• Journal of Institute of Convergence Technology
• /
• v.9 no.1
• /
• pp.1-6
• /
• 2019

Matrix multiplication is a fundamental mathematical operation that has numerous applications across most scientific fields. In this paper, we presents a parallel GPU computation algorithm for dense matrix-matrix multiplication using OpenGL compute shader, which can play a very important role as a fundamental building block for many high-performance computing applications. Experimental results on NVIDIA Quad 4000 show that the proposed algorithm runs about 208 times faster than previous CPU algorithm and achieves performance of 75 GFLOPS in single precision for dense matrices with matrix size 4,096. Such performance proves that our algorithm is practical for real applications.

• Journal of Institute of Control, Robotics and Systems
• /
• v.19 no.11
• /
• pp.968-973
• /
• 2013

In the transportation industry, especially in the shipbuilding process, 3D surface measurement of large-scale hull pieces is needed for fabrication and assembly. We suggest an efficient method for checking the shape of curved plates under the forming operation with short time by measuring 3D profiles along the multi lines of the target surface. For accurate profile reconstruction, 2D camera calibration and 3D calibration using gauge blocks were performed. The evaluation test shows that the measurement accuracy is within the boundary of tolerance required in the shipbuilding process.

• International Journal of Precision Engineering and Manufacturing
• /
• v.4 no.6
• /
• pp.50-60
• /
• 2003

A web-based dynamic simulation system, called O-DYN, for multibody systems is developed. All the interfaces of the system are accessible via web browsers, such as Netscape or Explorer. The system uses a block-diagram type O-DYN/Modeler developed in JAVA Applet as a preprocessor. The O-DYN postprocessor composed of O-DYN/Plotter and O-DYN/Animator is developed in JAVA Applet. The O-DYN/Solver for predicting the dynamic behavior is run on the web server. Anyone who wants to simulate the dynamics of multibody systems or share results data can access the analysis system over the internet regardless of their OS, platform, or location.

• Proceedings of the KOR-KST Conference
• /
• 1998.10a
• /
• pp.3-10
• /
• 1998

본 연구에서는 동적통행시간예측을 위한 하나의 방법으로 블록밀도법을 도입하여 가로상의 일정한 구간을 세분화하고 몇 개의 블록으로 분할한 후 교통류를 유체근사화시키고 각 블록의 밀도를 일정시간 마다 갱신해 나가는 방법을 채택하였다. 즉, 각 블록에서 주어진 밀도의 초기치와 최하류부의 블록에서 유출교통량을 이용하여 일정시간 간격으로 모든 블록에서의 밀도를 수정해 가는 유체의 연속방정식의 개념을 도입하였다. 또한, 본 연구에서는 첨단교통시스템에 적용될 동적통행시간 예측을 위해 기존의 연속교통류만을 대상으로 하던 것에서 벗어나 신호등을 포함한 단속교통류를 대상으로 하였다. 또한 교통류의 저밀도 구간과 고밀도 구간을 분리하여 Ele복합모형을 적용하여 교통류를 해석하고 가급적 실제 상황과 유사하게 근접시키고자 하였다. 이 이론을 근거로 구축된 모델에 실제 현장에서 얻어진 교통량, 밀도, 속도 등의 자료들을 투입하고 통행시간의 예측을 도모하였다.

• Journal of Korean Institute of Industrial Engineers
• /
• v.23 no.2
• /
• pp.359-370
• /
• 1997

We consider a facility layout problem with the objective of minimizing total transportation distance, which is the sum of rectilinear distances between facilities weighted by the frequency of trips between the facilities. It is assumed that facilities are required to have rectangular shapes and there is no empty space between the facilities in the layout. In this study, a graph theoretic heuristic is developed for the problem. In the heuristic, planar graphs are constructed to represent adjacencies between the facilities and then the graphs are converted to block layouts on a continual plane using a layout construction module. (Therefore, each graph corresponds to a layout.) An initial layout is obtained by constructing a maximal weighted planar graph and then the layout is improved by changing the planar graph. A simulated annealing algorithm is used to find a planar graph which gives the best layout. To show the performance of the proposed heuristic, computational experiments are done on randomly generated test problems and results are reported.

• Proceedings of the KSR Conference
• /
• 2005.11a
• /
• pp.953-957
• /
• 2005

Light Rail Transit(LRT) Systems with transport capacity between subway and bus(5,000-25,000 persons per hour) are being carried on over 100 lines around the world. In Korea, to solve the urban transportation problem, the introduction of LRT system has been proceeded positively. It is planned to develop the Korean standard LRT system in which safety, efficiency and cost effectiveness are emphasized. So we were able to make proto type of Rubber tired AGT system for LRT. This is capable of driverless operation using CBTC(Communication Based Train Control) of Moving Block System and is currently making an experiment for reliability in test-line. This study is focused on verifving the train tracking with CBTC in driverless and ATO mode through implementing the examination.

• Transactions of the Korean Society of Automotive Engineers
• /
• v.23 no.3
• /
• pp.247-253
• /
• 2015

Platform standardization is widely used in commercial auto industries and other industries to acquire cost reduction as well as high reliability. This paper reviews some application examples and estimates expected effect of Korean army's platform standardization for combat system and combat support system. Also, military platform standardization based on utilization of commercial vehicles has been studied with the standard vehicles used in Korean army. It is estimated that outstanding results may be achieved if a dual platform, the public and military platforms, is utilized. So, when the strategy of the military transportation system is decided, "dual use" concept is necessary and this could maximize the effect of the military strategy reinforcement.

• Journal of the Korean Institute of Navigation
• /
• v.18 no.4
• /
• pp.49-57
• /
• 1994

In marine transportation of bulk cargoes such as crude oil. ore, coal etc., a lot of full form ship which have poor manoeuvrability were presented in many countries. Since ship manoeuvrability depends upon many parameters namely hydrodynamic derivatives, interference factors etc., as external forces, it is of great importance that we investigate these values of parameters on analysis of manoeuvrability. In this paper, we investigated and analyzed interaction coefficients among hull-propeller-rudder for a full form ship by captive model test in circulating water channel, and then compared with experimental results by PMM test. A tanker model ship which has 0.83 as block coefficient and MMG mathematical models were used in this experiment. Almost same tendencies were found in qualitative analysis, even though more serial experiments were demanded in quantitative analysis.

• Proceedings of the KIEE Conference
• /
• 2005.10c
• /
• pp.264-269
• /
• 2005

Light Rail Transit(LRT) Systems with transport capacity between subway and bus(5,000-25,000 persons per hour) are being carried on over 100 lines around the world, In Korea, to solve the urban transportation problem, the introduction of LRT system has been proceeded positively. It is planned to develop the Korean standard LRT system in which safety, efficiency and cost effectiveness are emphasized. So we were able to make proto type of Rubber tired AGT system for LRT. This is capable of driverless operation using CBTC(Communication Based Train Control) of Moving Block System and is currently making an experiment for reliability in test-line. This study is focused on verifving the performance of CBTC system for driverless and ATO mode through implementing the examination.

• Journal of Institute of Convergence Technology
• /
• v.8 no.1
• /
• pp.15-20
• /
• 2018

Recently a GPU has acquired programmability to perform general purpose computation fast by running thousands of threads concurrently. This paper presents a parallel GPU computation algorithm for dense matrix-matrix addition and scalar multiplication using OpenGL compute shader. It can play a very important role as a fundamental building block for many high-performance computing applications. Experimental results on NVIDIA Quad 4000 show that the proposed algorithm runs 21 times faster than CPU algorithm and achieves performance of 16 GFLOPS in single precision for dense matrices with size 4,096. Such performance proves that our algorithm is practical for real applications.