• 정보통신설비학회논문지
• /
• 제10권1호
• /
• pp.7-13
• /
• 2011

Matrix multiplication is a fundamental operation of linear algebra and arises in many areas of science and engineering. This paper introduces an efficient parallel matrix multiplication scheme on N ${\times}$ N mesh-connected SIMD array processor, called multiple hierarchical SIMD architecture (HMSA). The architectural characteristic of HMSA is the hierarchically structured control units which consist of a global control unit, N local control units configured diagonally, and $N^2$ processing elements (PEs) arranged in an N ${\times}$ N array. PEs are communicating through local buses connecting four adjacent neighbor PEs in mesh-torus networks and global buses running across the rows and columns called horizontal buses and vertical buses, respectively. This architecture enables HMSA to have the features of diagonally indexed concurrent broadcast and the accessibility to either rows (row control mode) or columns (column control mode) of 2D array PEs alternately. An algorithmic mapping method is used for performance evaluation by mapping matrix multiplication on the proposed architecture. The asymptotic time complexities of them are evaluated and the result shows that paralle matrix multiplication on HMSA can provide significant performance improvement.

• 한국전자통신학회논문지
• /
• 제14권6호
• /
• pp.1207-1214
• /
• 2019

4차 산업혁명의 기술이 우리도 모르는 사이 우리의 삶 속으로 스며들고 있다. CNN이 이미지 인식 분야에서 탁월한 능력을 보여준 이후 많은 IoT 기반 홈보안 시스템은 침입자로부터 가족과 가정을 보호하며 얼굴을 인식하기 위한 좋은 생체인식 방법으로 CNN을 사용하고 있다. 본 논문에서는 2D와 2.5D 이미지에 대하여 여러 종류의 입력 이미지 크기와 필터를 가지고 있는 CNN의 구조를 연구한다. 실험 결과는 50*50 크기를 가진 2.5D 입력 이미지, 2 컨벌류션과 맥스풀링 레이어, 3*3 필터를 가진 CNN 구조가 0.966의 인식률을 보여 주었고, 1개의 입력 이미지에 대하여 가장 긴 CPU 소비시간은 0.057S로 나타났다. 홈보안 시스템은 좋은 얼굴 인식률과 짧은 연산 시간을 요구하므로 본 논문에서 제안한 구조의 CNN은 홈보안 시스템에서 얼굴인식을 기반으로 하는 액추에이터 제어 등에 적합한 방법이 될 것이다.

• 산업경영시스템학회지
• /
• 제47권2호
• /
• pp.155-167
• /
• 2024

Distribution and logistics industries contribute some of the biggest GDP(gross domestic product) in South Korea and the number of related companies are quarter of the total number of industries in the country. The number of retail tech companies are quickly increased due to the acceleration of the online and untact shopping trend. Furthermore, major distribution and logistics companies try to achieve integrated data management with the fulfillment process. In contrast, small and medium distribution companies still lack of the capacity and ability to develop digital innovation and smartization. Therefore, in this paper, a deep learning-based demand forecasting & recommendation model is proposed to improve business competitiveness. The proposed model is developed based on real sales transaction data to predict future demand for each product. The proposed model consists of six deep learning models, which are MLP(multi-layers perception), CNN(convolution neural network), RNN(recurrent neural network), LSTM(long short term memory), Conv1D-BiLSTM(convolution-long short term memory) for demand forecasting and collaborative filtering for the recommendation. Each model provides the best prediction result for each product and recommendation model can recommend best sales product among companies own sales list as well as competitor's item list. The proposed demand forecasting model is expected to improve the competitiveness of the small and medium-sized distribution and logistics industry.

• Structural Engineering and Mechanics
• /
• 제27권6호
• /
• pp.683-696
• /
• 2007

This article is concerned with the presentation of a time-domain BEM approach applied to the solution of the scalar wave equation for 2D problems. The basic idea is quite simple: the basic variables of the problem at time $t_n$ (potential and flux) are computed with the results related to the potential and to its time derivative at time $t_{n-1}$ playing the role of "initial conditions". This time-marching scheme needs the computation of the potential and its time derivative at all boundary nodes and internal points, as well as the entire discretization of the domain. The convolution integrals of the standard time-domain BEM formulation, however, are not computed; the matrices assembled, only at the initial time interval, are those related to the potential, flux and to the potential time derivative. Two examples are presented and discussed at the end of the article, in order to verify the accuracy and potentialities of the proposed formulation.

• 한국정보통신학회:학술대회논문집
• /
• 한국해양정보통신학회 2010년도 춘계학술대회
• /
• pp.176-179
• /
• 2010

In this paper, we propose a fast implementation of Bilateral filter for tone mapping algorithm. Bilateral filter is able to preserve detail while at the same time prevent halo-ing artifacts because of improper scale selection by ensuring image smoothed that not only depend on pixel closeness, but also similarity. We accelerate Bilateral filter by using a piecewise linear approximation and recursive Gaussian filter as its domain filter. Recursive Gaussian filter is scale independent filter that combines low cost 1D filter which makes this filter much faster than conventional convolution filter and filtering in frequency domain. The experiment results show that proposed method is simpler and faster than previous method without mortgaging the quality.

• 한국정보과학회:학술대회논문집
• /
• 한국정보과학회 2010년도 한국컴퓨터종합학술대회논문집 Vol.37 No.1(B)
• /
• pp.380-383
• /
• 2010

다양한 응용프로그램에서 효과적으로 2차원 영상을 처리하기 위해서는 여러 가지 기법들이 이용되는데 그 중 2차원 필터링은 가장 많이 사용되는 방법 중 하나이다. 2차원 필터링에서 회선처리는 수평과 수직 방향의 1차원 선형 필터를 이용하는 방법이다. 2차원 회선처리는 커다란 이미지 위를 커널이 움직이며 연산을 해야 하므로 연산량이 매우 많으며 메모리 접근을 많이 필요로 한다. 하지만 회선처리는 입력화소뿐 아니라 주변 화소 값까지 고려하는 지역적인 동작으로 인해 병렬화된 처리가 가능하다. 이에 본 논문에서는 메모리 접근을 줄이고 연산을 병렬적으로 처리함으로서 회선처리의 수행 시간을 개선하는 하드웨어 기반의 회선처리 방법을 제안한다.

• 한국지능시스템학회논문지
• /
• 제11권6호
• /
• pp.563-567
• /
• 2001

We generalize the widebane P0-weyl symbol (P0WS) and the widebane spreading function (WSF) using the generalized warping function . The new generalized P0WS and WSF are useful for analyzing system and communication channels producing generalized time shifts. We also investigated the relationship between the affine Wey1 symbol(AWS) and the P0WS. By using specific warping functions, we derive new P0WS and WSF as analysis tools for systems and communication channels with non-linear group delary characteristics. The new P0WS preserves specific types of changes imposed on random processes. The new WSF provides a new interpretation of output of system and communication channel as weighted superpositions of non-linear time shifts on the input. It is compared to the conventional method obtaining output of system and communication channel as a convention integration of the input with the impulse response of the system and the communication channel. The convolution integration can be interpreted as weighted superpositions of liner time shifts on the input where the weight is the impulse response of the system and the communication channel. Application examples in analysis and detection demonstrate the advantages of our new results.

• 한국우주과학회:학술대회논문집(한국우주과학회보)
• /
• 한국우주과학회 2009년도 한국우주과학회보 제18권2호
• /
• pp.36.3-37
• /
• 2009

This is an attempt to improve a formula to predict variations of geomagnetic storm indices (Dst) from solar wind parameters. A formula which is most widely accepted was given by Burton et al. (1975) over 30 years ago. Their formula is: dDst*/dt = Q(t) - Dst*(t)/$\tau$, where Q(t) is the Dst injection rate given by the convolution of dawn-to-dusk electric field generated by southward solar wind magnetic field and some response function. However, they did not clearly specify the response function. As a result, misunderstanding seems to be prevailing that the injection rate is proportional to the dawn-to-dusk electric field. In this study we tried to determine the response function by examining 12 intense geomagnetic storms with minimum Dst < -200 nT for which solar wind data are available. The method is as follows. First we assume the form of response function that is specified by several time constants, so that we can calculate the injection rate Q1(t) from the solar wind data. On the other hand, Burton et al. expression provide the observed injection rate Q2(t) = dDst*/dt + Dst*(t)/$\tau$. Thus, it is possible to determine the time constants of response function by a least-squares method to minimize the difference between Q1(t) and Q2(t). We have found this simple method successful enough to reproduce the observed Dst variations from the corresponding solar wind data. The present result provides a scheme to predict the development of Dst 30 minutes to 1 hour in advance by using the real time solar wind data from the ACE spacecraft.

• 충청수학회지
• /
• 제17권1호
• /
• pp.85-101
• /
• 2004

Let A be a Hopf algebra with a linear form ${\sigma}:k{\rightarrow}A{\otimes}A$, which is convolution invertible, such that ${\sigma}_{21}({\Delta}{\otimes}id){\tau}({\sigma}(1))={\sigma}_{32}(id{\otimes}{\Delta}){\tau}({\sigma}(1))$. We define Hopf algebras, ($A_{\sigma}$, m, u, ${\Delta}_{\sigma}$, ${\varepsilon}$, $S_{\sigma}$). If B and C are opposite skew copaired Hopf algebras and $A=B{\otimes}_kC$ then we find Hopf algebras, ($A_{[{\sigma}]}$, $m_B{\otimes}m_C$, $u_B{\otimes}u_C$, ${\Delta}_{[{\sigma}]}$, ${\varepsilon}B{\otimes}{\varepsilon}_C$, $S_{[{\sigma}]}$). Let H be a finite dimensional commutative Hopf algebra with dual basis $\{h_i\}$ and $\{h_i^*\}$, and let $A=H^{op}{\otimes}H^*$. We show that if we define ${\sigma}:k{\rightarrow}H^{op}{\otimes}H^*$ by ${\sigma}(1)={\sum}h_i{\otimes}h_i^*$ then ($A_{[{\sigma}]}$, $m_A$, $u_A$, ${\Delta}_{[{\sigma}]}$, ${\varepsilon}_A$, $S_{[{\sigma}]}$) is the dual space of Drinfeld double, $D(H)^*$, as Hopf algebra.

• International Journal of Ocean System Engineering
• /
• 제1권1호
• /
• pp.52-59
• /
• 2011

A prototype radar display unit was implemented using inexpensive off-the-shelf components, including a nonlinear estimation algorithm for the target tracking in a clutter environment. Two custom designed boards; an analog signal processing board and a DSP board, can be plugged into an expansion slot of a personal computer (PC) to form a maritime radar display unit. Our system provided all the functionality specified in the International Maritime Organization (IMO) resolution A422(XI). The analog signal processing board was used for A/D conversion as well as rain and sea clutter suppression. The main functions of the DSP board were scan conversion and video overlay operations. A host PC was used to run the tracking algorithm of targets in clutter, using the discrete-time Bayes optimal (nonlinear, and non-Gaussian) estimation method, and the graphic user interface (GUI) software for Automatic Radar Plotting Aid (ARPA). The proposed tracking method recursively found the entire probability density function of the target position and velocity by converting into linear convolution operations.