• The Journal of the Acoustical Society of Korea
• /
• v.16 no.5
• /
• pp.112-121
• /
• 1997

In this paper, we study a model parameter compensation method for noise-robust speech recognition. We study model parameter compensation on a sentence by sentence and no other informations are used. Parallel model combination(PMC), well known as a model parameter compensation algorithm, is implemented and used for a reference of performance comparision. We also propose a modified PMC method which tunes model parameter with an association factor that controls average variability of gaussian mixtures and variability of single gaussian mixture per state for more robust modeling. We obtain a re-estimation solution of environmental variables based on the expectation-maximization(EM) algorithm in the cepstral domain. To evaluate the performance of the model compensation methods, we perform experiments on speaker-independent isolated word recognition. Noise sources used are white gaussian and driving car noise. To get corrupted speech we added noise to clean speech at various signal-to-noise ratio(SNR). We use noise mean and variance modeled by 3 frame noise data. Experimental result of the VTS approach is superior to other methods. The scheme of the zero order VTS approach is similar to the modified PMC method in adapting mean vector only. But, the recognition rate of the Zero order VTS approach is higher than PMC and modified PMC method based on log-normal approximation.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.42 no.1
• /
• pp.14-24
• /
• 2017

In this paper, we propose a new EIM(erasure insertion method) based on the average-minimal-noise-power for HOM(higher order modulation) over AWGN(additive white Gaussian noise) under PBJ(partial-band jamming). Then we design SFH/SS(slow-frequency-hopping spread-spectrum) system by applying this method and formulate the PER(packet error rate) of the system. Based on this formula, we propose a new method to set the optimal threshold of the EIM and verify it at the designed 16-QAM SFH/SS system.

• Journal of the Korea Institute of Military Science and Technology
• /
• v.4 no.1
• /
• pp.147-154
• /
• 2001

The underwater explosion shock test is performed for the evaluation of the shock-resistant capability which is a very critical factor considering the survivability of the battle ship. Some measured signals have impulsive noise and gaussian white noise because of the unstable power supply system and the transient movement of cables during the underwater explosion shock test. The advanced shock signal analysis method which remove the noise of measured signal using the threshold policy of the median filter and the orthogonal wavelet coefficients are proposed. It is verified that the signal-to-noise ratio was improved about 30㏈ by the numerical simulation.

• Journal of the Korean Institute of Telematics and Electronics B
• /
• v.32B no.6
• /
• pp.876-877
• /
• 1995

The techniques for the estimation of motion vector from the image sequence assume implicitly that the intensity of image is constant through the time. But this assumption can be distored by such causes as the added noises and the sub-pel motion following the sampling, and the errors can be generated on the motion estimation by the change of intensity. In this paper, we analyzed theoretically the effect of the change of intensity by the noise on the motion estimation with the white Gaussian noise. We know a fact that the signal may be fluctuated to reduce the effect of the noise and so the sampling rate have to make down. Also we confirmed the theoretically analysis through the experiments which investigated the relation between the noises and the sampling rates.

• Proceedings of the KIEE Conference
• /
• 1992.07a
• /
• pp.305-307
• /
• 1992

This paper presents a technique, which identifies both process noise covariance and sensor noise covariance by using innovation correlation test. A correlation test, which checks whether the square root Kalman filter is workingly optimal or not, is given. The system is stochastic autoregressive moving-average model with auxiliary white noise Input. The linear quadratic Gaussian control is used for minimizing stochastic cost function. This paper indentifies Q, R, and estimates parametric matrics $A(q^{-1}),B(q^{-1}),C(q^{-1})$ by means of extended recursive least squares and model reference control. And The proposed technique has been validated in simulation results on the fourth order system.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.5 no.5
• /
• pp.853-860
• /
• 2001

This paper presents the new two denoising methods using wavelet. One is new spatially selective noise filtration(NSSNF) using spatial correlation and the other is undecimated discrete wavelet transform (UDWT) threshold-based. NSSNF got the flexible gain special property of SNR adding new parameter at the existing SSNF and UDWT had superior denosing effect than orthogonal wavelet transform(OWT) applied soft-threshold by applied hard-threshold. We selected additive white gaussian noise(AWGN) in this test environment. Also we analyzed and compared ousting denoising method using SNR as standard of judgement of improvemental effect.

• ETRI Journal
• /
• v.29 no.5
• /
• pp.607-613
• /
• 2007

Most signal-to-noise ratio (SNR) estimation techniques in digital communication channels derive the SNR estimates solely from samples of the received signal after the matched filter. They are based on symbol SNR and assume perfect synchronization and intersymbol interference (ISI)-free symbols. In severe channel distortion where ISI is significant, the performance of these estimators badly deteriorates. We propose an SNR estimator which can operate on data samples collected at the front-end of a receiver or at the input to the decision device. This will relax the restrictions over channel distortions and help extend the application of SNR estimators beyond system monitoring. The proposed estimator uses the characteristics of the second order moments of the additive white Gaussian noise digital communication channel and a linear predictor based on the modified-covariance algorithm in estimating the SNR value. The performance of the proposed technique is investigated and compared with other in-service SNR estimators in digital communication channels. The simulated performance is also compared to the Cram$\acute{e}$r-Rao bound as derived at the input of the decision circuit.

• Journal of Korean Society of Steel Construction
• /
• v.13 no.5
• /
• pp.525-533
• /
• 2001

The wind resistant design of long-span bridges has urged a special attention to the prevention of the flutter occurrence Therefore calculation of flutter derivatives is indispensable to this prediction. A used system identification method must identify all the flutter derivatives from noisy experimental data In this paper MITD(Modified Ibrahim Tim Domain) method and AKF (Adaptive Kalman Filter) method are applied to extract flutter derivatives from section-model tests. The robustness and reliability of proposal SI methods under a high signal-to-noise ratio is demonstrated through numerical simulation for windtunnel test.

• Proceedings of the IEEK Conference
• /
• 2004.08c
• /
• pp.645-648
• /
• 2004

In this paper, we drive a theoretical performance of complementary code keying (CCK) codes on additive white Gaussian noise (AWGN) channel. The CCK codes can be demodulated by the optimal maximum likelihood decoding method and sub-optimal correlation magnitude decoding algorithm. We calculate the bit error rate (BER) and symbol or codeword error rate (SER) of the CCK codes using the above mentioned two decoding algorithms. To derive the error performance, we use the weigh distributions and cross-correlation distributions of CCK codes.

• Proceedings of the IEEK Conference
• /
• 2004.08c
• /
• pp.663-666
• /
• 2004

In this paper, a modified-SOYA (soft output viterbi algorithm) of turbo codes is proposed for quantized channel receiver filter outputs. We derive optimum branch values for the Viterbi algorithm. Here we assume that received filter outputs are quantized and the channel is additive white Gaussian noise. The optimum non-uniform quantizer is used to quantize channel receiver filter outputs. To compare the BER (bit error rate) performance we perform simulations for the modified SOYA algorithm and the general SOYA