• Journal of the Society of Naval Architects of Korea
• /
• v.47 no.6
• /
• pp.831-835
• /
• 2010

Currently, engineering design is carried out in a distributed manner geographically or physically. This imposes new requirements on the computational environments, such as efficient integration and collaboration in the Internet and network environments. Meanwhile, Web-based distributed design has led new paradigms in design and manufacturing fields. For example, Web-based technologies have reduced the product development time and to ensure a competitive product in order to exchange and interact of real-time design information that integrates the distributed design environment between departments as well as companies via Internet and Web. Hence, efficient data communication for design information sharing is the basis for collaborative systems in the distributed environments. Design data communication techniques such as CORBA, DCOM, and JAVA RMI have been considered in the existing research, but these techniques have some disadvantages such as limitations of interoperability and firewall problems. This paper presents the application of engineering design problems in which distributed design information resources are integrated and exchanged using Web Service for supporting XML and HTTP without interoperability and firewall problems.

• Journal of KIISE:Software and Applications
• /
• v.27 no.2
• /
• pp.97-107
• /
• 2000

Incremental learning generally reduces training time and increases the generalization of a neural network by selecting training data incrementally during the training. However, the existing methods of incremental learning repeatedly evaluate the importance of training data every time they select additional data. In this paper, an incremental learning algorithm is proposed for pattern classification problems. It evaluates the importance of each piece of data only once before starting the training. The importance of the data depends on how close they are to the decision boundary. The current paper presents an algorithm which orders the data according to their distance to the decision boundary by using clustering. Experimental results of two artificial and real world classification problems show that this proposed incremental learning method significantly reduces the size of the training set without decreasing generalization performance.

• Journal of Mechanical Science and Technology
• /
• v.19 no.7
• /
• pp.1393-1404
• /
• 2005

Fatigue crack growth and life have been estimated based on established empirical equations. In this paper, an alternative method using artificial neural network (ANN) -based model developed to predict fatigue damages simultaneously. To learn and generalize the ANN, fatigue crack growth rate and life data were built up using in-plane bending fatigue test results. Single fracture mechanical parameter or nondestructive parameter can't predict fatigue damage accurately but multiple fracture mechanical parameters or nondestructive parameters can. Existing fatigue damage modeling used this merit but limited real-time damage monitoring. Therefore, this study shows fatigue damage model using backpropagation neural networks on the basis of X -ray half breadth ratio B / $B_o$, fractal dimension $D_f$ and fracture mechanical parameters can estimate fatigue crack growth rate da/ dN and cycle ratio N / $N_f$ at the same time within engineering limit error ($5\%$).

• Journal of the Korean Institute of Intelligent Systems
• /
• v.15 no.5
• /
• pp.552-559
• /
• 2005

Recurrent neural networks(RNNs) trained with gradient based such as real time recurrent learning(RTRL) has a drawback of slor convergence rate. This algorithm also needs the derivative calculation which is not trivialized in error back propagation process. In this paper a derivative free Kalman filter, so called the unscented Kalman filter(UKF), for training a fully connected RNN is presented in a state space formulation of the system. A derivative free Kalman filler learning algorithm makes the RNN have fast convergence speed and good tracking performance without the derivative computation. Through experiments of nonlinear channel equalization, performance of the RNNs with a derivative free Kalman filter teaming algorithm is evaluated.

• Advances in aircraft and spacecraft science
• /
• v.4 no.3
• /
• pp.297-316
• /
• 2017

With rapid growth in the complexity of large scale engineering systems, the application of multidisciplinary analysis and design optimization (MDO) in the engineering design process has garnered much attention. MDO addresses the challenge of integrating several different disciplines into the design process. Primary challenges of MDO include computational expense and poor scalability. The introduction of a distributed, collaborative computational environment results in better utilization of available computational resources, reducing the time to solution, and enhancing scalability. SORCER, a Java-based network-centric computing platform, enables analyses and design studies in a distributed collaborative computing environment. Two different optimization algorithms widely used in multidisciplinary engineering design-VTDIRECT95 and QNSTOP-are implemented on a SORCER grid. VTDIRECT95, a Fortran 95 implementation of D. R. Jones' algorithm DIRECT, is a highly parallelizable derivative-free deterministic global optimization algorithm. QNSTOP is a parallel quasi-Newton algorithm for stochastic optimization problems. The purpose of integrating VTDIRECT95 and QNSTOP into the SORCER framework is to provide load balancing among computational resources, resulting in a dynamically scalable process. Further, the federated computing paradigm implemented by SORCER manages distributed services in real time, thereby significantly speeding up the design process. Part 1 covers SORCER and the algorithms, Part 2 presents results for aircraft panel design with curvilinear stiffeners.

• Advances in aircraft and spacecraft science
• /
• v.4 no.3
• /
• pp.317-334
• /
• 2017

With rapid growth in the complexity of large scale engineering systems, the application of multidisciplinary analysis and design optimization (MDO) in the engineering design process has garnered much attention. MDO addresses the challenge of integrating several different disciplines into the design process. Primary challenges of MDO include computational expense and poor scalability. The introduction of a distributed, collaborative computational environment results in better utilization of available computational resources, reducing the time to solution, and enhancing scalability. SORCER, a Java-based network-centric computing platform, enables analyses and design studies in a distributed collaborative computing environment. Two different optimization algorithms widely used in multidisciplinary engineering design-VTDIRECT95 and QNSTOP-are implemented on a SORCER grid. VTDIRECT95, a Fortran 95 implementation of D. R. Jones' algorithm DIRECT, is a highly parallelizable derivative-free deterministic global optimization algorithm. QNSTOP is a parallel quasi-Newton algorithm for stochastic optimization problems. The purpose of integrating VTDIRECT95 and QNSTOP into the SORCER framework is to provide load balancing among computational resources, resulting in a dynamically scalable process. Further, the federated computing paradigm implemented by SORCER manages distributed services in real time, thereby significantly speeding up the design process. Part 1 covers SORCER and the algorithms, Part 2 presents results for aircraft panel design with curvilinear stiffeners.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.15 no.1
• /
• pp.20-34
• /
• 2021

Today, IoT devices are flooding, and traffic is increasing rapidly. The Internet of Things creates a variety of added value through connections between devices, while many devices are easily targeted by attackers due to security vulnerabilities. In the IoT environment, security diagnosis has problems such as having to provide different solutions for different types of devices in network situations where various types of devices are interlocked, personal leakage of security solutions themselves, and high cost, etc. To avoid such problems, a TCP-based active scan was presented. However, the TCP-based active scan has limitations that it is difficult to be applied to real-time systems due to long detection times. To complement this, this study uses UDP-based approaches. Specifically, a lightweight active scan algorithm that effectively identifies devices using UPnP protocols (SSDP, MDNS, and MBNS) that are most commonly used by manufacturers is proposed. The experimental results of this study have shown that devices can be distinguished by more than twice the true positive and recall at an average time of 1524 times faster than Nmap, which has a firm position in the field.

• ETRI Journal
• /
• v.44 no.5
• /
• pp.849-858
• /
• 2022

Dynamic voltage frequency scaling (DVFS) has been widely adopted for runtime power management of various processing units. In the case of neural processing units (NPUs), power management of neural network applications is required to adjust the frequency and voltage every layer to consider the power behavior and performance of each layer. Unfortunately, DVFS is inappropriate for layer-wise run-time power management of NPUs due to the long latency of voltage scaling compared with each layer execution time. Because the frequency scaling is fast enough to keep up with each layer, we propose a layerwise dynamic frequency scaling (DFS) technique for an NPU. Our proposed DFS exploits the highest frequency under the power limit of an NPU for each layer. To determine the highest allowable frequency, we build a power model to predict the power consumption of an NPU based on a real measurement on the fabricated NPU. Our evaluation results show that our proposed DFS improves frame per second (FPS) by 33% and saves energy by 14% on average, compared with DVFS.

• Journal of Korean Society for Quality Management
• /
• v.51 no.1
• /
• pp.119-129
• /
• 2023

Purpose: This study proposes a classification model for implementing condition-based maintenance (CBM) by monitoring the real-time status of a machine using acceleration sensor data collected from a vehicle. Methods: The classification model's performance was improved by applying Fourier transform to convert the acceleration sensor data from the time domain to the frequency domain. Additionally, the Generative Adversarial Network (GAN) algorithm was used to augment images and further enhance the classification model's performance. Results: Experimental results demonstrate that the GAN algorithm can effectively serve as an image augmentation technique to enhance the performance of the classification model. Consequently, the proposed approach yielded a significant improvement in the classification model's accuracy. Conclusion: While this study focused on the effectiveness of the GAN algorithm as an image augmentation method, further research is necessary to compare its performance with other image augmentation techniques. Additionally, it is essential to consider the potential for performance degradation due to class imbalance and conduct follow-up studies to address this issue.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2022.10a
• /
• pp.447-449
• /
• 2022

This study was based on Android, one of the representative mobile platforms with the characteristics of connecting to the network anytime, anywhere and flexible mobility. In addition, using a demand prediction algorithm that can know the data of defective products based on AI, we will study the real-time monitoring UI design method based on Android studio with demand prediction data and company time series data.