• 대한전기학회:학술대회논문집
• /
• 대한전기학회 2003년도 학술회의 논문집 정보 및 제어부문 A
• /
• pp.263-266
• /
• 2003

In this paper, we have proposed a realization method of tactile sense of virtual objects using multi-layer Neural Networks(NN). Inputs of the NN are position data of non-rigid objects and outputs of the NN are forces at that time and point. First, the position and forte data are measured from non-rigid objects (a sponge and a balloon) using two PHANToMS, one as a master and the other as a slave manipulator, then the data are used to train a multi-layer Neural Networks whose inputs and outputs are designed to represent tactile information. The trained Neural Networks is used to regenerate the tactile sense on the virtual objects graphically made by a computer, and one can feel a quite similar sense of touch by using the system while touching the virtual objects.

• 전자공학회논문지B
• /
• 제30B권2호
• /
• pp.69-77
• /
• 1993

Automatic image recognition system is essential for a better man-to machine interaction. Because of the noise and deformation due to the sensor operation, it is not simple to build an image recognition system even for the fixed images. In this paper neural network which has been reported to be adequate for pattern recognition task is applied to the fixed and variational(rotation, size, position variation for the fixed image)recognition with a hope that the problems of conventional pattern recognition techniques are overcome. At fixed image recognition system. ART model is trained with face images obtained by camera. When recognizing an matching score. In the test when wigilance level 0.6 - 0.8 the system has achievel 100% correct face recognition rate. In the variational image recognition system, 65 invariant moment features sets are taken from thirteen persons. 39 data are taken to train multi-layer perceptron and other 26 data used for testing. The result shows 92.5% recognition rate.

• 한국철도학회:학술대회논문집
• /
• 한국철도학회 2007년도 춘계학술대회 논문집
• /
• pp.1475-1480
• /
• 2007

Numerical analysis has been conducted to simulate pedestrian flow in the model of two-storied subway station. Because almost all the subway stations are two or three storied structure, simulations are conducted for the passengers those who get off the train and pass the wicket. Passenger flow analysis is very important factor to design the station and also to manage the operation of subway system. In the subway station, pedestrians move to the horizontal directions as well as vertical ones. Therefore, to consider the movement of pedestrians is necessary for the guarantee of safety and conveniency. As the up and down floors are connected with step, escalator and elevator, the entire movements in the multi-storied station should be simulated as like a 3-dimensional flow. Numerical schemes for the directional sweeping are developed to prevent the dependency on physical structure of station and to determine primary direction and secondary one. By using the developed program, we compared the simulation results of the effects of the location and size of exit and elapsed time.

• 한국자동차공학회논문집
• /
• 제5권3호
• /
• pp.106-116
• /
• 1997

In this paper, using MATLAB SIMULINK, a generalized design methodology was suggested for multi pass transmission(MPT) by classifying the vehicle power train as prime mover, MPT and vehicle dynamics. This approach enables a designer to investigate the influence of each transmission component by simple combination of system components without changes of overall program. Using the design methodology, a MPT consisting of CVT, 2, clutches and reduction gears was designed for a braking energy regenerative flywheel hybrid vehicle. The CVT is essential in order to connect the engine and flywheel speed with the vehicle speed. For the purpose of smooth clutch operation, control algorithm was suggested by introducing dead zone for the clutch engagement. Using the SIMULINK model, performance of the flywheel hybrid vehicle with MPT was investigated. It was observed from the simulation results that the MPT vehicle showed better fuel economy, 47% than that of AT vehicle, 27% than that of CVT vehicle for ECE-15 driving cycle. Especially destinct fuel efficiency improvement was obtained for city driving cycle requiring more frequent stop and start.

• 한국철도학회:학술대회논문집
• /
• 한국철도학회 2008년도 춘계학술대회 논문집
• /
• pp.2185-2190
• /
• 2008

In a high speed train, multi-layered panels for floor, side wall and roof are important sound insulating part. As these multi-layered panels require high bending strength vs. weight, corrugated steels or aluminium honeycomb panel are generally used. However, with some inevitable factors, these panels show lower sound insulation performance than that of the plate with the same weight. Transmission loss(TL) often severely decreases in a particular frequency range because of the decrease of the critical frequency, occurrence of local resonance modes and cavity resonance modes, which are not shown in a plate. In this study, frequency range and cause of the TL drop are investigated on the corrugated and honeycomb panels.

• 한국시뮬레이션학회논문지
• /
• 제9권3호
• /
• pp.77-89
• /
• 2000

This paper suggests a multi-attribute dispatching rule for an automated guided vehicle(AGV). The attributes to be considered are the number of queues in outgoing buffers of workstations, distance between an idle AGV and a workstation with a job waiting for the service of vehicle, and the number of queues in input buffers of the destination workstation of a job. The suggested rule is based on the simple additive weighting method using a normalized score for each attribute. A neural network approach is applied to obtain an appropriate weight vector of attributes based on the current status of the manufacturing system. Backpropagation algorithm is used to train the neural network model. The proposed dispatching rules and some single attribute rules are compared and analyzed by simulation technique. A number of simulation runs are executed under different experimental conditions to compare the several performance measures of the suggested rules and some existing single attribute dispatching rules each other.

• 농촌지도와개발
• /
• 제25권3호
• /
• pp.121-133
• /
• 2018

In South Korea, a few kinds of social farming practice are identified: care farming, labour integration, and training in farming sector. Although social farming is not a prevailing activity in rural communities, it attracts much attention from a range of actors in society. In Hongseong-gun, from a few years ago, two farms began to care and employ the mentally disabled and to train young new comers who want to grow crops in the way of organic farming. Both of them are cooperatives, which were established by the residents want to participate in. These movement has made some changes in the community. And now, it became the well-known cases of social farming as well as multi-functional agriculture. Social farming can be described as a praxis to deterritorialize the units of agricultural production and the rural community, where food empires imposed their ordering principle upon units of agricultural production in order to appropriate the value added by farming.

• 말소리와 음성과학
• /
• 제9권2호
• /
• pp.95-101
• /
• 2017

This paper proposes a new method to train Deep Neural Network (DNN)-based acoustic models for speech recognition of native and foreign speakers. The proposed method consists of determining multi-set state clusters with various acoustic properties, training a DNN-based acoustic model, and recognizing speech based on the model. In the proposed method, hidden nodes of DNN are shared, but output nodes are separated to accommodate different acoustic properties for native and foreign speech. In an English speech recognition task for speakers of Korean and English respectively, the proposed method is shown to slightly improve recognition accuracy compared to the conventional multi-condition training method.

• 한국컴퓨터정보학회논문지
• /
• 제23권11호
• /
• pp.43-49
• /
• 2018

In this paper, we propose a displacement measurement method based on deep learning using image data obtained from tensile tests of a material specimen. We focus on the fact that the sequential images during the tension are generated and the displacement of the specimen is represented in the image data. So, we designed sample generation model which makes sequential images of specimen. The behavior of generated images are similar to the real specimen images under tensile force. Using generated images, we trained and validated our model. In the deep neural network, sequential images are assigned to a multi-channel input to train the network. The multi-channel images are composed of sequential images obtained along the time domain. As a result, the neural network learns the temporal information as the images express the correlation with each other along the time domain. In order to verify the proposed method, we conducted experiments by comparing the deformation measuring performance of the neural network changing the displacement range of images.

• 한국군사과학기술학회지
• /
• 제27권4호
• /
• pp.474-484
• /
• 2024

Reinforcement learning, which are also studied in the field of defense, face the problem of sample efficiency, which requires a large amount of data to train. Transfer learning has been introduced to address this problem, but its effectiveness is sometimes marginal because the model does not effectively leverage prior knowledge. In this study, we propose a stochastic initial state randomization(SISR) method to enable robust knowledge transfer that promote generalized and sufficient knowledge transfer. We developed a simulation environment involving a cooperative robot transportation task. Experimental results show that successful tasks are achieved when SISR is applied, while tasks fail when SISR is not applied. We also analyzed how the amount of state information collected by the agents changes with the application of SISR.