• 한국정보통신학회논문지
• /
• 제15권11호
• /
• pp.2465-2473
• /
• 2011

WiMAX, WLAN 등의 무선통신 시스템에 사용되는 LDPC(low density parity check) 복호기의 핵심 기능블록인 DFU(decoding function unit)의 회로 최적화를 제안한다. DFU를 2의 보수 연산 대신에 sign-magnitude 연산 기반으로 설계함으로써 수체계 변환과정을 제거하였으며, 모바일 WiMAX용 다중모드 LDPC 복호기에 사용되는 96개 DFU 배열의 게이트 수를 18% 감소시켰다. 제안된 DFU 구조를 적용하여 모바일 WiMAX 표준을 지원하는 다중모드 LDPC 복호기를 설계하였다. 설계된 LDPC 복호기는 0.18-${\mu}m$ CMOS 셀 라이브러리를 이용하여 50 MHz 클록주파수로 합성한 결과 268,870 게이트와 71,424 비트의 메모리로 구현되었으며, FPGA 구현을 통해 하드웨어 동작을 검증 하였다.

• 한국인터넷방송통신학회논문지
• /
• 제8권2호
• /
• pp.15-21
• /
• 2008

본 논문에서는 2진논리의 확장을 Galis체상에서 해석하여 확장논리에 기초한 순차디지털논리시스템과 컴퓨터구조의 핵심인 연산알고리즘을 논의하였다. 순차디지털논리시스템은 Building Block으로서 T-gate를 사용하였으며, 차순상태함수, 출력함수를 도출하여 최종 궤환이 없는 Moore Model의 순차디지털논리시스템을 구성하였다. 그리고, 컴퓨터구조에서 중요한 연산알고리즘의 핵심인 가산, 감산, 승산 및 제산 알고리즘을 유한체의 수학적 성질을 토대로 각각 도출하였다. 특히, 유한체 GF($P^m$)상에서 P=2인 경우는 기존의 2진디지털논리시스템에 적용이 용이하다는 장점이 있으며, mod2의 성질에 의해 감산 알고리즘은 가산 알고리즘과 동일하다. 제안한 방법은 기존의 2진논리를 확장할 수 있어 좀 더 효율적으로 디지털논리시스템을 구성할 수 있을 것으로 사료된다.

• Journal of Mechanical Science and Technology
• /
• 제16권7호
• /
• pp.959-965
• /
• 2002

Effective hydraulic conductivity of a statistically anisotropic heterogeneous medium is obtained for steady two-dimensional flows employing stochastic analysis. Flow equations are solved up to second order and the effective conductivity is obtained in a semi-analytic form depending only on the spatial correlation function and the anisotropy ratio of the hydraulic conductivity field, hence becoming a true intrinsic property independent of the flow field. Results are obtained using a statistically anisotropic Gaussian correlation function where the anisotropy is defined as the ratio of integral scales normal and parallel to the mean flow direction. Second order results indicate that the effective conductivity of an anisotropic medium is greater than that of an isotropic one when the anisotropy ratio is less than one and vice versa. It is also found that the effective conductivity has upper and lower bounds of the arithmetic and the harmonic mean conductivities.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 1989년도 한국자동제어학술회의논문집; Seoul, Korea; 27-28 Oct. 1989
• /
• pp.653-662
• /
• 1989

Cerebellar Model Arithmetic Controller (CMAC) has been introduced as an adaptive control function generator. CMAC computes control functions referring to a distributed memory table storing functional values rather than by solving equations analytically or numerically. CMAC has a unique mapping structure as a coarse coding and supervisory delta-rule learning property. In this paper, learning aspects and a convergence of the CMAC were investigated. The efficient training algorithms were developed to overcome the limitations caused by the conventional maximum error correction training and to eliminate the accumulated learning error caused by a sequential node training. A nonlinear function generator and a motion generator for a two d.o.f. manipulator were simulated. The efficiency of the various learning algorithms was demonstrated through the cpu time used and the convergence of the rms and maximum errors accumulated during a learning process. A generalization property and a learning effect due to the various gains were simulated. A uniform quantizing method was applied to cope with various ranges of input variables efficiently.

• 한국기계연구소 소보
• /
• 통권19호
• /
• pp.93-115
• /
• 1989

Cerebellar Model Arithmetic Controller(CMAC) has been introduced as an adaptive control function generator. CMAC computes control functions referring to a distributed memory table storing functional values rather than by solving equations analytically or numerically. CMAC has a unique mapping structure as a coarse coding and supervisory delta-rule learning property. In this paper, learning aspects and a convergence of the CMAC were investigated. The efficient training algorithms were developed to overcome the limitations caused by the conventional maximum error correction training and to eliminate the accumulated learning error caused by a sequential node training. A nonlinear function generator and a motion generator for a two d. o. f. manipulator were simulated. The efficiency of the various learning algorithms was demonstrated through the cpu time used and the convergence of the rms and maximum errors accumulated during a learning process; A generalization property and a learning effect due to the various gains were simulated. A uniform quantizing method was applied to cope with various ranges of input variables efficiently.

• 제어로봇시스템학회논문지
• /
• 제19권10호
• /
• pp.879-885
• /
• 2013

By performing interval analysis on the system transfer function, we propose an improved method of control loop design for a DC/DC converter. In conventional design methods, the effect of system parameter change due to the specified range of operating conditions and production tolerances in power components should be checked a posteriori, because this may result in a transfer function shift and performance degradation. In the proposed method, a possible parameter change is considered a priori in the design step in order that the desired crossover frequency and sufficient phase margin can be achieved even in the worst case condition. As an illustrative example, a buck dc/dc converter is designed by two different methods and performance comparisons are performed to verify the feasibility of the proposed scheme.

• 한국초전도저온공학회:학술대회논문집
• /
• 한국초전도저온공학회 2003년도 추계학술대회 논문집
• /
• pp.117-119
• /
• 2003

We have designed fabricated, and tested an RSFQ(Rapid Single Flux Quantum) 1-bit ALU (Arithmetic Logic Unit). The 1-bit ALU was composed of a half adder and three SFQ DC switches. Three DC switches were attached to the two output ports of an ALU for the selection of each function from the available functions that were AND, OR, XOR and ADD. And we also attached two DC switches at the input ports of the half adder so that the input data were controlled using the function generators operating at low speed while we tested the circuit at high speed. The test bandwidth was from 1KHz to 5 ㎓. The chip was tested at the liquid helium temperature of 4.2 K.

• 대한수학회지
• /
• 제56권6호
• /
• pp.1599-1611
• /
• 2019

Let $m,n{\in}{\mathbb{N}}$. We represent the additive subgroups of the ring ${\mathbb{Z}}_m{\times}{\mathbb{Z}}_n$, which are also (unital) subrings, and deduce explicit formulas for $N^{(s)}(m,n)$ and $N^{(us)}(m,n)$, denoting the number of subrings of the ring ${\mathbb{Z}}_m{\times}{\mathbb{Z}}_n$ and its unital subrings, respectively. We show that the functions $(m,n){\mapsto}N^{u,s}(m,n)$ and $(m,n){\mapsto}N^{(us)}(m,n)$ are multiplicative, viewed as functions of two variables, and their Dirichlet series can be expressed in terms of the Riemann zeta function. We also establish an asymptotic formula for the sum $\sum_{m,n{\leq}x}N^{(s)}(m,n)$, the error term of which is closely related to the Dirichlet divisor problem.

• 충청수학회지
• /
• 제37권2호
• /
• pp.57-66
• /
• 2024

In this paper, we study how the Hilbert polynomial, associated with a reduced closed subscheme X of codimension 2 in ℙ^N, reveals geometric information about X. Although it is known that the Hilbert polynomial can tell us about the scheme's degree and arithmetic genus, we find additional geometric information it can provide for smooth varieties of codimension 2. To do this, we introduce the concept of Gotzmann coefficients, which helps to extract more information from the Hilbert polynomial. These coefficients are based on the binomial expansion of values of the Hilbert function. Our method involves combining techniques from initial ideals and partial elimination ideals in a novel way. We show how these coefficients can determine the degree of certain geometric features, such as the singular locus appearing in a generic projection, for smooth varieties of codimension 2.

• 전자공학회논문지
• /
• 제54권1호
• /
• pp.41-45
• /
• 2017

본 논문에서는 조합논리스템의 효율적인 다중출력스위칭함수 구성의 한 가지 방법을 제안하였다. 제안한 방법의 시간영역기반의 멀티플렉싱을 기반으로 공통다중종단노드확장논리결정도를 도출하여 최종 조합논리시스템의 다중출력스위칭함수를 구하므로 기존의 시간영역기반의 멀티플렉싱에 비해 최적화된 입력변수의 쌍과 출력변수 쌍을 상당히 줄일 수 있으며, 또한 코스트 면에서도 유리하다. 또한, 입출력단자 수의 감소, 회로구성의 간략화, 연산속도의 향상 등의 이점이 있으며 기존의 방법에 비해 좀 더 정규성과 확장성이 용이하다.