• Proceedings of the Acoustical Society of Korea Conference
• /
• 1994.06c
• /
• pp.358-363
• /
• 1994

This paper presents the improvement of the SCHMM using discrete VQ and One-Pass algorithm for keywords recognition in Korean continuous speech. The SCHMM using discrete VQ is a simple model that is composed of a variable mixture gaussian probability density function with dynamic mixture number. One-Pass algorithm is improved such that recognition rates are enhanced by fathoming any undesirable semisyllable with the low likelihood and the high duration penalty, and computation time is reduced by testing only the frame which is dissimilar to the previously testd frame. In recognition experiments for speaker-dependent case, the improved One-Pass algorithm has shown recognition rates as high as 99.7% and has reduced compution time by about 30% compared with the currently abailable one-pass algorithm.

• International Journal of Internet, Broadcasting and Communication
• /
• v.12 no.3
• /
• pp.46-55
• /
• 2020

Ad Hoc network is a special wireless network, mainly because the nodes are no control center, the topology is flexible, and the networking could be established quickly, which results the transmission stability is lower than other types of networks. In order to guarantee the transmission of data packets in the network effectively, an improved Queue Ad Hoc On-demand Distance Vector Routing protocol (Q-AODV) for node detection by using blockchain technology is proposed. In the route search process. Firstly, according to the node's daily communication record the cluster is formed by the source node using the smart contract and gradually extends to the path detection. Then the best optional path nodes are chained in the form of Merkle tree. Finally, the best path is chosen on the blockchain. Simulation experiments show that the stability of Q-AODV protocol is higher than the AODV protocol or the Dynamic Source Routing (DSR) protocol.

• Corrosion Science and Technology
• /
• v.17 no.2
• /
• pp.81-89
• /
• 2018

In this study, the effect of Nb alloying element on the corrosion fatigue properties of high strength steel is investigated by conducting fatigue experiments under corrosive condition and hydrogen induced condition, potentiodynamic polarization test, tensile test and surface analyses. Nb element is added to enhance the mechanical property of medium carbon steel. This element forms MX-type phases such as carbides and nitrides which are playing an important role in the grain refinement. The grain refinement is one of the effective way to improve mechanical property because both tensile strength and toughness can be improved at the same time. However, MX-type phase precipitates can be a susceptible site to localized corrosion in corrosive environment due to the potential difference between matrix and precipitate. The obtained results showed that Nb-added steel improved corrosion fatigue property by grain refinement. However, it is degraded for hydrogen-induced fatigue property due to Nb, Ti-inclusions acting as a stronger trap.

• The KSFM Journal of Fluid Machinery
• /
• v.10 no.2 s.41
• /
• pp.46-53
• /
• 2007

An improved performance characteristics analysis model of a regenerative pump is proposed in the present paper. For its low characteristic speed, a regenerative pump generates high head with low flow rate. However, the efficiency is fairly low due to the skin friction between impeller and casing. Also, the complexity of its internal flow pattern makes prediction of performance characteristics difficult. In the present research, a one-dimensional analysis model was improved with consideration of disc friction loss, minor loss, and modified flow length, and the result was proven to be close in range with the results from experiments.

• Proceedings of the Korean Society of Precision Engineering Conference
• /
• 2001.10a
• /
• pp.214-219
• /
• 2001

For navigation of a mobile robot, it is one of the essential tasks of find out its current position. Dead reckoning is the most frequently used method to estimate its position. However conventional dead reckoner is prone to give us false information on the robot position especially when the wheels are slipping. This paper proposes an improved dead reckoning scheme using neural networks. The network detects the instance of wheel slipping and estimates the linear velocity of the wheel ; thus it calculates current position and heading angel of a mobile robot. The structure and variables of the neural network are chosen based on the analysis of slip motion robot. The structure and variables of the neural network are chosen based on the analysis of slip motion characteristics. A series of experiments are performed to investigate the performance of the improved dead reckoning system.

• Proceedings of the Korean Society of Precision Engineering Conference
• /
• 2004.10a
• /
• pp.162-165
• /
• 2004

On this study, we improved diaphgram for micro speaker performance using Taguchi method in discrete design space. The design of diaphgram has an effect on performance of micro speaker such as, thickness of diaphgram, shape of diaphgram, etc. Therefore this study carried to decide shape of diaphgram and thickness of diaphgram for minimizing 2nd natural frequency of diaphgram using Taguchi method. we showed improved design factors that minimized 2nd natural frequency of diaphgram. Also, 2nd natural frequency of diaphgram for micro speaker is reduced up to 37 percent maintaining twist mode shape. From the results of ANOVA, 2nd natural frequency of diaphgram for micro speaker have an effect on position of the outer curved shape and thickness of diaphgram.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.10 no.10
• /
• pp.1848-1853
• /
• 2006

In this paper, we propose a method to reconstruct resolution-improved 3D images computationally based on lenslet array model in integral imaging technique and analyze its performance. While conventional view-point reconstruction method has low resolution based on pinhole array model, the proposed method can obtain resolution-improved 3D images because of extracting multiple pixels from elemental images by use of lenslet way model. To show the usefulness of proposed method, we perform computational experiments and present its results.

• Journal of the Korean Society of Manufacturing Technology Engineers
• /
• v.25 no.1
• /
• pp.36-41
• /
• 2016

A method is proposed for the improvement of deposition reproducibility in the selective electrodeposition process using laser masking and DC voltage. Selective electrodeposition using laser masking and DC voltage can achieve a deposited layer with micro patterns. However, selective electrodeposition using laser masking and DC voltage have a critical problem: the lack of reproducibility in selective deposition. The reproducibility of selective electrodeposition can be improved by a new process that consists of laser masking, two-step electro-deposition, laser scribing, and ultrasonic cleaning. The experiments in this study show that the reproducibility of selective deposition can be successfully improved by the combination of two-step electrodeposition and laser scribing.

• The Journal of the Acoustical Society of Korea
• /
• v.16 no.2E
• /
• pp.32-37
• /
• 1997

This paper proposes a isolated speech recognition method of Korean digits using a TDNN(Time Delay Neural Network) which is able to recognizc time-varying speech properties. We also make an investigation of effect on network parameter of TDNN ; hidden layers and time-delays. TDNNs in our experiments consist of 2 and 3 hidden layers and have several time-delays. From experiment result, TDNN structure which has 2 hidden-layers, gives a good result for speech recognition of Korean digits. Mis-recognition by time-delays can be improved by changing TDNN structures and mis-recognition separated from time-delays can be improved by changing input patterns.

• Journal of the Semiconductor & Display Technology
• /
• v.12 no.1
• /
• pp.1-7
• /
• 2013

Grey-level statistical models have been widely used in many applications for object location and identification. However, conventional models yield some problems in model refinement when training images are not properly aligned, and have difficulties for real-time recognition of arbitrarily rotated models. This paper presents improved grey-level statistical models that align training images using image or feature matching to overcome problems in model refinement of conventional models, and that enable real-time recognition of arbitrarily rotated objects using efficient hierarchical search methods. Edges or features extracted from a mean training image are used for accurate alignment of models in the search image. On the aligned position and orientation, fitness measure based on grey-level statistical models is computed for object recognition. It is demonstrated in various experiments in PCB inspection that proposed methods are superior to conventional methods in recognition accuracy and speed.