• 한국정밀공학회:학술대회논문집
• /
• 한국정밀공학회 1994년도 추계학술대회 논문집
• /
• pp.131-137
• /
• 1994

Detection of grinding trouble occuring during the grinding process is classified into two types, i.e, based on the quantitative and qualitative knowledge. But, since the grinding operation is especially related with a large amount of functional parameters, it is actually defficult to cope with the grinding troubles occuring during process. Therefore, grinding trouble-shooting has difficulty in satisfying the requirement from the user. To cope with the grinding troubles occuring during the process, the application of neural network is on effective way. In this study, we identify the four parameters derived from the AE(Acoustic Emission) signals and present the grinding trouble-shooting system utilizing a back-propagation model of the neural network.

• ETRI Journal
• /
• 제25권1호
• /
• pp.34-40
• /
• 2003

This paper considers a TMN-based management system for the management of public ATM switching networks using a four-level hierarchical structure consisting of one network management system, several element management systems, and several agent-ATM switch pairs. Using Jackson's queuing model, we analyze the effects of one TMN command on the performance of the component ATM switch in processing local calls. The TMN command considered is the permanent virtual call connection. We analyze four performance measures of ATM switches- utilization, mean queue length and mean waiting time for the processor directly interfacing with the subscriber lines and trunks, and the call setup delay of the ATM switch- and compare the results with those from Jackson's queuing model.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 2001년도 ICCAS
• /
• pp.113.1-113
• /
• 2001

The robust position control with the bound function of neural network structure is proposed for uncertain robot manipulators. The neural network structure presents the bound function and does not need the concave property of the bound function, The robust approach is to solve this problem as uncertainties are included in a model and the controller can achieve the desired properties in spite of the imperfect modeling. Simulation is performed to validate this law for four-axis SCARA type robot manipulators.

• Journal of Information Display
• /
• 제5권3호
• /
• pp.35-40
• /
• 2004

In this study, we collected data used to formulate the relationship between quantitative metrological parameters in CRT display and the perceived focus quality. Human perception of the focusing quality was evaluated in terms of user feedback scores regarding the character legibility from four highly trained inspectors. Thirteen CRT monitors from five different manufacturers were compared relatively with respect to the norm monitor. The profile of electron beam such as spot size and the shape of distribution made by electron beam, contrast, convergence of RGB beams, and luminance characteristics were measured using a precision measurement system. Linear regression analysis and artificial neural network models were used to formulate the relationship between human perception and the quantitative measurements. The accuracy of the formulated linear regression model ($R^2$=0.515) was not satisfactory but the nonlinear neural network model ($R^2$=0.716) was fairly convincing and robust even the utilized data included subjective differences.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제15권3호
• /
• pp.853-873
• /
• 2021

In recent years, deep reinforcement learning (DRL) models are enjoying great interest as their success in a variety of challenging tasks. Deep Q-Network (DQN) is a widely used deep reinforcement learning model, which trains an intelligent agent that executes optimal actions while interacting with an environment. This model is well known for its ability to surpass skilled human players across many Atari 2600 games. Although DQN has achieved excellent performance in practice, there lacks a clear understanding of why the model works. In this paper, we present a visual analytics system for understanding deep Q-network in a non-blind matter. Based on the stored data generated from the training and testing process, four coordinated views are designed to expose the internal execution mechanism of DQN from different perspectives. We report the system performance and demonstrate its effectiveness through two case studies. By using our system, users can learn the relationship between states and Q-values, the function of convolutional layers, the strategies learned by DQN and the rationality of decisions made by the agent.

• World Technopolis Review
• /
• 제5권2호
• /
• pp.108-118
• /
• 2016

There are very few studies on governance structure for the collaboration between HEIs and science and technology parks until today. Major activities between science parks and HEIs are R&D activities, collaborative researches, technology transfer, space provision for BIs and Technology BIs in the science parks, provision of technical, legal and financial services for start-ups and venture firms. Governance structure for the collaboration of high education institutes with science and technology parks is the handling of complexity and management of dynamic flows of collaboration between two groups. Three models on the governance structure for the collaboration are suggested in this study. The first model is a governance structure that links R&D system such as universities, public research institutes and private research institutes with industrial production cluster such as a group of companies and industrial parks. The second model is a governance structure that has four layers of hierarchy. This hierarchical governance model is composed of four levels of organizations such as central government, three actors, one center for collaboration and many individual research performers. The third model is a governance structure that networks all the stakeholders horizontally. Under this structure, governance is conducted by the network members with no separate and unique governance entity.

• 정보교육학회논문지
• /
• 제1권2호
• /
• pp.1-15
• /
• 1997

In this paper we present database model for Intelligent Peripheral (IP) of Advanced Intelligent Network (AIN). The new model is developed through four phase. (1) An information of AIN IP is classified that would be stored in AIN IP database as specialized resources, service. schema and system information. (2) The modeling criteria are developed that would be used to model information classified. (3) Object-oriented concepts are used in modelifl8 classified information according to modeling criteria captured. (4) Methods applied to developed model are grouped, and active-based mechnisms such as trigger and constraints are developed. These selected methods and attributes are encapsulated into objects. Consequently they compose an active object-oriented AIN IP database model.

• 한국산업융합학회 논문집
• /
• 제7권2호
• /
• pp.199-205
• /
• 2004

This paper describes the fault diagnosis method to order to construct CIM in complex system with hierarchical structure similar to human body structure. Complex system is divided into unit, item and component. For diagnosing this hierarchical complex system, it is necessary to implement a special neural network. Fault diagnosis system can forecast faults in a system and decide from the signal information of current machine state. Comparing with other diagnosis system for a single fault, the developed system deals with multiple fault diagnosis, comprising hierarchical neural network (HNN). HNN consists of four level neural network, i.e. first is fault symptom classification and second fault diagnosis for item, third is symptom classification and forth fault diagnosis for component. UNIX IPC is used for implementing HNN with multitasking and message transfer between processes in SUN workstation with X-Windows (Motif). We tested HNN at four units, seven items per unit, seven components per item in a complex system. Each one neural network represents a separate process in UNIX operating system, information exchanging and cooperating between each neural network was done by message queue.

• Geomechanics and Engineering
• /
• 제9권4호
• /
• pp.465-481
• /
• 2015

Three-dimensional simulation of flow through dam foundation is performed using finite element (Seep3D model) and artificial neural network (ANN) models. The governing and discretized equation for seepage is obtained using the Galerkin method in heterogeneous and anisotropic porous media. The ANN is a feedforward four layer network employing the sigmoid function as an activator and the back-propagation algorithm for the network learning, using the water level elevations of the upstream and downstream of the dam, as input variables and the piezometric heads as the target outputs. The obtained results are compared with the piezometric data of Shahid Abbaspour's Dam. Both calculated data show a good agreement with available measurements that demonstrate the effectiveness and accuracy of purposed methods.

• International Journal of Aeronautical and Space Sciences
• /
• 제17권1호
• /
• pp.64-72
• /
• 2016

The coverage area of a GNSS regional ionospheric delay model is mainly determined by the distribution of GNSS ground monitoring stations. Extrapolation of the ionospheric model data can extend the coverage area. An extrapolation algorithm, which combines observed ionospheric delay with the environmental parameters, is proposed. Neural network and least square regression algorithms are developed to utilize the combined input data. The bi-harmonic spline method is also tested for comparison. The IGS ionosphere map data is used to simulate the delays and to compute the extrapolation error statistics. The neural network method outperforms the other methods and demonstrates a high extrapolation accuracy. In order to determine the directional characteristics, the estimation error is classified into four direction components. The South extrapolation area yields the largest estimation error followed by North area, which yields the second-largest error.