tran-smultiplexer is decribed. Using the polyphase network approach a filter bank composed of only all-pasdigital filter sections was designed. The use of all-pass filters as basic building blocks is shown to provide a Transmultiplexer structure that has low computational requirements low quanization noise and hign modularity. A design of 1st order 2nd All pass filter is done using COMS 2um double metal.

• Journal of Advanced Marine Engineering and Technology
• /
• v.21 no.1
• /
• pp.66-70
• /
• 1997

In this paper a new learning algorithm for curvature smoothing and improved generalization for multi-layer neural networks is proposed. To enhance the generalization ability a constraint term of hidden neuron activations is added to the conventional output error, which gives the curvature smoothing characteristics to multi-layer neural networks. When the total cost consisted of the output error and hidden error is minimized by gradient-descent methods, the additional descent term gives not only the Hebbian learning but also the synaptic weight decay. Therefore it incorporates error back-propagation, Hebbian, and weight decay, and additional computational requirements to the standard error back-propagation is negligible. From the computer simulation of the time series prediction with Santafe competition data it is shown that the proposed learning algorithm gives much better generalization performance.

• Journal of the Korean Institute of Telematics and Electronics C
• /
• v.34C no.5
• /
• pp.33-42
• /
• 1997

In this paper, the modified inverse chebyshev low-pass function is analyzed in the frequency domain, time domain, and sensitivity characteristics as compared with the classical inverse chebyshev function. Unlike the classical function, the modified function exhibits progressively diminishing ripples in the stopband. So, the modified function has a great attenuation throughout the stopband except at the vicinity of a stop frequency and can be realizable in the passive doubly-terminated ladder network for the even order. The poles of the modified function move towards real axis by the effect of diminishing ripples. Thus the pole-Q, which is one of the valuable measurements to estimage the function characteristics, is reduced without increasing order. In the frequency and can be realizable in the passive doubly-terminated ladder network to examine the magnitude and pole-Q sensitivities.

• Journal of Korea Society of Industrial Information Systems
• /
• v.19 no.2
• /
• pp.45-49
• /
• 2014

In this work, a phase shifter radio-frequency integrated chip (RFIC) using a simple all-pass network is presented. As a tuning element of the phase shifter RFIC, tunable capacitors with a multi-layer dielectric of a para-/ferro-/para-electrics using a high tunable BST ferroelectric and a low-loss BZN paraelectric thin film were utilized. In order to evaluate and analyze the fabricated phase shifter RFIC, the same elements such as an inductor and capacitor integrated into it are also fabricated and tested. The designed phase shifter RFIC was fabricated on a quartz substrate in the size of $1.16{\times}1.21mm^2$. As the test results, the maximum phase difference of $350^{\circ}$ is obtained at 15 V and its tuning frequency bandwidth is 90 MHz from 2.72 to 2.81GHz.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.7 no.1
• /
• pp.29-36
• /
• 1982

This paper investigates the utilization of the concept of FDNR first proposed by L. T. Bruton to eliminate the necessity of inductance in a network. A new FDNR with single OP-Amp is also introduces to simplify the network in this paper. The efficiency of the new contrivance is also proved through experimented confirmation of its frequency when applied to on active eliptic low pass filter.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.14 no.8
• /
• pp.1843-1848
• /
• 2010

The Key exchange protocols are very crucial tools to provide the secure communication in the broadband satellite access network. They should be required to satisfy various requirements such as security, Key confirmation, and Key freshness. In this paper, we propose Two authenticated key exchange protocols Two Pass EKE-E(Encrypted Key Exchange-Efficient) and Two Pass EKE-S(Encrypted Key Exchange-Secure) are introduced. A basic idea of the protocols is that a password can be represented by modular addition N, and the number of possible modular addition N representing the password is $2^N$ The Two Pass EKE-E is secure against the attacks including main-in-the-middle attack and off-line dictionary attack, and the performance is excellent so as beyond to comparison with other authenticated key exchange protocols. The Two Pass EKE-S is a slight modification of the Two Pass EKE-E. The Two Pass EKE-S provides computational in feasibility for learning the password without having performed off line dictionary attack while preserving the performance of the Two Pass EKE-E.

• Journal of Biomedical Engineering Research
• /
• v.10 no.3
• /
• pp.343-350
• /
• 1989

A new algorithm classifying ECG Arrhythmia signals using Back-propagation network is proposed. The base-line of ECG signal is detected by high pass filter and probability density function then input data are normalized for learning and classifying. In addition, ECG data are scanned to classify Arrhythmia signal which is hard to find R-wave. A two-layer perceptron with one hidden layer along with error back-propagation learning rule is utilized as an artificial neural network. The proposed algorithm shows outstanding performance under circumstances of amplitude variation, baseline wander and noise contamination.