• Journal of the Korean Wood Science and Technology
• /
• v.45 no.6
• /
• pp.746-752
• /
• 2017

Cross-laminated timber (CLT) is a relatively new engineered wood for timber construction. It is a great shear wall material. It was known that the shear performance of the CLT wall depends on the performance of connections. In connection, nail or screw has to be installed with a certain distance from the end of the timber. Current building code specifies the distance on the name of end distance. The end distance was decided as a minimum distance not to make splitting or tearing out in lumber or glued laminated timber. As a relatively new engineered wood, the end distance of CLT connection need to be identified because CLT is cross-wisely glued lumber products like plywood. Different from glued laminated timber or lumber, cross layer of CLT may prevent wood from splitting or tearing-out. As a result, the end distance of CLT was expected to be reduced than glued laminated timber. The shorter end distance may let more versatile connector design possible. In this study, prior to developing novel connection for CLT, the end distance of CLT connection was experimentally investigated to identify the end distance limitation. The experiments showed that the end distance can be reduced from 7D to 6D, in case of the tested CLT combination and screw in this study.

• IEMEK Journal of Embedded Systems and Applications
• /
• v.15 no.2
• /
• pp.51-60
• /
• 2020

Optical character recognition (OCR) has been studied for decades because it is very useful in a variety of places. Nowadays, OCR's performance has improved significantly due to outstanding deep learning technology. Thus, there is an increasing demand for commercial-grade but affordable OCR systems. We have developed a low-cost, high-performance OCR system for the industry with the cheapest embedded developer kit that supports GPU acceleration. To achieve high accuracy for industrial use on limited computing resources, we chose a state-of-the-art text recognition algorithm that uses an end-to-end deep learning network as a baseline model. The model was then improved by replacing the feature extraction network with the best one suited to our conditions. Among the various candidate networks, EfficientNet-B3 has shown the best performance: excellent recognition accuracy with relatively low memory consumption. Besides, we have optimized the model written in TensorFlow's Python API using TensorFlow-TensorRT integration and TensorFlow's C++ API, respectively.

• IEMEK Journal of Embedded Systems and Applications
• /
• v.18 no.4
• /
• pp.165-172
• /
• 2023

Heterogeneous computing is a technology that utilizes different types of processors to perform parallel processing. It maximizes task processing and energy efficiency by leveraging various computing resources such as CPUs, GPUs, and FPGAs. On the other hand, edge computing has developed with IoT and 5G technologies. It is a distributed computing that utilizes computing resources close to clients, thereby offloading the central server. It has evolved to intelligent edge computing combined with artificial intelligence. Intelligent edge computing enables total data processing, such as context awareness, prediction, control, and simple processing for the data collected on the edge. If heterogeneous computing can be successfully applied in the edge, it is expected to maximize job processing efficiency while minimizing dependence on the central server. In this paper, experiments were conducted to verify the feasibility of various parallel programming models on high-end and low-end edge devices by using benchmark applications. We analyzed the performance of five parallel programming models on the Raspberry Pi 4 and Jetson Orin Nano as low-end and high-end devices, respectively. In the experiment, OpenACC showed the best performance on the low-end edge device and OpenSYCL on the high-end device due to the stability and optimization of system libraries.

• The KIPS Transactions:PartC
• /
• v.16C no.3
• /
• pp.363-372
• /
• 2009

For large scale VOD service, cluster servers are spotlighted to their high performance and low cost. A cluster server usually consists of a front-end node and multiple back-end nodes. Though increasing the number of back-end nodes can result in the more QoS streams for clients, the possibility of failures in back-end nodes is proportionally increased. The failure causes not only the stop of all streaming service but also the loss of the current playing positions. In this paper, when a back-end node becomes a failed state, the recovery mechanisms are studied to support the unceasing streaming service. For the actual VOD service environment, we implement a cluster-based VOD servers composed of general PCs and adopt the parallel processing for MPEG movies. From the implemented VOD server, a video block recovery mechanism is designed on parity algorithms. However, without considering the architecture of cluster-based VOD server, the application of the basic technique causes the performance bottleneck of the internal network for recovery and also results in the inefficiency CPU usage of back-end nodes. To address these problems, we propose a new failure recovery mechanism based on the pipeline computing concept.

• Journal of KIISE:Information Networking
• /
• v.31 no.6
• /
• pp.608-619
• /
• 2004

Application layer multicast(ALM) has been developed as an alternative to support the multi-receiver applications due to the slow deployment of IP multicast. Unlike routers in IP multicast, end systems participating in a ALM session are not optimized for relaying data and have various processing and network capacity. Therefore, with the increased role of end systems, the delay incurred at the end systems occupy large portion of the end-to-end delay. In this paper, we model the end system delay of ALM sessions and analyze its impact on the performance of the existing ALM schemes. Furthermore, we propose an enhanced scheme, based on the existing mesh-based scheme, which reflects the end system delay We also evaluate the performance of the proposed scheme by simulation and show that it reduces the end-to-end delay compared with the existing scheme.

• Journal of Advanced Marine Engineering and Technology
• /
• v.41 no.1
• /
• pp.1-7
• /
• 2017

When implementing butt joint welding of two plates, it is useful to attach end-tabs made of a metal with high heat conductivity (e.g., copper) at the front and back sides of the welded plates to prevent the bead from rolling down and prevent defects that may occur at the tips of the weld zone. In this study, the fin shape, which is known to have good heat discharging characteristics by natural convection, has been applied to enhance the cooling performance of the end-tab. From both experiment and numerical analysis, it was confirmed that end-tabs with fin-shaped holes have better heat discharging performance than end-tabs without holes. Through thermal and fluid flow analysis, the cooling rates of end-tabs with different hole shapes were estimated in order to figure out characteristics of shape factor that are important for the heat discharging performance. As a result, we found that the structure including vertical fins with optimal fin gap was the best-performing shape.

• Korean Management Science Review
• /
• v.16 no.1
• /
• pp.25-38
• /
• 1999

기존의 데이터망과는 달리 B-ISDN은 데이터, variable-rate 비디오 등과 같은 bursty하고 variable bit rate를 갖는 트래픽원으로부터 smooth하고 constant bit rate를 갖는 트래픽원까지 광범위한 계층의 트래픽을 모두 수용해야만 한다. 게다가 이 망은 다양한 계층의 트래픽에게 일정 수준 이상의 품질, 즉 QoS(Quality of Service)를 보장해 주어야만 한다. 셀 손실이나 지연시간 그리고 지연시간 jittere등을 포함하는 QoS 성능척도들은 응용서비스 별로 서로 다를 것이고 특히 실시간(real-time) 응용 서비스를 지원하는 망에서의 QoS문제는 특별히 중요성을 더하게 된다. 따라서 통계적 다중화에 의해 얻어지는 망자원의 효율성을 확보하면서 end-to-end QoS 보장을 해주는 문제는 B-ISDN에서 매우 중요하나 아직 대부분 미해결 문제로 남아 있는 분야이다. 본 연구에서는 end-to-end 분석에 가장 큰 에로점으로 부각되는 multi-hop망 환경하에서 트래픽 분석에 초점을 맞추어 기존 연구결과등을 폭넓게 비교 분석하고 최근 연구 동향 및 추후 연구 사항을 제시하고자 한다.

• Journal of Digital Contents Society
• /
• v.3 no.1
• /
• pp.101-111
• /
• 2002

Lately development of network around Internet expands range of data traffic to multimedia information, and so for the guarantee of multimedia services end-to-end QoS(Quality of Service) must service because comparing with existing Internet service can not support For satisfying those QoS requirements, network have to guarantee not only network on parameter, such as delay, jitter, throughput but also system resources like CPU utilization, memory usage. Therefore it is urgent to develop QoS based middleware to distribute multimedia data and maximize network utilization in the limited resource environment. And it must be necessary of network to provide end-to-end QoS(Quality of Service) for multimedia applications. Multimedia applications want that QoS which satisfy their own service properties be guaranteed Then, We must analyze those necessary QoS requirements md define QoS parameter which specify as two viewpoint, user's and network's perspective. Therefore network provider supplying network for usual user and university, enterprise must want to find about their own network performance and problem. It is essential for network manager to want to use a tool like this. On the basis of technique about QoS test-bed in the AIM network, We studied on the method of QoS measurement and management about end-to-end connection in the Internet. We measured network status about end-to-end connection and analyze the result of performance.

• Phonetics and Speech Sciences
• /
• v.12 no.4
• /
• pp.63-71
• /
• 2020

Speech recognition is one of the areas actively commercialized using deep learning and machine learning techniques. However, the majority of speech recognition systems on the market are developed on data with limited diversity of speakers and tend to perform well on typical adult speakers only. This is because most of the speech recognition models are generally learned using a speech database obtained from adult males and females. This tends to cause problems in recognizing the speech of the elderly, children and people with dialects well. To solve these problems, it may be necessary to retain big database or to collect a data for applying a speaker adaptation. However, this paper proposes that a new end-to-end speech recognition method consists of an acoustic augmented recurrent encoder and a transformer decoder with linguistic prediction. The proposed method can bring about the reliable performance of acoustic and language models in limited data conditions. The proposed method was evaluated to recognize Korean elderly and children speech with limited amount of training data and showed the better performance compared of a conventional method.

• The Journal of Korea Robotics Society
• /
• v.13 no.4
• /
• pp.205-212
• /
• 2018

This paper deals with the paddle type end of arm tool for rescue robot instead of rescue worker in dangerous environments such as fire, earthquake, national disaster and defense. It is equipped at the dual arm manipulator of the rescue robot to safely lift up an injured person. It consists of the paddle for lifting person, sensors for detecting insertion of person onto the paddle, sensor for measuring the tilting angle of the paddle, and mechanical compliance part for preventing incidental injuries. The electronics is comprised of the DAQ module to acquire the sensors data, the control module to treat the sensors data and to manage the errors, and the communication module to transmit the sensors data. After optimally designing the mechanical and electronical parts, we successfully made the paddle type end of arm tool and evaluated its performance by using specially designed jigs. The developed paddle type end of arm tool is going to be applied to the rescue robot for performance verification through field testing.