<?xml version="1.0" encoding="UTF-8"?>
<?xml-stylesheet type="text/xsl" href="/resources/xsl/jats-html.xsl"?>
<article article-type="research-article" dtd-version="1.1" xml:lang="ko" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
<front>
	<journal-meta>
		<journal-id journal-id-type="publisher-id">jkits</journal-id>
		<journal-title-group>
		<journal-title>한국지식정보기술학회 논문지</journal-title>
		<journal-title xml:lang="en">Journal of Knowledge Information Technology and Systems</journal-title>
		</journal-title-group>
		<issn pub-type="ppub">1975-7700</issn>
		<publisher>
		<publisher-name>한국지식정보기술학회</publisher-name>
		<publisher-name xml:lang="en">Korea Knowledge Information Technology Society</publisher-name>
		</publisher>
	</journal-meta>
	<article-meta>
		<article-id pub-id-type="publisher-id">jkits_2019_14_04_359</article-id>
		<article-id pub-id-type="doi">10.34163/jkits.2019.14.4.006</article-id>
		<article-categories>
			<subj-group>
				<subject>Research Article</subject>
			</subj-group>
		</article-categories>
		<title-group>
			<article-title>딥러닝을 이용한 시스템식별에 관한 연구</article-title>
			<trans-title-group xml:lang="en">
				<trans-title>A Study on System Identification Using Deep Learning</trans-title>
			</trans-title-group>
		</title-group>
		<contrib-group>
			<contrib contrib-type="author" xlink:type="simple">
				<name-alternatives>
					<name name-style="eastern">
						<surname>정</surname><given-names>홍근</given-names>
					</name>
					<name name-style="western" xml:lang="en">
						<surname>Joung</surname><given-names>Houng-Kun</given-names>
					</name>
					</name-alternatives>
					<xref ref-type="aff" rid="A1"><sup>1</sup></xref>
			</contrib>
			<contrib contrib-type="author" xlink:type="simple">
				<name-alternatives>
					<name name-style="eastern">
						<surname>오</surname><given-names>원근</given-names>						
					</name>
					<name name-style="western" xml:lang="en">
						<surname>Oh</surname><given-names>Won-Guen</given-names>
					</name>
					</name-alternatives>
					<xref ref-type="aff" rid="A2"><sup>2</sup></xref>
					<xref ref-type="fn" rid="fn01"><sup>*</sup></xref>
			</contrib>
			</contrib-group>
		<aff-alternatives id="A1">
			<aff><sup>1</sup><italic>청주대학교 융합전자공학부 조교수</italic></aff>
			<aff xml:lang="en"><italic>Division of Converged Electronic Engineering, Cheongju University,</italic></aff>
		</aff-alternatives>
		<aff-alternatives id="A2">
			<aff><sup>2</sup><italic>순천대학교 멀티미디어공학전공 교수</italic></aff>
			<aff xml:lang="en"><italic>Department of Multimedia Engineering, Sunchon National University</italic></aff>
		</aff-alternatives>
				<author-notes>
			<fn id="fn01"><label>*</label><p>Corresponding author is with the Department of Multimedia Engineering, Sunchon National University, 255 Jungang-ro, Suncheon, Jellanam-do, 57922, KOREA. </p>
			<p><italic>E-mail address</italic>: <email>owg@scnu.ac.kr</email></p></fn>
		</author-notes>
		<pub-date pub-type="ppub">
			<month>8</month>
			<year>2019</year>
		</pub-date>
		<volume>14</volume>
		<issue>4</issue>
		<fpage>359</fpage>
		<lpage>368</lpage>
		<history>
			<date date-type="received">
				<day>10</day>
				<month>6</month>
				<year>2019</year>
			</date>
			<date date-type="rev-recd">
				<day>26</day>
				<month>6</month>
				<year>2019</year>
			</date>
			<date date-type="accepted">
				<day>9</day>
				<month>8</month>
				<year>2019</year>
			</date>
		</history>
		<permissions>
			<copyright-statement>&#x00A9; 2019 KKITS All rights reserved</copyright-statement>
			<copyright-year>2019</copyright-year>
		</permissions>
		<abstract>
			<title>요약</title>
			<p>본 논문에서는 시간지연이 존재하는 시스템을 위한 제어기기 동조 문제에서 딥러닝을 이용한 시스템 식별에 관한 연구이다. 시스템 식별을 위한 제어기기 동조에 관현 연구 중 Yunwana와 Seborg(1982)가 제시한 제어기의 동조방법은 Pade’ 근사에 의한 위상 오차때문에 시간 지연이 없거나 작은 시간 지연에 관하여는 제어가 잘되는 장점이 있으나 큰 시간 지연이 존해 할수록 시간 지연이 크게 추정되고 시스템에서는 적용할 수 없는 단점이 있다. 그리고 산업현장에서 많이 쓰이고 있는 Zigler-Nichols가 제안한 시행착오 방법은 제어기기 동조를 위하여 많은 시간이 소비되는 단점이 있다. Cohen-Coon이 제안한 프로세스 반응곡선을 이용한 제어기기 동조방법은 Zigler-Nichols가 제안한 방법보다는 제어기기 동조에 걸리는 시간이 단축되는 장점이 있지만 개루프 시스템에서만 적용할 수 있고 폐루프 시스템에서는 적용할 수 없는 단점이 있다.이러한 단점들을 보완하기 위하여Suh가 제안한 방법은 폐루프 시스템에서도 적용할 수 있는 장점이 있고 또한 시간지연 항을 Pade’근사로 변환시 발생하는 위상오차에 대하여 Pade’근사에서 조절인자의 설정으로 위상오차를 줄여 제어기기의 최적 동조 방법에 대하여 제시하였다. 그러나 이러한 방법은 시간 지연항의 상수 값과 비례적으로 조절인자를 대입한 방법이기 때문에 해석적이지 못한 단점이 있다. 본 논문에서는 위상 오차에 대한 이론적 분석을 통해 이와 같은 원인이 Pade’근사에 의한 위상오차와 시간 지연 항이 크게 추정되는 문제점을 해결 할 수 있는 해석적인 방법으로 딥 러닝 알고리즘 중에 깊은 믿음 네트워크 알고리즘을 이용한 조절인자의 설정으로 위상 오차를 최적으로 줄일 수 있는 새로운 최적 동조방법을 기존의 본 연구자 Suh(1984)의 연구결과를 바탕으로 제시하였다. 그리고 시물레이션을 통하여 Zielger-Nichols의 시행착오 방법과 Yunwana-Seborg가 제시한 제어기의 동조방법을 비교하여 본 논문에서 제시한 방법의 타당성을 입증하였다.</p>
		</abstract>
		<trans-abstract xml:lang="en">
			<title>ABSTRACT</title>
			<p>This paper deals with a study on a system identification using deep learning in the case of a controller tuning for the system where a time delay exists. Of studies on the controller tuning for the system identification, the controller tuning method suggested by Yunwana and Seborg(1982) has an advantage of taking a good control over either none or small time delays due to phase error by Pade' approximation, whereas it comes with a disadvantage of having a greater estimated of time delay over the presence of a large time delay and of being unable to be used in a system. Furthermore, the trial-and-error method suggested by Zigler-Nichols and commonly used in industrial fields shows a disadvantage which is time consuming for a controller tuning. The controller tuning using a process response curve suggested by Cohen-Coon has a benefit of cutting more time taken for a controller tuning than the method by Zigler-Nichols does. It also faces a limitation of being applicable only to the open loop system but not applicable to the close loop system. To make up for these disadvantages, the Suh-suggested method, as its benefit, is applicable even to the close loop system. On top of this, it proposed a controller's optimal tuning method by reducing phase error through setting up control factors in the Pade' approximation with respect to the phase error generated in converting time delay into Pade' approximation. This method, however, involves putting control factors in proportion to time-delay constant values, which is therefore - as a disadvantage - not analytical. This paper went through a theoretical analysis on phase error as an analytical method to solve an issue involving the large estimation of phase error by Pade' approximation and time delay with the use of deep learning. Presented based on the findings of this existing researcher Suh (1984) was a new optimal tuning method dedicated to reducing phase error to a optimal level by setting control factors using the deep belief network algorithm out of deep learning algorithms. Besides, a related simulation was performed to compare the trial-and-error method by Zielger-Nichols and the tuning method for a controller suggested by Yunwana-Seborg, and the validity of the methods suggested in this paper was verified, accordingly.
</p>
		</trans-abstract>
		<kwd-group kwd-group-type="author" xml:lang="en">
			<kwd>Deep learing</kwd>
			<kwd>Deep belief network</kwd>
			<kwd>PID control</kwd>
			<kwd>Time delay</kwd>
			<kwd>Pade’ Aproximaion</kwd>
			<kwd>System identification</kwd>
			<kwd>Controller tuning</kwd>
		</kwd-group>
	</article-meta>
</front>
<body>
<sec id="sec001" sec-type="intro">
	<title>1. 서 론</title>
<p>일반적으로 제어기는 선택되어지는 입력들에 의해서 시스템의 응답이 요구되는 기준에 따르도록 하는데 목적이 있고 제어기의 동조방법은 알고리즘 변수들을 최적화 시키는 것이라고 할 수 있다. 이에 대한 본격적인 연구는 1941년 Ziegler와 Nichols<xref ref-type="bibr" rid="B001">[1]</xref>에 의해 이루어 졌으며 루프 동조방법으로 불리고 산업현장에서 많이 쓰이고 있는 방법중에 하나이다. 그러나 Zigler와 Nichols가 제안한 방법은 많은 최종 이득과 최종 주기를 얻기 위해 많은 시행착오를 거치기 때문에 시간이 소비되는 단점이 존재한다. 이러한 시행착오 방법을 보완하기 위한 문제점을 해결하기 위하여 Cohen과 Coon(1953)에 의하여 제안된 프로세스 반응 곡선을 이용한 방법에 제시되었다<xref ref-type="bibr" rid="B002">[2]</xref>. 그러나 이 방법은 Ziegler와 Nichols가 제안한 방법보다는 편리한 방법이지만 추정 모델의 함수를 구하기 위하여 프로세서 반응곡선의 정확한 기울기를 결정해야 하고 개루프 시스템에서만 적용할 수 있는 단점이 있다. 이러한 단점을 보완하기 위하여 Yunwana와 Seborg(1982)<xref ref-type="bibr" rid="B003">[3]</xref>에 의해 해석적인 방법이 제시되었고 이 방법은 폐루프 시스템에서 비례 파라미터만 입력하여 응답곡선을 얻고 이로부터 추정된 모델의1차 전달함수의 파라미터(시간 지연 항, 시간 상수, 안정상태 이득)를 결정하여 관계식을 유도하여 PID 제어기 동조 파라미터들의 값을 최적화하는 방법이다. 그러나 위의 방법은 식에 오류가 있었으며,이의 해결 방안이 Suh(1984)[<xref ref-type="bibr" rid="B004">4</xref>,<xref ref-type="bibr" rid="B005">5</xref>]에의해 제시되었다. 이 방법에서는 시간 지연항과 pade’ 근사에 대한 오차를 줄이기 위하여 조절인자를 도입하여 PID제어기 동조 파라미터들의 값을 최적화하는 기법이 사용되었다. 그러나 이러한 방법은 해석적이지 못하고 시간 지연항의 상수와 비례적으로 조절인자를 설정하여 안정성과 빠른 응답을 보장하였지만 해석적으로 증명하지 못한 단점이 있다.</p>
<p>해석적인 방법을 이용한 시스템 식별에 관한 PID 제어기 동조 연구는 릴레이 피드백을 이용한 자동 동조 방법이 있다[<xref ref-type="bibr" rid="B006">6</xref>-<xref ref-type="bibr" rid="B008">8</xref>]. 그러나 릴레이 피드백에 의한 자동 동조 방법은 제한된 모델에만 적용할 수 있고 광범위한 시스템 식별 문제에서는 적용하지 못하는 단점이 있다.</p>
<p>본 논문에서는 Pade’ 근사로 발생하는 오차를 줄이기 위하여 딥 러닝 알고리즘 중에서 깊은 믿음 네트워크(Deep belief network : DBN) 알고리즘<xref ref-type="bibr" rid="B009">[9]</xref><xref ref-type="bibr" rid="B011">[11]</xref>을 이용하여 Suh(1984)<xref ref-type="bibr" rid="B003">[3]</xref>방법에서 제안한 조절인자의 설정을 해석적인 방법으로 위상 오차를 효과적으로 줄이는 방법을 제시한다.</p>
<p>본 논문의 구성은 다음과 같다. 2장에서는 Pade’ 근사와 조절인자와 딥 러닝을 이용한 조절인자의 설정에 대하여 소개하고 3장에서는 조절인자를 고려한 제어기 동조에 대하여 설명하며 4장에서는 본 논문에서 제안하는 방법으로 모의실험을 진행하고 결과를 보여주며 5장에서 결론을 보여 논문을 마무리한다.</p>
</sec>
<sec id="sec002">
<title>2. Pade’ 근사와 조절인자</title>
<sec id="sec002-1">
<title>2.1 Pade’ 근사에 의한 위상오차</title>
<p>Yunwana와 Seborg(1982)가 제시한 제어기의 동조 방법에서 시간 지연 항 <mml:math id="m001-1"><mml:msup><mml:mi>e</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:msub><mml:mi>d</mml:mi><mml:msup><mml:mi>m</mml:mi><mml:mi>s</mml:mi></mml:msup></mml:msub></mml:mrow></mml:msup><mml:mspace linebreak="newline"/><mml:mspace linebreak="newline"/><mml:mspace linebreak="newline"/></mml:math>을 식 (<xref ref-type="disp-formula" rid="dm01">1</xref>)과 같이 Pade’ 근사 방법을 이용하여 제어기 동조에 사용하였다.</p>
<disp-formula-group>
	<disp-formula id="dm01">
		<label>(1)</label>
<mml:math id="dm01-1"><mml:msup><mml:mi>e</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:msub><mml:mi>d</mml:mi><mml:msup><mml:mi>m</mml:mi><mml:mi>s</mml:mi></mml:msup></mml:msub></mml:mrow></mml:msup><mml:mo>&#x2245;</mml:mo><mml:mfrac><mml:mrow><mml:mn>1</mml:mn><mml:mo>-</mml:mo><mml:mn>0</mml:mn><mml:mo>.</mml:mo><mml:mn>5</mml:mn><mml:msub><mml:mi>d</mml:mi><mml:msup><mml:mi>m</mml:mi><mml:mi>s</mml:mi></mml:msup></mml:msub></mml:mrow><mml:mrow><mml:mn>1</mml:mn><mml:mo>+</mml:mo><mml:mn>0</mml:mn><mml:mo>.</mml:mo><mml:mn>5</mml:mn><mml:msub><mml:mi>d</mml:mi><mml:msup><mml:mi>m</mml:mi><mml:mi>s</mml:mi></mml:msup></mml:msub></mml:mrow></mml:mfrac><mml:mspace linebreak="newline"/><mml:mspace linebreak="newline"/><mml:mspace linebreak="newline"/></mml:math>
	</disp-formula>
</disp-formula-group>
<p>시간 지연 항 <mml:math id="m001-2"><mml:msup><mml:mi>e</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:msub><mml:mi>d</mml:mi><mml:msup><mml:mi>m</mml:mi><mml:mi>s</mml:mi></mml:msup></mml:msub></mml:mrow></mml:msup><mml:mspace linebreak="newline"/><mml:mspace linebreak="newline"/><mml:mspace linebreak="newline"/></mml:math>과 Pade’ 근사 방법을 비교 하면 이득(Gain)에 대한 오차는 발생하지는 않지만 위상 의 경우에는 시간 지연이 (Phase) 커질수록 큰 오차가 발생한다. 시간 지연이 커진다는 의미는 시간 지연 항의 <mml:math id="m001-3"><mml:msup><mml:mi>e</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:msub><mml:mi>d</mml:mi><mml:msup><mml:mi>m</mml:mi><mml:mi>s</mml:mi></mml:msup></mml:msub></mml:mrow></mml:msup><mml:mspace linebreak="newline"/><mml:mspace linebreak="newline"/><mml:mspace linebreak="newline"/></mml:math>에서 <italic>d</italic><sub><italic>m</italic></sub>의 값이 커진다는 것을 의미하고 이러한 Pade’ 근사에 의하여 발생하는 오차를 &lt;<xref ref-type="fig" rid="f001">그림 1</xref>&gt;에 나타내었다. 이러한 위상오차는 시간 지연이 증가함에 따라 발생하는 응답 데이터로부터 유도된 관계식에 의하여 추정된 시간 지연 <italic>d</italic><sub><italic>m</italic></sub>이 실제 제어 대상의 시간 지연 항보다 상당히 크게 추정되는 결과를 보여주고 따라서 Pade’ 근사 방법을 시간 지연항이 비교적으로 큰 시스템에서는 적용하지 못하는 이유가 된다.</p>
<fig id="f001" orientation="portrait" position="float">
	<label>그림 1.</label>
	<caption>
		<title> <mml:math id="dm0001"><mml:msup><mml:mi>e</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:msub><mml:mi>d</mml:mi><mml:msup><mml:mi>m</mml:mi><mml:mi>s</mml:mi></mml:msup></mml:msub></mml:mrow></mml:msup><mml:mspace linebreak="newline"/><mml:mspace linebreak="newline"/><mml:mspace linebreak="newline"/></mml:math>의 Pade’ 근사에 의한 주파수 응답</title>
		<p>Figure 1. The Frequency Response of Pade’ Approximation to  <mml:math id="m001-4"><mml:msup><mml:mi>e</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:msub><mml:mi>d</mml:mi><mml:msup><mml:mi>m</mml:mi><mml:mi>s</mml:mi></mml:msup></mml:msub></mml:mrow></mml:msup><mml:mspace linebreak="newline"/><mml:mspace linebreak="newline"/><mml:mspace linebreak="newline"/></mml:math></p>
	</caption>
	<graphic xlink:href="../ingestImageView?artiId=ART002496983&amp;imageName=jkits_2019_14_04_359_f001.jpg" position="float" orientation="portrait" xlink:type="simple"></graphic>
</fig>
</sec>
<sec id="sec002-2">
<title>2.2 Suh에 의한 조절인자의 설정</title>
<p>&lt;<xref ref-type="fig" rid="f001">그림 1</xref>&gt;에서 Pade’근사에 의한 위상 오차를 줄이기 위하여 Suh는 조절인자 <italic>α</italic>를 고려하여 새로운 근사방법을 식 (<xref ref-type="disp-formula" rid="dm02">2</xref>)와 같이 제안하였다.</p>
<disp-formula-group>
	<disp-formula id="dm02">
		<label>(2)</label>
<mml:math id="dm02-1"><mml:msup><mml:mi>e</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:msub><mml:mi>d</mml:mi><mml:msup><mml:mi>m</mml:mi><mml:mi>s</mml:mi></mml:msup></mml:msub></mml:mrow></mml:msup><mml:mo>&#x2245;</mml:mo><mml:mfrac><mml:mrow><mml:mn>1</mml:mn><mml:mo>-</mml:mo><mml:mn>0</mml:mn><mml:mo>.</mml:mo><mml:mn>5</mml:mn><mml:mi>&#x3B1;</mml:mi><mml:msub><mml:mi>d</mml:mi><mml:msup><mml:mi>m</mml:mi><mml:mi>s</mml:mi></mml:msup></mml:msub></mml:mrow><mml:mrow><mml:mn>1</mml:mn><mml:mo>+</mml:mo><mml:mn>0</mml:mn><mml:mo>.</mml:mo><mml:mn>5</mml:mn><mml:mi>&#x3B1;</mml:mi><mml:msub><mml:mi>d</mml:mi><mml:msup><mml:mi>m</mml:mi><mml:mi>s</mml:mi></mml:msup></mml:msub></mml:mrow></mml:mfrac><mml:mspace linebreak="newline"/><mml:mspace linebreak="newline"/><mml:mspace linebreak="newline"/></mml:math>
	</disp-formula>
</disp-formula-group>
<p>Suh 방법에 의한 조절인자 <italic>α</italic>의 설정방법은 시간 지연항 <mml:math id="m001-5"><mml:msup><mml:mi>e</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:msub><mml:mi>d</mml:mi><mml:msup><mml:mi>m</mml:mi><mml:mi>s</mml:mi></mml:msup></mml:msub></mml:mrow></mml:msup><mml:mspace linebreak="newline"/><mml:mspace linebreak="newline"/><mml:mspace linebreak="newline"/></mml:math>에서 <italic>d</italic><sub><italic>m</italic></sub>의 값과 동일하게 설정하여 Yunwana와 Seborg가 제안한 해석적인 방법에서의 잘못된 유도식을 보완하였으나 비례적인 방법으로 조절인자 값을 설정하였기 때문에 시간 지연항 <mml:math id="dm002"><mml:msup><mml:mi>d</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:msub><mml:mi>d</mml:mi><mml:msup><mml:mi>m</mml:mi><mml:mi>s</mml:mi></mml:msup></mml:msub></mml:mrow></mml:msup><mml:mspace linebreak="newline"/><mml:mspace linebreak="newline"/><mml:mspace linebreak="newline"/></mml:math>가 작은 시스템부터 큰 시스템에서는 해석적이지 못한 단점이 있다.</p>
</sec>
<sec id="sec002-3">
<title>2.3 딥 러닝을 이용한 조절인자의 설정</title>
<p>최근 많은 연구가 이루어지고 있는 딥 러닝은 다층 신경망보다 더 많은 은닉층과 뉴런을 가진 아키텍처를 가지고 있으며, 영상이나 음향, 그리고 데이터 분석 등에 활발히 적용되고 있다.</p>
<p>Pade’ 근사에서 발생하는 오차를 줄이기 위하여 딥 러닝의 알고리즘 중 DBN 알고리즘을 사용하였다.[<xref ref-type="bibr" rid="B016">16</xref>,<xref ref-type="bibr" rid="B020">20</xref>] 전체적인 딥 러닝 알고리즘을 &lt;<xref ref-type="fig" rid="f002">그림 2</xref>&gt;에 나타내었다.</p>
<fig id="f002" orientation="portrait" position="float">
	<label>그림 2.</label>
	<caption>
		<title>딥 러닝 과정</title>
		<p>Figure 2 Process of deep learning</p>
	</caption>
	<graphic xlink:href="../ingestImageView?artiId=ART002496983&amp;imageName=jkits_2019_14_04_359_f002.jpg" position="float" orientation="portrait" xlink:type="simple"></graphic>
</fig>
</sec>
</sec>
<sec id="sec003">
<title>3. 조절인자를 고려한 제어기 동조</title>
<fig id="f003" orientation="portrait" position="float">
	<label>그림 2.</label>
	<caption>
		<title>폐루프시스템의 블록다이어그램</title>
		<p>Fig. 2. Block Diagram of a Feedback Control System</p>
	</caption>
	<graphic xlink:href="../ingestImageView?artiId=ART002496983&amp;imageName=jkits_2019_14_04_359_f003.jpg" position="float" orientation="portrait" xlink:type="simple"></graphic>
	</fig>
<p>&lt;<xref ref-type="fig" rid="f002">그림 2</xref>&gt;에서 볼 수 있듯이 시스템에서 프로세서 전달함수 <italic>G</italic><sub><italic>p</italic></sub>(<italic>s</italic>)와 부하 전달함수는 모른다고 가정하고,프로세서의 추정된 모델은 식(<xref ref-type="disp-formula" rid="dm03">3</xref>)과 같이 가정한다.</p>
<disp-formula-group>
	<disp-formula id="dm03">
		<label>(3)</label>
<mml:math id="dm03-1"><mml:msub><mml:mi>G</mml:mi><mml:mi>m</mml:mi></mml:msub><mml:mfenced><mml:mi>s</mml:mi></mml:mfenced><mml:mo>=</mml:mo><mml:mfrac><mml:mrow><mml:msub><mml:mi>k</mml:mi><mml:mi>m</mml:mi></mml:msub><mml:msup><mml:mi>e</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:msub><mml:mi>d</mml:mi><mml:mi>m</mml:mi></mml:msub><mml:mi>s</mml:mi></mml:mrow></mml:msup></mml:mrow><mml:mrow><mml:msub><mml:mi>&#x3C4;</mml:mi><mml:msup><mml:mi>m</mml:mi><mml:mi>s</mml:mi></mml:msup></mml:msub><mml:mo>+</mml:mo><mml:mn>1</mml:mn></mml:mrow></mml:mfrac><mml:mspace linebreak="newline"/><mml:mspace linebreak="newline"/><mml:mspace linebreak="newline"/></mml:math>
	</disp-formula>
</disp-formula-group>
<p><italic>k</italic><sub><italic>m</italic></sub> : 안정 상태 이득, <italic>&#x3C4;</italic><sub><italic>m</italic></sub> : 시간상수</p>
<p>&lt;<xref ref-type="fig" rid="f002">그림 2</xref>&gt;에서 제어기 전달함수 <italic>G</italic><sub><italic>c</italic></sub>(<italic>s</italic>)를 비례모드 <italic>k</italic><sub><italic>c</italic></sub>로 가정하면 폐루프 전달함수는 식 (<xref ref-type="disp-formula" rid="dm04">4</xref>)와 같이 표현할 수 있다.</p>
<disp-formula-group>
	<disp-formula id="dm04">
		<label>(4)</label>
<mml:math id="dm04-1"><mml:mfrac><mml:mrow><mml:mi>C</mml:mi><mml:mfenced><mml:mi>s</mml:mi></mml:mfenced></mml:mrow><mml:mrow><mml:mi>R</mml:mi><mml:mfenced><mml:mi>s</mml:mi></mml:mfenced></mml:mrow></mml:mfrac><mml:mo>=</mml:mo><mml:mfrac><mml:mrow><mml:mi>K</mml:mi><mml:msup><mml:mi>e</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:msub><mml:mi>d</mml:mi><mml:msup><mml:mi>m</mml:mi><mml:mi>s</mml:mi></mml:msup></mml:msub></mml:mrow></mml:msup></mml:mrow><mml:mrow><mml:mn>1</mml:mn><mml:mo>+</mml:mo><mml:msub><mml:mi>&#x3C4;</mml:mi><mml:msup><mml:mi>m</mml:mi><mml:mi>s</mml:mi></mml:msup></mml:msub><mml:mo>+</mml:mo><mml:mi>K</mml:mi><mml:msup><mml:mi>e</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:msub><mml:mi>d</mml:mi><mml:msup><mml:mi>m</mml:mi><mml:mi>s</mml:mi></mml:msup></mml:msub></mml:mrow></mml:msup></mml:mrow></mml:mfrac><mml:mspace linebreak="newline"/><mml:mspace linebreak="newline"/><mml:mspace linebreak="newline"/></mml:math>
	</disp-formula>
</disp-formula-group>
<p><italic>K</italic>=<italic>k</italic><sub><italic>c</italic></sub><italic>k</italic><sub><italic>m</italic></sub> : 개루프 이득</p>
<p>계단입력 변화에 대한 프로세서의 응답곡선은 &lt;<xref ref-type="fig" rid="f004">그림 3</xref>&gt;과 같다.</p>
<fig id="f004" orientation="portrait" position="float">
	<label>그림 3.</label>
	<caption>
		<title>계단입력 변화에 의한 프로세서의 응답곡선</title>
		<p>Figure 3. Response Curve of Process to a Step input Change</p>
	</caption>
	<graphic xlink:href="../ingestImageView?artiId=ART002496983&amp;imageName=jkits_2019_14_04_359_f004.jpg" position="float" orientation="portrait" xlink:type="simple"></graphic>
	</fig>
<p>&lt;<xref ref-type="fig" rid="f004">그림 3</xref>&gt;에서 표현되는 응답곡선에서 추정된 모델의 변수를 결정하기 위해서 식 (<xref ref-type="disp-formula" rid="dm02">2</xref>)의 조절인자가 고려된 Pade ‘근사 방법을 사용하여 식 (<xref ref-type="disp-formula" rid="dm05">4</xref>)에 대입하면 식 (<xref ref-type="disp-formula" rid="dm06">5</xref>)와 같이 표현할 수 있다.</p>
<disp-formula-group>
	<disp-formula id="dm05">
		<label>(4)</label>
<mml:math id="dm05-1"><mml:mfrac><mml:mrow><mml:mi>C</mml:mi><mml:mfenced><mml:mi>s</mml:mi></mml:mfenced></mml:mrow><mml:mrow><mml:mi>R</mml:mi><mml:mfenced><mml:mi>s</mml:mi></mml:mfenced></mml:mrow></mml:mfrac><mml:mo>=</mml:mo><mml:mfrac><mml:mrow><mml:mi>K</mml:mi><mml:mo>'</mml:mo><mml:mfenced><mml:mrow><mml:mn>1</mml:mn><mml:mo>+</mml:mo><mml:mn>0</mml:mn><mml:mo>.</mml:mo><mml:mn>5</mml:mn><mml:mi>&#x3B1;</mml:mi><mml:msub><mml:mi>d</mml:mi><mml:msup><mml:mi>m</mml:mi><mml:mi>s</mml:mi></mml:msup></mml:msub></mml:mrow></mml:mfenced><mml:msup><mml:mi>e</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mi>d</mml:mi><mml:mi>m</mml:mi><mml:mi>s</mml:mi></mml:mrow></mml:msup></mml:mrow><mml:mrow><mml:msup><mml:mi>&#x3C4;</mml:mi><mml:mn>2</mml:mn></mml:msup><mml:msup><mml:mi>s</mml:mi><mml:mn>2</mml:mn></mml:msup><mml:mo>+</mml:mo><mml:mn>2</mml:mn><mml:mi>&#x3B6;</mml:mi><mml:mi>&#x3C4;</mml:mi><mml:mi>s</mml:mi><mml:mo>+</mml:mo><mml:mn>1</mml:mn></mml:mrow></mml:mfrac><mml:mspace linebreak="newline"/><mml:mspace linebreak="newline"/><mml:mspace linebreak="newline"/></mml:math>
	</disp-formula>
	<disp-formula id="dm06">
<mml:math id="dm01-6"><mml:mi>K</mml:mi><mml:mo>'</mml:mo><mml:mo>=</mml:mo><mml:mfrac><mml:mi>K</mml:mi><mml:mrow><mml:mi>K</mml:mi><mml:mo>+</mml:mo><mml:mn>1</mml:mn></mml:mrow></mml:mfrac><mml:mo>,</mml:mo><mml:mo>&#xA0;</mml:mo><mml:mi>&#x3C4;</mml:mi><mml:mo>=</mml:mo><mml:msup><mml:mfenced open="[" close="]"><mml:mfrac><mml:mrow><mml:mi>&#x3B1;</mml:mi><mml:msub><mml:mi>d</mml:mi><mml:mi>m</mml:mi></mml:msub><mml:msub><mml:mi>&#x3C4;</mml:mi><mml:mi>m</mml:mi></mml:msub></mml:mrow><mml:mrow><mml:mn>2</mml:mn><mml:mi>K</mml:mi><mml:mo>+</mml:mo><mml:mn>1</mml:mn></mml:mrow></mml:mfrac></mml:mfenced><mml:mrow><mml:mn>2</mml:mn><mml:mo>,</mml:mo></mml:mrow></mml:msup><mml:mspace linebreak="newline"/><mml:mi>&#x3B6;</mml:mi><mml:mo>=</mml:mo><mml:mfrac><mml:mrow><mml:msub><mml:mi>&#x3C4;</mml:mi><mml:mi>m</mml:mi></mml:msub><mml:mo>+</mml:mo><mml:mn>0</mml:mn><mml:mo>.</mml:mo><mml:mn>5</mml:mn><mml:mi>&#x3B1;</mml:mi><mml:msub><mml:mi>d</mml:mi><mml:mi>m</mml:mi></mml:msub><mml:mfenced><mml:mrow><mml:mn>1</mml:mn><mml:mo>-</mml:mo><mml:mi>K</mml:mi></mml:mrow></mml:mfenced></mml:mrow><mml:msup><mml:mfenced open="[" close="]"><mml:mrow><mml:mn>2</mml:mn><mml:mi>&#x3B1;</mml:mi><mml:msub><mml:mi>d</mml:mi><mml:mi>m</mml:mi></mml:msub><mml:msub><mml:mi>&#x3C4;</mml:mi><mml:mi>m</mml:mi></mml:msub><mml:mfenced><mml:mrow><mml:mi>K</mml:mi><mml:mo>+</mml:mo><mml:mn>1</mml:mn></mml:mrow></mml:mfenced></mml:mrow></mml:mfenced><mml:mfrac><mml:mn>1</mml:mn><mml:mn>2</mml:mn></mml:mfrac></mml:msup></mml:mfrac><mml:mspace linebreak="newline"/><mml:mspace linebreak="newline"/><mml:mspace linebreak="newline"/></mml:math>
	</disp-formula>
</disp-formula-group>
<p>그리고 &lt;<xref ref-type="fig" rid="f004">그림 3</xref>&gt;의 응답 결과를 가지고 추정된 모델의 변수인 안정상태 이득과 시간상수, 그리고 추정된 시간 지연항을 계산하면 식 (<xref ref-type="disp-formula" rid="dm07">7</xref>), (<xref ref-type="disp-formula" rid="dm08">8</xref>), (<xref ref-type="disp-formula" rid="dm09">9</xref>) 와 같다.</p>
<disp-formula-group>
	<disp-formula id="dm07">
		<label>(7)</label>
<mml:math id="dm07-1"><mml:mover><mml:msub><mml:mi>k</mml:mi><mml:mi>m</mml:mi></mml:msub><mml:mo>^</mml:mo></mml:mover><mml:mo>=</mml:mo><mml:mfrac><mml:msub><mml:mi>C</mml:mi><mml:mo>&#x221E;</mml:mo></mml:msub><mml:mrow><mml:msub><mml:mi>k</mml:mi><mml:mi>c</mml:mi></mml:msub><mml:mfenced><mml:mrow><mml:mi>A</mml:mi><mml:mo>-</mml:mo><mml:msub><mml:mi>C</mml:mi><mml:mo>&#x221E;</mml:mo></mml:msub></mml:mrow></mml:mfenced></mml:mrow></mml:mfrac><mml:mspace linebreak="newline"/><mml:mspace linebreak="newline"/><mml:mspace linebreak="newline"/></mml:math>
	</disp-formula>
	<disp-formula id="dm08">
		<label>(8)</label>
<mml:math id="dm08-1"><mml:mover><mml:msub><mml:mi>&#x3C4;</mml:mi><mml:mi>m</mml:mi></mml:msub><mml:mo>^</mml:mo></mml:mover><mml:mo>=</mml:mo><mml:mfrac><mml:mrow><mml:mo>&#x2206;</mml:mo><mml:mi>t</mml:mi></mml:mrow><mml:mi>&#x3C0;</mml:mi></mml:mfrac><mml:mfenced open="[" close="]"><mml:mrow><mml:mi>&#x3B6;</mml:mi><mml:msqrt><mml:mi>K</mml:mi><mml:mo>+</mml:mo><mml:mn>1</mml:mn></mml:msqrt><mml:mo>+</mml:mo><mml:msqrt><mml:msup><mml:mi>&#x3B6;</mml:mi><mml:mn>2</mml:mn></mml:msup><mml:mfenced><mml:mrow><mml:mi>K</mml:mi><mml:mo>+</mml:mo><mml:mn>1</mml:mn></mml:mrow></mml:mfenced><mml:mo>+</mml:mo><mml:mi>K</mml:mi><mml:mo>-</mml:mo><mml:mn>1</mml:mn></mml:msqrt></mml:mrow></mml:mfenced><mml:msup><mml:mfenced open="[" close="]"><mml:mrow><mml:mfenced><mml:mrow><mml:mn>1</mml:mn><mml:mo>-</mml:mo><mml:msup><mml:mi>&#x3B6;</mml:mi><mml:mn>2</mml:mn></mml:msup></mml:mrow></mml:mfenced><mml:mfenced><mml:mrow><mml:mi>K</mml:mi><mml:mo>+</mml:mo><mml:mn>1</mml:mn></mml:mrow></mml:mfenced></mml:mrow></mml:mfenced><mml:mstyle displaystyle="false"><mml:mfrac><mml:mn>1</mml:mn><mml:mn>2</mml:mn></mml:mfrac></mml:mstyle></mml:msup><mml:mspace linebreak="newline"/><mml:mspace linebreak="newline"/><mml:mspace linebreak="newline"/></mml:math>
	</disp-formula>
	<disp-formula id="dm09">
		<label>(9)</label>
<mml:math id="dm09-1"><mml:mover><mml:msub><mml:mi>d</mml:mi><mml:mi>m</mml:mi></mml:msub><mml:mo>^</mml:mo></mml:mover><mml:mo>=</mml:mo><mml:mi>&#x3B1;</mml:mi><mml:mfrac><mml:mrow><mml:mn>2</mml:mn><mml:mo>&#x2206;</mml:mo><mml:mi>t</mml:mi><mml:mfenced open="[" close="]"><mml:mrow><mml:mfenced><mml:mrow><mml:mn>1</mml:mn><mml:mo>-</mml:mo><mml:msup><mml:mi>&#x3B6;</mml:mi><mml:mn>2</mml:mn></mml:msup></mml:mrow></mml:mfenced><mml:msup><mml:mfenced><mml:mrow><mml:mi>K</mml:mi><mml:mo>+</mml:mo><mml:mn>1</mml:mn></mml:mrow></mml:mfenced><mml:mstyle displaystyle="true"><mml:mfrac><mml:mn>1</mml:mn><mml:mn>2</mml:mn></mml:mfrac></mml:mstyle></mml:msup></mml:mrow></mml:mfenced></mml:mrow><mml:mrow><mml:mi mathvariant="normal">&#x3C0;</mml:mi><mml:mfenced open="[" close="]"><mml:mrow><mml:mi mathvariant="normal">&#x3B6;</mml:mi><mml:msqrt><mml:mi>K</mml:mi><mml:mo>+</mml:mo><mml:mn>1</mml:mn></mml:msqrt><mml:mo>+</mml:mo><mml:msqrt><mml:msup><mml:mi mathvariant="normal">&#x3B6;</mml:mi><mml:mn>2</mml:mn></mml:msup><mml:mfenced><mml:mrow><mml:mi>K</mml:mi><mml:mo>+</mml:mo><mml:mn>1</mml:mn></mml:mrow></mml:mfenced><mml:mo>+</mml:mo><mml:mi>K</mml:mi><mml:mo>+</mml:mo><mml:mn>1</mml:mn></mml:msqrt></mml:mrow></mml:mfenced></mml:mrow></mml:mfrac><mml:mspace linebreak="newline"/><mml:mspace linebreak="newline"/><mml:mspace linebreak="newline"/></mml:math>
	</disp-formula>
</disp-formula-group>
<p><italic>A</italic> : 계단 입력의 크기, <italic>C</italic><sub>&#x221E;</sub> : 안정상태 값, &#x2206;<italic>t</italic> : 응답의 반주기</p>
</sec>
<sec id="sec004">
<title>4. 모의실험</title>
<p>작은 시간 지연부터 큰 시간 지연을 갖는 시스템을 모의실험 대상으로 하여 다음과 같은 전달함수를 고려하였다. 그리고 Pade’근사에 의한 시간지연항에 대한 오차를 줄이기 위하여 RBN 알고리즘을 사용하였고 제안된 알고리즘의 설정을 위하여 은닉층은 2개의 층으로, 뉴런의 수는 각각 50개로 설정했다. 학습 횟수는 400번, 학습률은 0.01, 드롭아웃률은 0.6로 설정했고, 제한된 볼츠만 머신(Restricted Boltzmann Machine : RBM)은 볼츠만 머신(Boltzmann machine : BM)에서 진화된 형태의 알고리즘으로 확률적 회귀 신경망으로써 홉필드 네트워크에서 확장된 형태의 네트워크이다[<xref ref-type="bibr" rid="B017">17</xref>,<xref ref-type="bibr" rid="B018">18</xref>]. 그리고 RBM에서 BBPRBM(Beta-Bernoulli Process RBM)을 사용했다<xref ref-type="bibr" rid="B010">[10]</xref>. 그리고 Deep Neural Network Matlab Toolbox를 참고하여 모의실험을 진행하였다<xref ref-type="bibr" rid="B012">[12]</xref>. 그리고활성화 함수로는 시그모이드 함수를 사용했다. 그리고 모의실험을 위하여 Maltab를 사용하였다.[<xref ref-type="bibr" rid="B013">13</xref>-<xref ref-type="bibr" rid="B016">16</xref>]</p>
<sec id="sec004-1">
<title>4.1 Example 1</title>
<disp-formula-group>
	<disp-formula id="dm10">
<mml:math id="dm10-1"><mml:msub><mml:mi>G</mml:mi><mml:mi>p</mml:mi></mml:msub><mml:mfenced><mml:mi>s</mml:mi></mml:mfenced><mml:mo>=</mml:mo><mml:mfrac><mml:mrow><mml:mn>2</mml:mn><mml:mo>.</mml:mo><mml:mn>08</mml:mn><mml:msup><mml:mi>e</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn>0</mml:mn><mml:mo>.</mml:mo><mml:mn>2</mml:mn><mml:mi>s</mml:mi></mml:mrow></mml:msup></mml:mrow><mml:mrow><mml:mfenced><mml:mrow><mml:mn>1</mml:mn><mml:mo>.</mml:mo><mml:mn>2</mml:mn><mml:mi>s</mml:mi><mml:mo>+</mml:mo><mml:mn>1</mml:mn></mml:mrow></mml:mfenced><mml:mfenced><mml:mrow><mml:msup><mml:mi>s</mml:mi><mml:mn>2</mml:mn></mml:msup><mml:mo>+</mml:mo><mml:mn>2</mml:mn><mml:mo>.</mml:mo><mml:mn>5</mml:mn><mml:mi>s</mml:mi><mml:mo>+</mml:mo><mml:mn>1</mml:mn><mml:mo>.</mml:mo><mml:mn>6</mml:mn></mml:mrow></mml:mfenced></mml:mrow></mml:mfrac></mml:math>
	</disp-formula>
</disp-formula-group>
<p>추정된 전달함수에 대한 시스템의 변수 값을 계산하기 위하여 폐루프 전달함수에서 시간 지연 항 <mml:math id="m001-6"><mml:msup><mml:mi>e</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:msub><mml:mi>d</mml:mi><mml:msup><mml:mi>m</mml:mi><mml:mi>s</mml:mi></mml:msup></mml:msub></mml:mrow></mml:msup><mml:mspace linebreak="newline"/><mml:mspace linebreak="newline"/><mml:mspace linebreak="newline"/></mml:math>에 대한 정확한 계산이 필요하기 때문에 Suh방법에 근거한 조절인자 <italic>α</italic>를 딥 러닝 방법으로 구한 결과 1.0이 계산되었고 조절인자로부터 얻어질 수 있는 계산된 추정된 전달함수의 파라미터 값은 &lt;<xref ref-type="table" rid="t001">표 1</xref>&gt;과 같다. 그리고 추정된 모델에 대한 계단응답곡선은 &lt;<xref ref-type="fig" rid="f005">그림 4</xref>&gt;와 같다.</p>
<table-wrap id="t001">
<label>표 1.</label>
<caption>
<title>Example 1에 대한 추정된 모델 파라미터</title>
<p>Table 1. Estimated Model Parameters for Example 1</p>
</caption>
<table frame="box" rules="all" width="100%">
<tbody>
<tr align="center">
<td><italic>R</italic></td>
<td><italic>k</italic><sub>c</sub></td>
<td><italic>k</italic><sub>m</sub></td>
<td><italic>τ</italic><sub>m</sub></td>
<td><italic>d</italic><sub>m</sub></td>
</tr>
<tr align="center">
<td>1.0</td>
<td>1.0</td>
<td>2.0</td>
<td>2.47</td>
<td>1.01</td>
</tr>
</tbody>
</table>
</table-wrap>
<table-wrap id="t002">
<label>표 2.</label>
<caption>
<title>Example 1에 대한 PID제어기 파라미터</title>
<p>Table 2. PID Parameters for Example 1.</p>
</caption>
<table frame="box" rules="all" width="100%">
<tbody>
<tr align="center">
<td></td>
<td><italic>K</italic><sub>u</sub></td>
<td><italic>P</italic><sub>u</sub></td>
<td><italic>K</italic><sub>c</sub></td>
<td><italic>T</italic><sub>i</sub></td>
<td><italic>T</italic><sub>d</sub></td>
</tr>
<tr align="center">
<td>Suh</td>
<td>3.46</td>
<td>4.42</td>
<td>2.21</td>
<td>2.29</td>
<td>0.72</td>
</tr>
<tr align="center">
<td>Z-N</td>
<td>3.72</td>
<td>4.2</td>
<td>2.23</td>
<td>2.10</td>
<td>0.53</td>
</tr>
</tbody>
</table>
</table-wrap>
<p>추정된 전달함수의 파라미터 값들을 이용하여 비례이득 <italic>K</italic><sub><italic>c</italic></sub>, 적분 시간 <italic>T</italic><sub><italic>i</italic></sub>, 미분 시간 <italic>T</italic><sub><italic>d</italic></sub>의 계산된 결과를 &lt;<xref ref-type="table" rid="t002">표 2</xref>&gt;에 나타내었다. 그리고 Ziegler와 Nichols(Z-N)의 제어기 동조 방법에 의한 계산된 결과도 같이 비교하였다</p>
<fig id="f005" orientation="portrait" position="float">
	<label>그림 4.</label>
	<caption>
		<title>Example 1의 모델에 대한 계단 응답 곡선</title>
		<p>Figure 4. Step Response Curve for Example 1</p>
	</caption>
	<graphic xlink:href="../ingestImageView?artiId=ART002496983&amp;imageName=jkits_2019_14_04_359_f005.jpg" position="float" orientation="portrait" xlink:type="simple"></graphic>
	</fig>
<p>딥 러닝을 이용한 Suh 방법에 의해 결정된 PID 제어기의 동조 파라미터 값을 가지고 모의실험의 결과를 &lt;<xref ref-type="fig" rid="f006">그림 5</xref>&gt; 에 나타내었다. &lt;<xref ref-type="fig" rid="f006">그림 5</xref>&gt;에서 볼 수 있듯이 본 논문에서 제안된 제어기의 동조 방법이 Ziegler와 Nichols가 제안한 방법보다 우수함을 알 수 있다.</p>
<fig id="f006" orientation="portrait" position="float">
	<label>그림 5.</label>
	<caption>
		<title>Example 1의 동조효과 비교</title>
		<p>Figure 5. Comparison of Tuning Effects for Example 1</p>
	</caption>
	<graphic xlink:href="../ingestImageView?artiId=ART002496983&amp;imageName=jkits_2019_14_04_359_f006.jpg" position="float" orientation="portrait" xlink:type="simple"></graphic>
	</fig>
</sec>
<sec id="sec004-2">
<title>4.2 Example 2</title>
<disp-formula-group>
	<disp-formula id="dm11">
<mml:math id="dm11-1"><mml:msub><mml:mi>G</mml:mi><mml:mi>p</mml:mi></mml:msub><mml:mfenced><mml:mi>s</mml:mi></mml:mfenced><mml:mo>=</mml:mo><mml:mfrac><mml:msup><mml:mi>e</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mi>s</mml:mi></mml:mrow></mml:msup><mml:mrow><mml:mfenced><mml:mrow><mml:mi>s</mml:mi><mml:mo>+</mml:mo><mml:mn>1</mml:mn></mml:mrow></mml:mfenced><mml:mfenced><mml:mrow><mml:mn>2</mml:mn><mml:mi>s</mml:mi><mml:mo>+</mml:mo><mml:mn>1</mml:mn></mml:mrow></mml:mfenced><mml:mfenced><mml:mrow><mml:mn>0</mml:mn><mml:mo>.</mml:mo><mml:mn>5</mml:mn><mml:mi>s</mml:mi><mml:mo>+</mml:mo><mml:mn>1</mml:mn></mml:mrow></mml:mfenced></mml:mrow></mml:mfrac></mml:math>
	</disp-formula>
</disp-formula-group>
<p>시간지연이 <italic>e</italic><sup><italic>-s</italic></sup>인 전달함수에서 추정된 파라미터 값을 시간 지연 항 <mml:math id="m001-7"><mml:msup><mml:mi>e</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:msub><mml:mi>d</mml:mi><mml:msup><mml:mi>m</mml:mi><mml:mi>s</mml:mi></mml:msup></mml:msub></mml:mrow></mml:msup><mml:mspace linebreak="newline"/><mml:mspace linebreak="newline"/><mml:mspace linebreak="newline"/></mml:math>의 근접한 계산을 위해 Example 1과 동일한 방법으로 구한 조절인자 <italic>α</italic>는0.85가 계산되고 추정된 파라미터 값과 Yunwana와 Seborg(Y-S)가 제안한 방법에 의하여 계산 된 파라미터는 &lt;<xref ref-type="table" rid="t003">표 3</xref>&gt;과 같고 추정된 모델의 계단 응답곡선은 &lt;<xref ref-type="fig" rid="f007">그림 6</xref>&gt;과 같다.</p>
<table-wrap id="t003">
<label>표 3.</label>
<caption>
<title>Example 2에 대한 추정된 모델 파라미터</title>
<p>Table 3. Estimated Model Parameters for Example 2</p>
</caption>
<table frame="box" rules="all" width="100%">
<tbody>
<tr align="center">
<td></td>
<td><italic>R</italic></td>
<td><italic>k</italic><sub>c</sub></td>
<td><italic>k</italic><sub>m</sub></td>
<td><italic>τ</italic><sub>m</sub></td>
<td><italic>d</italic><sub>m</sub></td>
</tr>
<tr align="center">
<td>Suh</td>
<td>1.0</td>
<td>2.0</td>
<td>2.0</td>
<td>2.78</td>
<td>0.92</td>
</tr>
<tr align="center">
<td>Y-S</td>
<td>1.0</td>
<td>2.0</td>
<td>1.0</td>
<td>2.63</td>
<td>3.27</td>
</tr>
</tbody>
</table>
</table-wrap>
<fig id="f007" orientation="portrait" position="float">
	<label>그림 6.</label>
	<caption>
		<title>Example 2의 모델에 대한 계단 응답 곡선</title>
		<p>Figure 6. Step Response Curve for Example 2</p>
	</caption>
	<graphic xlink:href="../ingestImageView?artiId=ART002496983&amp;imageName=jkits_2019_14_04_359_f007.jpg" position="float" orientation="portrait" xlink:type="simple"></graphic>
	</fig>
<p>그리고 &lt;<xref ref-type="table" rid="t003">표 3</xref>&gt;에서 추정된 파라미터 값들을 이용하여 구한 PID 제어기 파라미터 값들과 Y-S, 그리고 Z-N 방법에 의하여 결정된 PID 제어기의 초기 값들은 &lt;<xref ref-type="table" rid="t004">표 4</xref>&gt;와 같다.</p>
<table-wrap id="t004">
<label>표 4.</label>
<caption>
<title>Example 2에 대한 PID제어기 파라미터</title>
<p>Table 4. PID Parameters for Example 2.</p>
</caption>
<table frame="box" rules="all" width="100%">
<tbody>
<tr align="center">
<td></td>
<td><italic>K</italic><sub>u</sub></td>
<td><italic>P</italic><sub>u</sub></td>
<td><italic>K</italic><sub>c</sub></td>
<td><italic>T</italic><sub>i</sub></td>
<td><italic>T</italic><sub>d</sub></td>
</tr>
<tr align="center">
<td>Suh</td>
<td>2.72</td>
<td>7.51</td>
<td>1.51</td>
<td>3.75</td>
<td>0.89</td>
</tr>
<tr align="center">
<td>Z-N</td>
<td>3.08</td>
<td>6.91</td>
<td>1.85</td>
<td>3.45</td>
<td>0.86</td>
</tr>
<tr align="center">
<td>Y-S</td>
<td>1.97</td>
<td>9.75</td>
<td>1.18</td>
<td>4.88</td>
<td>1.22</td>
</tr>
</tbody>
</table>
</table-wrap>
<p>딥 러닝을 이용한 Suh 방법과 Z-N방법, 그리고 Y-S 방법에 의한 동조 효과를 비교한 모의실험 결과는 &lt;<xref ref-type="fig" rid="f008">그림 7</xref>&gt;과 같고 본 논문에서 제안한 방법이 비교를 위한 다른 방법들 보다 좋은 결과를 볼 수 있다.</p>
<fig id="f008" orientation="portrait" position="float">
	<label>그림 7.</label>
	<caption>
		<title>Example 2의 모델에 대한 계단 응답 곡선</title>
		<p>Figure 7. Step Response Curve for Example 2</p>
	</caption>
	<graphic xlink:href="../ingestImageView?artiId=ART002496983&amp;imageName=jkits_2019_14_04_359_f008.jpg" position="float" orientation="portrait" xlink:type="simple"></graphic>
	</fig>
</sec>
<sec id="sec004-3">
<title>4.3 Example 3</title>
<disp-formula-group>
	<disp-formula id="dm12">
<mml:math id="dm12-1"><mml:msub><mml:mi>G</mml:mi><mml:mi>p</mml:mi></mml:msub><mml:mfenced><mml:mi>s</mml:mi></mml:mfenced><mml:mo>=</mml:mo><mml:mfrac><mml:msup><mml:mi>e</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn>3</mml:mn><mml:mi>s</mml:mi></mml:mrow></mml:msup><mml:mrow><mml:msup><mml:mfenced><mml:mrow><mml:mi>s</mml:mi><mml:mo>+</mml:mo><mml:mn>1</mml:mn></mml:mrow></mml:mfenced><mml:mn>2</mml:mn></mml:msup><mml:mfenced><mml:mrow><mml:mn>2</mml:mn><mml:mi>s</mml:mi><mml:mo>+</mml:mo><mml:mn>1</mml:mn></mml:mrow></mml:mfenced></mml:mrow></mml:mfrac></mml:math>
	</disp-formula>
</disp-formula-group>
<p>시간 지연이 <mml:math id="dm003"><mml:msup><mml:mi>e</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:msub><mml:mn>3</mml:mn><mml:mi>s</mml:mi></mml:msub></mml:mrow></mml:msup><mml:mspace linebreak="newline"/><mml:mspace linebreak="newline"/><mml:mspace linebreak="newline"/></mml:math>인 전달함수에서 추정된 파라미터의 계산을 위한 시간 지연 항 <mml:math id="m001-8"><mml:msup><mml:mi>e</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:msub><mml:mi>d</mml:mi><mml:msup><mml:mi>m</mml:mi><mml:mi>s</mml:mi></mml:msup></mml:msub></mml:mrow></mml:msup><mml:mspace linebreak="newline"/><mml:mspace linebreak="newline"/><mml:mspace linebreak="newline"/></mml:math>의 근접한 계산을 위한 딥 러닝을 이용하여 계산된 조절인자 <italic>α</italic>는 2.85가 계산되었고 추정된 시스템의 파라미터 값들과 Y-S방법에 의하여 계산된 파라미터 값들은 &lt;<xref ref-type="table" rid="t005">표 5</xref>&gt;와 같다. 추정된 모델에 대한 계단 응답 곡선은 &lt;<xref ref-type="fig" rid="f009">그림 8</xref>&gt;과 같다. &lt;<xref ref-type="table" rid="t005">표 5</xref>&gt;를 이용하여 딥 러닝을 이용한 Suh방법, Y-S방법, 그리고 Z-N 방법에 의하여 계산된 제어 파라미터 값은 &lt;<xref ref-type="table" rid="t006">표 6</xref>&gt;과 같고 제어기 동조 효과에 대한 계단 응답 곡선은 &lt;<xref ref-type="fig" rid="f010">그림 9</xref>&gt;와 같다.</p>
<table-wrap id="t005">
<label>표 5.</label>
<caption>
<title>Example 3에 대한 추정된 모델 파라미터</title>
<p>Table 5. Estimated Model Parameters for Example 3</p>
</caption>
<table frame="box" rules="all" width="100%">
<tbody>
<tr align="center">
<td></td>
<td><italic>R</italic></td>
<td><italic>k</italic><sub>c</sub></td>
<td><italic>k</italic><sub>m</sub></td>
<td><italic>τ</italic><sub>m</sub></td>
<td><italic>d</italic><sub>m</sub></td>
</tr>
<tr align="center">
<td>Suh</td>
<td>1.0</td>
<td>1.0</td>
<td>1.1</td>
<td>1.45</td>
<td>3.13</td>
</tr>
<tr align="center">
<td>Y-S</td>
<td>1.0</td>
<td>1.0</td>
<td>1.0</td>
<td>1.54</td>
<td>11.75</td>
</tr>
</tbody>
</table>
</table-wrap>
<table-wrap id="t006">
<label>표 6.</label>
<caption>
<title>Example 3에 대한 PID제어기 파라미터</title>
<p>Table 6. PID Parameters for Example 3</p>
</caption>
<table frame="box" rules="all" width="100%">
<tbody>
<tr align="center">
<td></td>
<td><italic>K</italic><sub>u</sub></td>
<td><italic>P</italic><sub>u</sub></td>
<td><italic>K</italic><sub>c</sub></td>
<td><italic>T</italic><sub>i</sub></td>
<td><italic>T</italic><sub>d</sub></td>
</tr>
<tr align="center">
<td>Suh</td>
<td>1.39</td>
<td>11.2</td>
<td>0.96</td>
<td>5.25</td>
<td>1.54</td>
</tr>
<tr align="center">
<td>Z-N</td>
<td>1.73</td>
<td>12.9</td>
<td>1.04</td>
<td>6.45</td>
<td>1.61</td>
</tr>
<tr align="center">
<td>Y-S</td>
<td>1.1</td>
<td>26.4</td>
<td>0.66</td>
<td>13.2</td>
<td>3.3</td>
</tr>
</tbody>
</table>
</table-wrap>
<fig id="f009" orientation="portrait" position="float">
	<label>그림 8.</label>
	<caption>
		<title>Example 3에 대한 계단 응답 곡선</title>
		<p>Figure 8. Step Response Curve for Example 3</p>
	</caption>
	<graphic xlink:href="../ingestImageView?artiId=ART002496983&amp;imageName=jkits_2019_14_04_359_f009.jpg" position="float" orientation="portrait" xlink:type="simple"></graphic>
	</fig>
<fig id="f010" orientation="portrait" position="float">
	<label>그림 9.</label>
	<caption>
		<title>Example 3의 동조효과 비교</title>
		<p>Figure 9. Comparison of Tuning Effects for Example 3</p>
	</caption>
	<graphic xlink:href="../ingestImageView?artiId=ART002496983&amp;imageName=jkits_2019_14_04_359_f010.jpg" position="float" orientation="portrait" xlink:type="simple"></graphic>
	</fig>
<p>&lt;<xref ref-type="fig" rid="f010">그림 9</xref>&gt;에서 알 수 있듯이 본 논문에서 제안한 방법이 다른 두 방법과 비교하여 우수한 결과를 볼 수 있고 Y-S 방법의 경우에는 시간 지연이 큰 시스템의 경우에는 안정성이 떨어지고 응답이 발산하는 결과를 볼 수 있다.</p>
</sec>
</sec>
<sec id="sec005" sec-type="conclusions">
<title>5. 결 론</title>
<p>Suh(1984)가 제시한 동조 방법은 Y다른 동조방법보다 우수함이 입증되었으나 시스템의 전달함수를 알 수 없는 상태에서 Suh방법의 조절인자의 결정을 시간지연 항의 상수 값과 과 비례하여 선택하는 방법을 이용하였다. 그러나 이러한 방법은 비례적으로 조절인자의 상수 값을 대입하는 방법이므로 해석적이지 못한 단점이 있다. 따라서 본 논문에서는 조절인자의 결정을 비례적인 방법이 아닌 딥 러닝 중에서 DBN 알고리즘을 이용한 조절인자의 결정으로 최적의 조절인자의 상수 값을 결정하여 해석적인 방법으로 조절인자 값 <italic>α</italic>를 구하면 딥 러닝을 이용한 최적의 제어기기 동조 방법의 효용성을 검증하였다.</p>
<p>본 논문에서 제시된 동조 방법은 해석적인 방법이므로 Ziegler와 Nichols의 루프 동조 방법과 Cohen과 Coon의 프로세서 반응곡선 방법들의 단점인 시행착오 방법으로 인한 시간이 많이 소모되는 것과 개루프 시스템에서만 고려되어야 하는 제한성이 부여되는 문제점을 해결해주었다. 그리고 폐루프 시스템에서 조절인자의 설정을 비례적인 방법이 아닌 딥 러닝을 이용한 해석적인 방법으로 정확한 제어를 할 수 있었다.</p>
</sec>
</body>
<back>
<ref-list>
<title>References</title>
<!-- [1] J. G. Ziegler, and N. B. Nichols, Optimum settings for automatic controllers, Trans American Society of Mechanical Engineers, Vol. 64, No. 11, p. 759, 1942.-->
<ref id="B001">
<label>[1]</label>
<element-citation publication-type="journal">
<person-group>
<name><surname>Ziegler</surname><given-names>J. G.</given-names></name>
<name><surname>Nichols</surname><given-names>N. B.</given-names></name>
</person-group>
<year>1942</year>
<article-title>Optimum settings for automatic controllers</article-title>
<source>Trans American Society of Mechanical Engineers</source>
<volume>64</volume><issue>11</issue>
<fpage>759</fpage>
</element-citation>
</ref>
<!-- [2] G. H. Cohen, and G. A. Coon, Theoretical considerations of retarded control, Taylor Instrument Companies Bulletin TDS-10A102, 1953.-->
<ref id="B002">
<label>[2]</label>
<element-citation publication-type="journal">
<person-group>
<name><surname>Cohen</surname><given-names>G. H.</given-names></name>
<name><surname>Coon</surname><given-names>G. A.</given-names></name>
</person-group>
<year>1953</year>
<article-title>Theoretical considerations of retarded control</article-title>
<source>Taylor Instrument Companies Bulletin TDS-10A102</source>
</element-citation>
</ref>
<!-- [3] M. Yuwana, and D. E. Seborg, A new method for an-line controller tuning, American Institute of Chemical E Journal, Vol. 28, No.3, pp. 434-440, 1982.-->
<ref id="B003">
<label>[3]</label>
<element-citation publication-type="journal">
<person-group>
<name><surname>Yuwana</surname><given-names>M.</given-names></name>
<name><surname>Seborg</surname><given-names>D. E.</given-names></name>
</person-group>
<year>1982</year>
<article-title>A new method for an-line controller tuning</article-title>
<source>American Institute of Chemical E Journal</source>
<volume>28</volume><issue>3</issue>
<fpage>434</fpage><lpage>440</lpage>
<pub-id pub-id-type="doi">10.1002/aic.690280311</pub-id>
</element-citation>
</ref>
<!-- [4] H. G. Park, and B. S. Suh, A study on the controller tuning, Institute of Electronics Engineers of Korea, Vol. 5, No. 8, pp. 46-54,1984.-->
<ref id="B004">
<label>[4]</label>
<element-citation publication-type="journal">
<person-group>
<name><surname>Park</surname><given-names>H. G.</given-names></name>
<name><surname>Suh</surname><given-names>B. S.</given-names></name>
</person-group>
<year>1984</year>
<article-title>A study on the controller tuning</article-title>
<source>Institute of Electronics Engineers of Korea</source>
<volume>5</volume><issue>8</issue>
<fpage>46</fpage><lpage>54</lpage>
</element-citation>
</ref>
<!-- [5] B. S. Suh, A controller tuning method by pade second order approximation, Vol. 23, No. 8, pp. 647-652, 1986.-->
<ref id="B005">
<label>[5]</label>
<element-citation publication-type="journal">
<person-group>
<name><surname>Suh</surname><given-names>B. S.</given-names></name>
</person-group>
<year>1986</year>
<article-title>A controller tuning method by pade second order approximation</article-title>
<volume>23</volume><issue>8</issue>
<fpage>647</fpage><lpage>652</lpage>
</element-citation>
</ref>
<!-- [6] J. H. Youn, D. G. Koo, and S. W. Sung and J. T. Lee, System identification of the hammerstein process for automatic tuning of PID controller using relay feedback, Proceedings of the International Conference on Control, Automation and Systems, pp. 2266-2269. 2001.-->
<ref id="B006">
<label>[6]</label>
<element-citation publication-type="paper">
<person-group>
<name><surname>Youn</surname><given-names>J. H.</given-names></name>
<name><surname>Koo</surname><given-names>D. G.</given-names></name>
<name><surname>Sung</surname><given-names>S. W.</given-names></name>
<name><surname>Lee</surname><given-names>J. T.</given-names></name>
</person-group>
<year>2001</year>
<article-title>System identification of the hammerstein process for automatic tuning of PID controller using relay feedback</article-title>
<conf-name>Proceedings of the International Conference on Control, Automation and Systems</conf-name>
<fpage>2266</fpage><lpage>2269</lpage>
</element-citation>
</ref>
<!-- [7] S. K. Lee, S. H. Yang, and K. Y. Lee, Improved response of time-delay system with integrator, Korean Institute of Electrical Engineers, pp. 930-932, 1999.-->
<ref id="B007">
<label>[7]</label>
<element-citation publication-type="other">
<person-group>
<name><surname>Lee</surname><given-names>S. K.</given-names></name>
<name><surname>Yang</surname><given-names>S. H.</given-names></name>
<name><surname>Lee</surname><given-names>K. Y.</given-names></name>
</person-group>
<year>1999</year>
<source>Improved response of time-delay system with integrator</source>
<publisher-name>Korean Institute of Electrical Engineers</publisher-name>
<fpage>930</fpage><lpage>932</lpage>
</element-citation>
</ref>
<!-- [8] N. L. V. Truong, H. G. Le, T. N. Dang, L. Le, T. L. Doan, T. T. Nguyen, and M. Y Lee, Fractional-order PI controllers design based on IMC scheme for enhanced performance of dead-time processes, Institute of Control, Robotics and Systems, pp. 1407-1410, 2013.-->
<ref id="B008">
<label>[8]</label>
<element-citation publication-type="other">
<person-group>
<name><surname>Truong</surname><given-names>N. L. V.</given-names></name>
<name><surname>Le</surname><given-names>H. G.</given-names></name>
<name><surname>Dang</surname><given-names>T. N.</given-names></name>
<name><surname>Le</surname><given-names>L.</given-names></name>
<name><surname>Doan</surname><given-names>T. L.</given-names></name>
<name><surname>Nguyen</surname><given-names>T. T.</given-names></name>
<name><surname>Lee</surname><given-names>M. Y.</given-names></name>
</person-group>
<year>2013</year>
<source>Fractional-order PI controllers design based on IMC scheme for enhanced performance of dead-time processes</source>
<comment>Institute of Control, Robotics and Systems</comment>
<fpage>1407</fpage><lpage>1410</lpage>
</element-citation>
</ref>
<!-- [9] M. Tanaka, and M. Okutomi, A Novel inference of a restricted boltzmann machine, International Conference on Pattern Recognition (ICPR2014), Aug. 2014.-->
<ref id="B009">
<label>[9]</label>
<element-citation publication-type="paper">
<person-group>
<name><surname>Tanaka</surname><given-names>M.</given-names></name>
<name><surname>Okutomi</surname><given-names>M.</given-names></name>
</person-group>
<year>2014</year>
<month>Aug.</month>
<article-title>A Novel inference of a restricted boltzmann machine</article-title>
<conf-name>International Conference on Pattern Recognition (ICPR2014)</conf-name>
<pub-id pub-id-type="doi">10.1109/icpr.2014.271</pub-id>
</element-citation>
</ref>
<!-- [10] R. Mittelman, H. Lee, and B Kuipers, Weakly supervised learning of mid-level features with beta-bernoulli process restricted boltzmann machines, 2013 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Jun. 2013.-->
<ref id="B010">
<label>[10]</label>
<element-citation publication-type="paper">
<person-group>
<name><surname>Mittelman</surname><given-names>R.</given-names></name>
<name><surname>Lee</surname><given-names>H.</given-names></name>
<name><surname>Kuipers</surname><given-names>B.</given-names></name>
</person-group>
<year>2013</year>
<month>Jun.</month>
<article-title>Weakly supervised learning of mid-level features with beta-bernoulli process restricted boltzmann machines</article-title>
<conf-name>2013 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)</conf-name>
</element-citation>
</ref>
<!-- [11] D. Mo, A survey on deep Learning: one small step toward AI, 2012.-->
<ref id="B011">
<label>[11]</label>
<element-citation publication-type="other">
<person-group>
<name><surname>Mo</surname><given-names>D.</given-names></name>
</person-group>
<year>2012</year>
<source>A survey on deep Learning: one small step toward AI</source>
</element-citation>
</ref>
<!-- [12] M. Tanaka, and M. Okutomi, A novel inference of a restricted boltzmann machine, International Conference on Pattern Recognition(ICPR2014), 2014.-->
<ref id="B012">
<label>[12]</label>
<element-citation publication-type="paper">
<person-group>
<name><surname>Tanaka</surname><given-names>M.</given-names></name>
<name><surname>Okutomi</surname><given-names>M.</given-names></name>
</person-group>
<year>2014</year>
<article-title>A novel inference of a restricted boltzmann machine</article-title>
<conf-name>International Conference on Pattern Recognition(ICPR2014)</conf-name>
</element-citation>
</ref>
<!-- [13] G. J. Balas, J. C. Doyle, K. Glover, A. Packard, and R. Smith, User's guide to the analysis and synthesis toolbox for MATLAB, The Mathworks Inc., 1991.-->
<ref id="B013">
<label>[13]</label>
<element-citation publication-type="book">
<person-group>
<name><surname>Balas</surname><given-names>G. J.</given-names></name>
<name><surname>Doyle</surname><given-names>J. C.</given-names></name>
<name><surname>Glover</surname><given-names>K.</given-names></name>
<name><surname>Packard</surname><given-names>A.</given-names></name>
<name><surname>Smith</surname><given-names>R.</given-names></name>
</person-group>
<year>1991</year>
<source>User's guide to the analysis and synthesis toolbox for MATLAB</source>
<publisher-name>The Mathworks Inc.</publisher-name>
</element-citation>
</ref>
<!-- [14] F. M. Callier, and C. A. Desoer, Linear system theory, Springer-Verlag New York, Inc., 1991.-->
<ref id="B014">
<label>[14]</label>
<element-citation publication-type="book">
<person-group>
<name><surname>Callier</surname><given-names>F. M.</given-names></name>
<name><surname>Desoer</surname><given-names>C. A.</given-names></name>
</person-group>
<year>1991</year>
<source>Linear system theory</source>
<publisher-name>Springer-Verlag New York, Inc.</publisher-name>
</element-citation>
</ref>
<!-- [15] R. Y. Chiang, and M. O. Safonov, User's guide to the robust control toolbox for MATLAB, The Mathworks Inc., 1988.-->
<ref id="B015">
<label>[15]</label>
<element-citation publication-type="book">
<person-group>
<name><surname>Chiang</surname><given-names>R. Y.</given-names></name>
<name><surname>Safonov</surname><given-names>M. O.</given-names></name>
</person-group>
<year>1988</year>
<source>User's guide to the robust control toolbox for MATLAB</source>
<publisher-name>The Mathworks Inc.</publisher-name>
</element-citation>
</ref>
<!-- [16] H. Saadat, computational aids in control systems using MATLAB, McGraw-Hill, 1993.-->
<ref id="B016">
<label>[16]</label>
<element-citation publication-type="book">
<person-group>
<name><surname>Saadat</surname><given-names>H.</given-names></name>
</person-group>
<year>1993</year>
<source>computational aids in control systems using MATLAB</source>
<publisher-name>McGraw-Hill</publisher-name>
</element-citation>
</ref>
<!-- [17] G. E. Hinton, S. Osindero, and Y. W. Teh, A fast learning algorithm for deep belief nets, neural computation, Vol .18, No. 7, pp. 1527-1554, 2006.-->
<ref id="B017">
<label>[17]</label>
<element-citation publication-type="journal">
<person-group>
<name><surname>Hinton</surname><given-names>G. E.</given-names></name>
<name><surname>Osindero</surname><given-names>S.</given-names></name>
<name><surname>Teh</surname><given-names>Y. W.</given-names></name>
</person-group>
<year>2006</year>
<article-title>A fast learning algorithm for deep belief nets</article-title>
<source>neural computation</source>
<volume>18</volume><issue>7</issue>
<fpage>1527</fpage><lpage>1554</lpage>
<pub-id pub-id-type="doi">10.1162/neco.2006.18.7.1527</pub-id>
</element-citation>
</ref>
<!-- [18] R. Mittelman, H. Lee, and B. Kuipers, Weakly supervised learning of mid-level features with beta-bernoulli process restricted boltzmann machines, IEEE Conference on Computer Vision and Patter n Recognition (CVPR)2013, pp. 476-483, 2013.-->
<ref id="B018">
<label>[18]</label>
<element-citation publication-type="paper">
<person-group>
<name><surname>Mittelman</surname><given-names>R.</given-names></name>
<name><surname>Lee</surname><given-names>H.</given-names></name>
<name><surname>Kuipers</surname><given-names>B.</given-names></name>
</person-group>
<year>2013</year>
<article-title>Weakly supervised learning of mid-level features with beta-bernoulli process restricted boltzmann machines</article-title>
<conf-name>IEEE Conference on Computer Vision and Patter n Recognition (CVPR)2013</conf-name>
<fpage>476</fpage><lpage>483</lpage>
</element-citation>
</ref>
<!-- [19] D. H. Ackley, G. E. Hinton, and T. J. Sejnowski, A learning algorithm for boltz mann machines, Cognitive Science, Vol. 9, No. 1, pp. 147-169, 1985.-->
<ref id="B019">
<label>[19]</label>
<element-citation publication-type="journal">
<person-group>
<name><surname>Ackley</surname><given-names>D. H.</given-names></name>
<name><surname>Hinton</surname><given-names>G. E.</given-names></name>
<name><surname>Sejnowski</surname><given-names>T. J.</given-names></name>
</person-group>
<year>1985</year>
<article-title>A learning algorithm for boltz mann machines</article-title>
<source>Cognitive Science</source>
<volume>9</volume><issue>1</issue>
<fpage>147</fpage><lpage>169</lpage>
<pub-id pub-id-type="doi">10.1207/s15516709cog0901_7</pub-id>
</element-citation>
</ref>
<!-- [20] I. Arel, D. C. Rose, and T. P. Karnowski, Deep machine learning-A new frontier in artificial intelligence research, IEEE Computational Intelligence Magazine, Vol .5, No.4, pp. 13-18, 2010.-->
<ref id="B020">
<label>[20]</label>
<element-citation publication-type="journal">
<person-group>
<name><surname>Arel</surname><given-names>I.</given-names></name>
<name><surname>Rose</surname><given-names>D. C.</given-names></name>
<name><surname>Karnowski</surname><given-names>T. P.</given-names></name>
</person-group>
<year>2010</year>
<article-title>Deep machine learning-A new frontier in artificial intelligence research</article-title>
<source>IEEE Computational Intelligence Magazine</source>
<volume>5</volume><issue>4</issue>
<fpage>13</fpage><lpage>18</lpage>
<pub-id pub-id-type="doi">10.1109/mci.2010.938364</pub-id>
</element-citation>
</ref>
</ref-list>
<ack>
<title>감사의 글</title>
<p>본 논문은 2016년 순천대학교 학술연구비로 연구 되었음.</p>
</ack> 
<bio>
<p><graphic xlink:href="../ingestImageView?artiId=ART002496983&amp;imageName=jkits_2019_14_04_359_f011.jpg"></graphic><bold>Houng-Kun Joung</bold> received the bachelor’s degree in the Department of Electrical Engineering from the Semyung University in 2002. He received the M.S. degree and the Ph.D. degree in the Department of Electrical Engineering from Hanyang University in 2004 and 2017, respectively. From 2011 to 2017, he was a Assistant professor at Korea Polytechnic College. He has been a Assistant professor in the Division of Converged Electronic Engineering at Cheongju University since 2019. His current research interests include Control System, Motor Control System, Power Electronics. He is a Regular member of the KKITS.</p>
<p><italic>E-mail address</italic>: <email>21c8545@cju.ac.kr</email></p>
<p><graphic xlink:href="../ingestImageView?artiId=ART002496983&amp;imageName=jkits_2019_14_04_359_f012.jpg"></graphic><bold>Won-Guen Oh</bold> received a bachelor’s degree, the M.S. degree and the Ph.D. degree in Electrical Communication Engineering from the Hanyang University, Korea, in 1989, 1992 and 1997, respectively. He is currently professor of Department of Multimedia engineering at the Sunchon National University. His research interests include the applications of machine learning to sound systems and control systems, digital signal processing, sound systems, and active noise control.</p>
<p><italic>E-mail address</italic>: <email>owg@scnu.ac.kr</email></p>
</bio>
</back>
</article>
