• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.27 no.3B
• /
• pp.202-211
• /
• 2002

Because wireless links are subject to disturbances and failures, it is important to support disconnected operations in mobile computing. Many schemes have been proposed to support the efficient disconnected operations. In this paper, we analyze and measure the computation time including disconnected operations to evaluate the performance of mobile computing on error prone wireless links. Mobile computation consists of three states: data hoarding, disconnected operation, and block states. We estimate the computation time using various parameters; error rate and recovery rate of wireless link, hoarding overhead, logging overhead, and reintegration overhead, etc. We can choose efficient strategies for disconnected operations and predict the performance using the results of this analysis.

• The Journal of the Acoustical Society of Korea
• /
• v.25 no.5
• /
• pp.213-221
• /
• 2006

Phone segmentation of speech waveform is especially important for concatenative text to speech synthesis which uses segmented corpora for the construction of synthetic units. because the quality of synthesized speech depends critically on the accuracy of the segmentation. In the beginning. the phone segmentation was manually performed. but it brings the huge effort and the large time delay. HMM-based approaches adopted from automatic speech recognition are most widely used for automatic segmentation in speech synthesis, providing a consistent and accurate phone labeling scheme. Even the HMM-based approach has been successful, it may locate a phone boundary at a different position than expected. In this paper. we categorized adjacent phoneme pairs and analyzed the mismatches between hand-labeled transcriptions and HMM-based labels. Then we described the dominant error patterns that must be improved for the speech synthesis. For the experiment. hand labeled standard Korean speech DB from ETRI was used as a reference DB. Time difference larger than 20ms between hand-labeled phoneme boundary and auto-aligned boundary is treated as an automatic segmentation error. Our experimental results from female speaker revealed that plosive-vowel, affricate-vowel and vowel-liquid pairs showed high accuracies, 99%, 99.5% and 99% respectively. But stop-nasal, stop-liquid and nasal-liquid pairs showed very low accuracies, 45%, 50% and 55%. And these from male speaker revealed similar tendency.

• The Transactions of the Korean Institute of Power Electronics
• /
• v.15 no.5
• /
• pp.361-368
• /
• 2010

This paper presents high speed position controller using stepping motor. The proposed position controller has close loop and open loop mode. In the high speed region, torque angle which is controlled by PI controller and memory based look-up table, is used to keep the reference position. The memory based look-up table produces a torque angle according to motor speed, and the PI controller can compensate the torque angle error. So, the fast dynamic response can be expected in the same position error. The open loop control mode which is divided by 3-modes control the actual position in the low speed and small position error. Each open loop modes are designed to reduce position error and dynamic brake in the stop command. The proposed position control scheme is verified by the practical stepping motor.

• Journal of Intelligence and Information Systems
• /
• v.5 no.1
• /
• pp.95-101
• /
• 1999

In this paper, a comparison between different methods to combine predictions from neural networks will be given. These methods are bagging, bumping, and balancing. Those are based on the analysis of the ensemble generalization error into an ambiguity term and a term incorporating generalization performances of individual networks. Neural Networks and AI machine learning models are prone to overfitting. A strategy to prevent a neural network from overfitting, is to stop training in early stage of the learning process. The complete data set is spilt up into a training set and a validation set. Training is stopped when the error on the validation set starts increasing. The stability of the networks is highly dependent on the division in training and validation set, and also on the random initial weights and the chosen minimization procedure. This causes early stopped networks to be rather unstable: a small change in the data or different initial conditions can produce large changes in the prediction. Therefore, it is advisable to apply the same procedure several times starting from different initial weights. This technique is often referred to as training ensembles of neural networks. In this paper, we presented a comparison of three statistical methods to prevent overfitting of neural network.

• Proceedings of the KSR Conference
• /
• 2010.06a
• /
• pp.1685-1690
• /
• 2010

This paper describes the calculation and application of the speed profile for the train speed control when the train is operated by ATO(Automatic Train Operation). We propose a speed profile calculation method considering the speed limitation and the jerk limitation, in order to maintain the quality of the train automatic operation. In previous works, ATO calculates a desired speed profile along the time, and controls the train to follow the profile. In this case, it may be hard to follow the restrictive speed along the location or to stop the train with a low location error, because of the difference between the desired location at each time and the real location of the train due to the control error. In the proposed method in this paper, we calculate a desired speed profile along the time considering the speed and jerk limitation first, and derive a speed profile along the location using it. If the restrictive speed profile is changed or the train speed strayed from the speed profile, ATO system calculates new speed profile and applies it immediately. Because ATO system controls the train speed based on the train location, the accuracy of the train location control can be improved. A simulation system for the test of the automatic train operation using this method is designed.

• The Transactions of the Korean Institute of Electrical Engineers D
• /
• v.51 no.2
• /
• pp.78-85
• /
• 2002

We introduce a rotational decision-directed joint algorithm of blind equalization coupled with carrier recovery for 32-QAM demodulation with high symbol rate. The proposed carrier recovery, which we call a rotational decision-directed carrier recovery(RDDCR), removes the residual phase difference by rotating the decision boundary for the kth received symbol by the frequency detector output of the (k-1)th received symbol. Since the RDDCR includes the function of PLL loop filter by rotating the decision boundary, it gives a simpler demodulator structure. The rotational decision-directed blind equalization(RDDBE) with the rotated decision boundary based on the Stop-and-Go Algorithm(SGA) operated during tracking the frequency offset by the RDDCR and removes intersymbol interference due to multipaths and channel noise. Test results show that symbol error rate of $10^{-3}$ is obtained before the forward error correction when SNR equals 15dB with 150KHz of carrier frequency offset and two multipaths, which is the channel condition for 32-QAM receiver.

• Laser Solutions
• /
• v.14 no.2
• /
• pp.24-29
• /
• 2011

The design of a telecentric f-${\theta}$ lens with a field of view (FOV) $30^{\circ}$ and an effective focal length of 1000mm is presented. The optical stop is placed at the front plane and the design is based on a geometric ray tracing technique, and the designed system consists of a series of convex and concave lenses. The designed f-${\theta}$ lens showed a considerable reduction in weight with a simplified structure and resulted in a good performance in the designated FOV. Detail analysis of rays is also presented. 653nm (red laser), 586nm (green laser), and 468nm (blue laser) were simulated as a light source and image illuminating source. The developed optical design requires 7 pieces of lenses made of SF1, N-FK56, N-LAK33, and BK7 glass materials. With optimal parametric design, the effective focal length was calculated to be 974.839mm which is very close to the initial design target. For the manufacturing purpose, the dimensions of lens curvature and thickness were truncated with error ranging 0.1% to 3.2%. As a result, the overall error was calculated to be 3.2% which can be still tolerable for display, laser material, and machining processing.

• Journal of the Korean Society for Precision Engineering
• /
• v.12 no.4
• /
• pp.67-77
• /
• 1995

In modern industry the requirement of automation of manufacturing process increases so that unmanned system has been popular as an ultimate goal of modern manufacturing process. In unmanned manufacturing process the tool fracture is a very serious problem because it results in the damage of workpieces and can stop the operation of whole manufa- turing system. In this study, image processing technique is used to detect the fracture of insert tip of face milling using multi-point-tool. In order to acquire the image information of fracture shape of rotation insert tip. We set up the optical system using a light beam chopper. In this system we can reduce the image degradation generated from stopped image of rotating insert tip using image restoration technique. We calculated the mean square error to diagnose the condition of tool fracture, and determind the criteria of tool fracture using experimental and staticstical method. From the results of this study we've developed non- contact detection technique of tool fracture using image processing method and proposed the fracture direction of automation and unmanned system considering the optimal time of tool change milling.

• Geomechanics and Engineering
• /
• v.33 no.4
• /
• pp.353-367
• /
• 2023

The construction of a protective embankment is a suitable strategy to stop and control high-energy rock blocks' impacts during the rockfall phenomenon. In this paper, based on the discrete element numerical method, by modeling an existing embankment reinforced with geogrid, its stability status under the impact of a rock block with two types of low and high kinetic energy, namely 2402 and 4180 kJ, respectively, has been investigated. The modeling results show that the use of geogrid has caused the displacement in the front and back of the embankment to decrease by more than 30%. In this case, the reinforced embankment has stopped the rock block earlier. The displacements obtained from the DEM modeling are compared with the displacements measured from an actual practical experiment to evaluate the results' validity. Comparison between the results shows that the displacement values are close together, while the maximum percentage error in previous studies by an analytical method and the finite element method was 76.4% and 36.6%, respectively. Therefore, the obtained results indicate the discrete numerical method's high ability compared to other numerical and analytical methods to simulate and design the geogrid-reinforced soil embankment under natural disasters such as rockfall with a minor error.

• ETRI Journal
• /
• v.18 no.2
• /
• pp.53-60
• /
• 1996

Blind equalization is a technique for adaptive equalization of a communication channel without the aid of training sequences. This paper proposes a new blind equalization algorithm. The advantage of the new algorithm is that it has the lower residual error than the GA (proposed by Godard) and Sign_GA (proposed by Weerackody et al.). The superior performance of the proposed algorithm is illustrated for the 16-QAM signal constellation. A Rummler channel model is assumed as a transmission medium. The performance of the proposed algorithm is compared to the GA, Sign_GA and Stop & Go Algorithm (SGA). The simulation results demonstrate that an improvement in performance is achieved with the proposed equalization algorithm.