• The Journal of the Korea institute of electronic communication sciences
• /
• 제5권4호
• /
• pp.357-362
• /
• 2010

This paper proposes the speech and noise recognition system by using a neural network in order to detect the speech and noise sections at each frame. The proposed neural network consists of a layered neural network training by back-propagation algorithm. First, a power spectrum obtained by fast Fourier transform and linear predictive coefficients are used as the input to the neural network for each frame, then the neural network is trained using these power spectrum and linear predictive coefficients. Therefore, the proposed neural network can train using clean speech and noise. The performance of the proposed recognition system was evaluated based on the recognition rate using various speeches and white, printer, road, and car noises. In this experiment, the recognition rates were 92% or more for such speech and noise when training data and evaluation data were the different.

• Proceedings of the Korean Biophysical Society Conference
• /
• 한국생물물리학회 2003년도 정기총회 및 학술발표회
• /
• pp.77-77
• /
• 2003

Gene expression in a cell is regulated by mutual activations or repressions between genes. Identifying the gene regulation network will be one of the most important research topics in the post genomic era. We propose a linear dynamic model of gene regulation for the yeast cell cycle. A small gene network consisting of about 40 genes is reconstructed from the analysis of micro-array gene expression data of yeast S. cerevisiae published by P. Spellman et al. We show that the network construction is consistent with the result of the hierarchical cluster analysis.

• IE interfaces
• /
• 제17권4호
• /
• pp.472-481
• /
• 2004

In this paper, we propose a linear programming model of supply planning process for the supply chain collaboration strategy of a company. The amount of its supplying quantity relies on outsourcing suppliers heavily. Conversely, the revenues of those suppliers are highly dependent on the supplying quota from the supply network planning of the company. In order to keep the supply stable through collaboration, the company builds such a policy to guarantee the fairness on revenue between the supplies. For this, the supply network plan should keep the capacity utilization ratio even for all the suppliers. But the production capacities are different and the distribution of molds is disproportional through suppliers, so the supply network plan is not easily established with simple arithmetic processes. Therefore, we developed the linear programming model with those target function and constraints minimizing the costs for holding inventory and penalty of delayed delivery, simultaneously guaranteeing the even capacity utilization through suppliers. The proposed model has been applied to real case and the evaluation for the planning result from the model would be followed in order to make sure that our model guarantee on extracting the supply network plan subordinated to the policy. Also we mention about further studies for improvement of the model.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• 제41권6호
• /
• pp.611-619
• /
• 2016

Success in modern war depends on electronic warfare. Therefore, It is very important to identify the kind of Radar PRI modulations in a lot of Radar electromagnetic waves. In this paper, I propose an algorithm to identify Linear up Sliding PRI, Non-Linear up Sliding PRI and Linear Down Sliding PRI, Non-Linear Down Sliding PRI among many Radar pulses. We applied not only the TDOA(Time Difference Of Arrival) concept of Radar pulse signals incoming to antennas but also a rising and falling curve characteristics of those PRI's. After making a program by such algorithm, we input each 40 data to those PRI's identification programs and as a result, those programs fully processed the data in according to expectations. In the future, those programs can be applied to the ESM, ELINT system.

• The Transactions of the Korean Institute of Electrical Engineers D
• /
• 제52권12호
• /
• pp.702-707
• /
• 2003

This study describes a new technique for human sensibility evaluation using electroencephalogram(EEG). Production of EEG is assumed to be linear. The linear predictor coefficients and the linear cepstral coefficients of EEG are used as the feature parameters of sensibility and pattern classification performances of them are compared. Using the better parameter, a human sensibility evaluation algorithm is designed. The obtained results are as follows. The linear predictor coefficients showed the better performance in pattern classification than the linear cepstral coefficients. Then, using the linear predictor coefficients as the feature parameter, a human sensibility evaluation algorithm is developed at the base of a multi-layer neural network. This algorithm showed 90% of accuracy in comfortableness evaluation in spite of fluctuations in statistics of EEG signal.

• Journal of Institute of Control, Robotics and Systems
• /
• 제2권2호
• /
• pp.88-95
• /
• 1996

It is inevitable for local systems to have deviations which represent interactions and modeling errors originated from the decomposition process of a large scale system. This paper presents a decentralized control scheme for interconnected systems using local linear models and a neuro-coordinator. In the proposed method, the local system is composed of a linear model and unknown deviations caused by linearizing the subsystems around operating points or by estimating parameters of the subsystems. Because the local system has unmeasurable deviations we define a local reference model which consists of a local linear model and a neural network to estimate the deviations indirectly. The reference model is reformed into a linear model which has no deviations through a transformation of input variables and we obtain an optimum feedback control law which minimizes a local performance index. Finally, we derive a decentralized feedback control law which consists of local linear states and neural network outputs. In the decentralized control, the neuro-coordinator generates a corrective control signal to cancel the effect of deviations through backpropagation learning with the errors obtained from the differences of the local system outputs and reference model outputs. Also, the stability of local system is proved by the degree of learning of the neural network under an assumption on a neural network learning index. It is shown by computer simulations that the proposed control scheme can be applied successfully to the control of a biased two-link planar robot manipulator.

• Proceedings of the Korean Society of Precision Engineering Conference
• /
• 한국정밀공학회 2002년도 춘계학술대회 논문집
• /
• pp.73-77
• /
• 2002

The requirements for higher productivity call for high speed of the machine tool axes. Iron core type linear DC motor is growly accepted far a viable candidate of the high speed machine tool feed unit. LDM, however, has inherent disturbance force components: cogging and force ripple. These disturbance force directly affects tracking accuracy of the carrage and must be eliminated or reduced. Reducing motor ripple, this paper adapted the feed forward compensation method and neural network control. Experiments carried 7ut on the linear motor test setup show that this control methods is usable in order to reduce the motor ripple.

• Proceedings of the KIEE Conference
• /
• 대한전기학회 1995년도 하계학술대회 논문집 B
• /
• pp.710-713
• /
• 1995

Hopfield networks have been applied to the problem of linear system identification. In this paper, Hopfield network based parameter identification scheme of non-linear dynamic systems is proposed. Simulation results demonstrate that Hopfield network can be used effectively for the identification of non-linear systems assuming that the system states and their time derivatives are available. Therefore, the proposed scheme can be applied in fault detection and isolation(FDI) and adaptive control of non-linear systems where the Hopfield networks perform on-line identification of system parameters.

• The Journal of Information Technology
• /
• 제7권2호
• /
• pp.43-54
• /
• 2004

Since Dr. John J. Hopfield has proposed the HOpfield network, it has been widely applied to the pattern recognition and the routing optimization. The method of Jian-Hua Li improved efficiency of Hopfield network which input pattern's weights are regenerated by SVD(singluar value decomposition). This paper deals with Li's Hopfield Network by linear pre-processing. Linear pre-processing is used for increasing orthogonality of input pattern set. Two methods of pre-processing are used, Hadamard method and random method. In manner of success rate, radom method improves maximum 30 percent than the original and hadamard method improves maximum 15 percent. In manner of success time, random method decreases maximum 5 iterations and hadamard method decreases maximum 2.5 iterations.

• Journal of Communications and Networks
• /
• 제15권4호
• /
• pp.430-438
• /
• 2013

This paper addresses the issue of throughput-delay of one-to-many wireless multi-hop flows based on random linear network coding (RLNC). Existing research results have been focusing on the single-hop model which is not suitable for wireless multi-hop networks. In addition, the conditions of related system model are too idealistic. To address these limitations, we herein investigate the performance of a wireless multi-hop network, focusing on the one-to-many flows. Firstly, a system model with multi-hop delay was constructed; secondly, the transmission schemes of system model were gradually improved in terms of practical conditions such as limited queue length and asynchronous forwarding way; thirdly, the mean delay and the mean throughput were quantified in terms of coding window size K and number of destination nodes N for the wireless multi-hop transmission. Our findings show a clear relationship between the multi-hop transmission performance and the network coding parameters. This study results will contribute significantly to the evaluation and the optimization of network coding method.