• 한국통신학회논문지
• /
• 제19권9호
• /
• pp.1812-1823
• /
• 1994

본 논문에서는 DS/CDMA 방식을 채택하는 Q-CDMA 시스템 역방향링크 송수신 모뎀에서 기지국 수신기의 구조 해석과 성능분석에 초점을 두고 시뮬레이션을 수행하였다. 수신기 알고리즘은 열악한 이동통신 채널환경하에서 신뢰성있는 데이터전송을 위하여 많은 데이터를 고속으로 처리해야 하는 특징이 있다. Q-CDMA 방식의 적용에 따른 코드획득회로, 코드추적회로, 그리고 다경로에 따른 시간 다이버시티를 이용한 레이크 시스템을 이용한 복조회로를 연결시켜 동작하도록 하였다. 각 블럭별 회로의 모델에 따른 해석을 하고 AWAGN환경과 다경로 페이딩 환경하에서 임의의 한 사용자에 대한 비트에러 확률을 구한다.

• 전기학회논문지
• /
• 제58권7호
• /
• pp.1411-1417
• /
• 2009

In this paper, we consider the design of $H_{\infty}$ control of linear discrete-time systems having no mathematical model. The basic approach is to use Q-learning which is a reinforcement learning method based on actor-critic structure. The model-free control design is to use not the mathematical model of the system but the informations on states and inputs. As a result, the derived iterative algorithm is expressed as linear matrix inequalities(LMI) of measured data from system states and inputs. It is shown that, for a sufficiently rich enough disturbance, this algorithm converges to the standard $H_{\infty}$ control solution obtained using the exact system model. A simple numerical example is given to show the usefulness of our result on practical application.

• Journal of Communications and Networks
• /
• 제5권2호
• /
• pp.174-183
• /
• 2003

Space-Time (ST) codes are known to provide high transmission rates, diversity and coding gains. In this paper, a tight upper bound on the error probability of ST codes over rapid fading channels is presented. Moreover, ST codes suitable for rapid fading channels are presented. These codes are designed using the QPSK and 16-QAM signal constellations. The proposed codes are based on two different encoding schemes. The first scheme uses a single trellis encoder, whereas the second scheme uses the I-Q encoding technique. Code design is achieved via partitioning the signal space such that the design criteria are maximized. As a solution for the decoding problem of I-Q ST codes, the paper introduces a low-complexity decoding algorithm. Results show that the I-Q ST codes using the proposed decoding algorithm outperform singleencoder ST codes with equal complexity. The proposed codes are tested over fading channels with different interleaving conditions, where it is shown that the new codes are robust under such imperfect interleaving conditions.

• International Journal of Control, Automation, and Systems
• /
• 제5권5호
• /
• pp.570-576
• /
• 2007

This paper presents a hexagon-based Q-leaning algorithm to find a hidden targer object with multiple robots. An experimental environment was designed with five small mobile robots, obstacles, and a target object. Robots went in search of a target object while navigating in a hallway where obstacles were strategically placed. This experiment employed two control algorithms: an area-based action making (ABAM) process to determine the next action of the robots and hexagon-based Q-learning to enhance the area-based action making process.

• 대한수학회논문집
• /
• 제25권4호
• /
• pp.655-669
• /
• 2010

In this paper we propose new primal-dual interior point methods (IPMs) for $P_*(\kappa)$ linear complementarity problems (LCPs) and analyze the iteration complexity of the algorithm. New search directions and proximity measures are defined based on a class of kernel functions, $\psi(t)=\frac{t^2-1}{2}-{\int}^t_1e{^{q(\frac{1}{\xi}-1)}d{\xi}$, $q\;{\geq}\;1$. If a strictly feasible starting point is available and the parameter $q\;=\;\log\;\(1+a{\sqrt{\frac{2{\tau}+2{\sqrt{2n{\tau}}+{\theta}n}}{1-{\theta}}\)$, where $a\;=\;1\;+\;\frac{1}{\sqrt{1+2{\kappa}}}$, then new large-update primal-dual interior point algorithms have $O((1\;+\;2{\kappa})\sqrt{n}log\;n\;log\;{\frac{n}{\varepsilon}})$ iteration complexity which is the best known result for this method. For small-update methods, we have $O((1\;+\;2{\kappa})q{\sqrt{qn}}log\;{\frac{n}{\varepsilon}})$ iteration complexity.

• Journal of Power Electronics
• /
• 제11권6호
• /
• pp.922-930
• /
• 2011

In this paper, two radial basis function neural networks (RBFNNs) are used to dynamically identify harmonics content in converter waveforms based on the p-q (real power-imaginary power) theory. The converter waveforms are analyzed and the types of harmonic content are identified over a wide operating range. Constant power and sinusoidal current compensation strategies are investigated in this paper. The RBFNN filtering training algorithm is based on a systematic and computationally efficient training method called the hybrid learning method. In this new methodology, the RBFNN is combined with the p-q theory to extract the harmonics content in converter waveforms. The small size and the robustness of the resulting network models reflect the effectiveness of the algorithm. The analysis is verified using MATLAB simulations.

• 전력전자학회:학술대회논문집
• /
• 전력전자학회 2007년도 하계학술대회 논문집
• /
• pp.328-330
• /
• 2007

This paper deals with the estimation method on the initial pole position of a z-axis permanent magnet linear synchronous motor(PMLSM) without magnetic pole sensors such as Hall sensors. The proposed method takes account of the z-axis conditions such as the gravitational force and also the load conditions. The algorithm consists of two steps. The first step is to estimate the initial q-axis approximately by monitoring the movements at predefined different test q-axes. The second step is to estimate the real q-axis as accurately as possible based on the results at three different test q-axes. Experimental results on the z-axis PMLSM show good estimation characteristics of the proposed method.

• 한국경영과학회지
• /
• 제22권1호
• /
• pp.75-85
• /
• 1997

A stable algorithm for computing the fundamental matrix (I-Q)$^{-1}$ of a Markov chain is proposed, where Q is a substochastic matrix. The proposed algorithm utilizes the GTH algorithm (Grassmann, Taskar and Heyman, 1985) which is turned out to be stable for finding the steady state distribution of a finite Markov chain. Our algorithm involves no subtractions and therefore loss of significant digits due to concellation is ruled out completely while Gaussian elimination involves subtractions and thus may lead to loss of accuracy due to cancellation. We present numerical evidence to show that our algorithm achieves higher accuracy than the ordinagy Gaussian elimination.

• 대한용접접합학회:학술대회논문집
• /
• 대한용접접합학회 2004년도 추계학술발표대회 개요집
• /
• pp.234-236
• /
• 2004

• 전기전자학회논문지
• /
• 제24권4호
• /
• pp.991-997
• /
• 2020

강화학습은 주로 순차적인 의사 결정 문제에 적용되어 왔다. 특히 최근에는 신경망과 결합한 형태로 기존에는 해결하지 못한 분야에서도 성공적인 결과를 내고 있다. 하지만 신경망을 이용하는 강화학습은 현장에서 즉각적으로 사용하기엔 너무 복잡하다는 단점이 있다. 본 논문에서는 학습이 쉬운 강화학습 알고리즘 중 하나인 Q-learning을 이용하여 이동 로봇의 경로를 생성하는 알고리즘을 구현하였다. Q-table을 미리 만드는 방식의 Q-learning은 명확한 한계를 가지기 때문에 실시간으로 Q-table을 업데이트하는 실시간 Q-learning을 사용하였다. 탐험 전략을 조정하여 실시간 Q-learning에 필요한 학습 속도를 얻을 수 있었다. 마지막으로 실시간 Q-learning과 DQN의 성능을 비교하였다.