Control engineering glossary
Every control term used in the lectures, labs and calculators, with the Korean name the site uses, the other names you will meet in Korean textbooks, a short definition and where to learn it.
A
- Ackermann's formula애커만 공식
A closed-form expression for the state-feedback gain of a single-input system: evaluate the desired characteristic polynomial at the matrix A and multiply by the last row of the inverse controllability matrix.
- actuator구동기
Also in Korean textbooks액추에이터, 작동기
The device that turns the controller's command into a physical action, such as a motor driver or a heater. Its limits, not the controller's equations, decide how hard the loop can push.
Korean usage구동기 is the textbook term and appeared about 140 times; the few uses of 액추에이터 now read 구동기.
- algebraic Riccati equationARE대수 리카티 방정식
Also in Korean textbooks리카티 방정식
The matrix equation whose positive-definite solution P gives the steady-state LQR gain K = R⁻¹BᵀP.
- aliasing앨리어싱
Also in Korean textbooks에일리어싱
The effect where a sinusoid above half the sampling frequency produces exactly the same samples as one of lower frequency, so the samples alone cannot tell the two apart.
- angle condition각도 조건
A point s lies on the root locus exactly when the phase of the loop transfer function there is an odd multiple of 180°. The gain that puts a pole there then follows from the magnitude condition.
- angle of arrival도착각
Also in Korean textbooks도달각
The direction in which a root-locus branch enters a complex zero, found by applying the angle condition at a point just next to that zero.
- angle of departure출발각
Also in Korean textbooks이탈각
The direction in which a root-locus branch leaves a complex pole as the gain rises from zero, found from the angle condition at a point just next to the pole.
- anti-windup안티와인드업
Also in Korean textbooks와인드업 방지
Any scheme that stops the integral term from growing while the actuator is saturated, for example by pausing integration or by feeding the clipped part of the command back to the integrator.
Korean usageThe lab button said 적분 누적 방지 and a figure said 와인드업 방지 while the lectures said 안티와인드업; all now use 안티와인드업.
- asymptotes (root locus)점근선
Straight lines that the n − m branches of a root locus approach as the gain grows without bound. They meet the real axis at the centroid σa and leave it at equally spaced angles.
B
- back EMF역기전력
The voltage a spinning DC motor generates in proportion to its speed. It opposes the supply voltage, so the motor carries a built-in speed feedback.
- bandwidth대역폭
The frequency up to which a closed loop follows its reference well, commonly where |T(jω)| falls 3 dB below its low-frequency value. A wider bandwidth means a faster response.
- BIBO stabilityBIBO 안정성
Also in Korean textbooks유계입력·유계출력 안정성
A system is bounded-input, bounded-output stable when every bounded input gives a bounded output. For an LTI system this holds exactly when the impulse response is absolutely integrable.
- block diagram블록선도
Also in Korean textbooks블록 다이어그램
A drawing of a system as transfer-function blocks joined by signal arrows, summing junctions and branch points. Series, parallel and feedback rules reduce it to one transfer function.
- Bode plot보드 선도
Also in Korean textbooks보드선도, Bode 선도
A pair of graphs of a frequency response against log frequency: magnitude in decibels and phase in degrees. Products of factors become sums, so the curves can be sketched from straight-line asymptotes.
Korean usageLecture titles and about 75 passages write 보드 선도 with a space; two search titles that wrote 보드선도 now match.
- breakaway point이탈점
Also in Korean textbooks분지점, 분리점
A point on the real axis where two root-locus branches meet and leave into the complex plane, or arrive from it. It solves dK/ds = 0 with a positive gain.
C
- causality인과성
The property that the output at any time depends only on present and past inputs, never on future ones. Every controller that runs in real time must be causal.
- characteristic equation특성방정식
The equation whose roots are the closed-loop poles: 1 + L(s) = 0 for a single feedback loop, det(sI − A) = 0 for a state model.
- characteristic polynomial특성다항식
The polynomial det(sI − A), or the numerator of 1 + L(s) once the fractions are cleared. Its coefficients feed the Routh test and pole placement.
- closed-loop control폐루프 제어
Also in Korean textbooks폐회로 제어, 닫힌 루프 제어
Control in which the measured output is fed back and compared with the reference, so the input keeps correcting disturbances and model error.
- closed-loop transfer function폐루프 전달함수
The transfer function from reference to output with the loop closed. For unity negative feedback it is L/(1 + L).
- compensator보상기
A dynamic element added to a loop, usually with its own poles and zeros, to shape a response that a single gain cannot reach.
- complementary sensitivity function상보 민감도 함수
Also in Korean textbooks상보 감도 함수
T = L/(1 + L): the transfer function from reference to output and, with a minus sign, from sensor noise to output. Because S + T = 1, the two cannot both be small at the same frequency.
- controllability제어가능성
Also in Korean textbooks가제어성
The property that the input can move the state from any value to any other in finite time. For an LTI model the controllability matrix must have full rank.
Korean usageThe lectures name the property 제어가능성; only a search title used 가제어성, which now appears in parentheses so searches still find it.
- controllable canonical form가제어 표준형
Also in Korean textbooks제어가능 표준형, 가제어 정준형
A state model in which the characteristic polynomial's coefficients sit in one row of A and the input enters only the last state, so state-feedback gains can be matched coefficient by coefficient.
Korean usageThe property is 제어가능성, but the canonical form keeps its established textbook name 가제어 표준형.
- controller제어기
Also in Korean textbooks컨트롤러
The part of a loop that computes the input from the reference and the measurements, for example a PID law or a state-feedback gain.
- convolution합성곱
Also in Korean textbooks컨볼루션
The integral that builds an LTI system's output by adding up shifted, weighted copies of its impulse response. In the s-domain it becomes a plain product.
Korean usageExercises and answers say 합성곱; a search title and a figure said 컨볼루션, which now stays only in parentheses in the search title.
- corner frequency절점 주파수
Also in Korean textbooks꺾임 주파수, 코너 주파수
The frequency 1/τ of a first-order factor, or ωn of a quadratic, where the straight-line Bode asymptotes meet and the slope changes by 20 dB per decade for each order.
- cost function비용함수
Also in Korean textbooks성능 지수, 평가 함수, 이차 비용
A number that scores a whole response. LQR uses the integral of weighted squares of state and input and picks the gain that makes it smallest.
- covariance공분산
A matrix describing how widely random variables spread and how they vary together. The Kalman filter propagates the covariance of its estimation error to decide how much to trust each measurement.
- critically damped임계감쇠
The second-order case ζ = 1: two equal real poles and the quickest return to rest that never overshoots.
D
- damped natural frequency감쇠 고유진동수
The frequency at which an underdamped second-order response actually oscillates: the imaginary part of its poles.
- damping ratio감쇠비
Also in Korean textbooks제동비
The dimensionless number ζ that sets how oscillatory a second-order response is: below 1 it rings, at 1 it is critically damped, above 1 it creeps.
- DC gain직류 이득
Also in Korean textbooks정적 이득, DC 이득
The ratio of the final output to a constant input for a stable system, G(0).
- DC motorDC 모터
Also in Korean textbooks직류 전동기, 직류 모터
An electric motor whose torque is proportional to current and whose back EMF is proportional to speed. It is the main plant of the lab.
- dead time (time delay)시간 지연
Also in Korean textbooks데드타임, 순수 지연, 수송 지연
A pure delay between an input and its first effect on the output. It adds phase lag that grows with frequency without changing the gain, so it eats phase margin.
Korean usageThe lectures say 시간 지연; the lab setting said 데드타임 and a few notes said 순수 지연. 수송 지연 remains only where it names a physical cause, transport through a pipe or a sensor.
- decibeldB데시벨
A logarithmic unit for gain: 20 log₁₀ of an amplitude ratio. A factor of 10 is 20 dB, and multiplying gains adds decibels.
- derivative action미분 동작
Also in Korean textbooks미분 제어, D 동작
The PID term proportional to the rate of change of the error, or of the measurement. It anticipates motion and adds damping, but it also amplifies sensor noise.
- derivative filter미분 필터
A first-order low-pass factor in the derivative term that stops its gain rising at high frequency, which limits the noise passed to the actuator.
- derivative kick미분 킥
Also in Korean textbooks목표값 킥, 설정값 킥
The spike in a PID output when the reference steps and the derivative acts on the error. Differentiating the measurement only (PI-D) removes it.
- digital control디지털 제어
Control computed by a processor at discrete sampling instants, with converters between the continuous plant and the sampled controller.
- discrete-time pole이산 극점
A pole in the z-plane. Sampling maps a continuous pole s to z = e^{sT}, so a stable discrete pole lies inside the unit circle.
- discretization이산화
Turning a continuous-time model or controller into a difference equation for a chosen sampling period, for example with a zero-order-hold equivalent.
- disturbance외란
Also in Korean textbooks교란
An unwanted input that acts on the plant from outside the controller, such as a load torque or a heat loss.
- disturbance rejection외란 억제
Also in Korean textbooks외란 제거
How well a loop keeps its output on target despite disturbances. Feedback scales an output disturbance by the sensitivity function, so high loop gain at low frequency helps.
- dominant poles지배극점
Also in Korean textbooks우세극점, 주극점
The slowest poles of a system, usually a complex pair nearest the imaginary axis, which govern the response when the other poles are much faster.
Korean usageThree spellings were in use (지배극점, 지배 극점, 우세극점); the lectures used 지배극점 most.
- duality쌍대성
The correspondence between control and estimation: (A, C) is observable exactly when (Aᵀ, Cᵀ) is controllable, so observer gains can be designed with state-feedback tools.
E
- eigenvalue고유값
Also in Korean textbooks고윳값, 고유치
A number λ with Av = λv for some nonzero vector v. The eigenvalues of A are the poles of a state model and set the exponents of its modes.
- encirclement count감싸는 횟수
Also in Korean textbooks회전수
The net number N of clockwise turns the Nyquist plot makes around −1. With P open-loop poles in the right half-plane, Z = N + P closed-loop poles are unstable.
- equilibrium point평형점
Also in Korean textbooks평형 상태
A state and constant input at which every derivative is zero, so the system can stay there indefinitely.
- error오차
Also in Korean textbooks편차
The difference between the reference and the measured output, which a feedback controller acts on.
- estimation error추정 오차
The difference between the true state and an observer's or filter's estimate. A good design makes it decay quickly without amplifying noise too much.
F
- feedback피드백
Also in Korean textbooks되먹임, 귀환
Using a measurement of the output to adjust the input, so the system corrects itself instead of relying on a perfect model.
Korean usage피드백 appears more than 350 times; the block-diagram bench, a lab label and a journal note still said 귀환 (귀환 경로 for the feedback path) and now say 피드백.
- feedforward피드포워드
Also in Korean textbooks앞먹임, 전향 보상
An input computed directly from the reference or a measured disturbance, without waiting for an error. It speeds up tracking but cannot correct model error by itself.
- final value theorem최종값 정리
Also in Korean textbooks최종치 정리
A shortcut for the limit of a signal as t → ∞ from its Laplace transform. It is valid only when that limit exists, that is, when every pole of sY(s) lies in the open left half-plane.
- first-order system1차 시스템
Also in Korean textbooks1차 지연 요소
A system with one pole. Its step response rises exponentially and covers 63% of its final change after one time constant.
- frequency response주파수 응답
Also in Korean textbooks주파수응답, 주파수 특성
How a stable LTI system scales and shifts a steady sinusoidal input at each frequency, given by G(jω).
G
- gain이득
Also in Korean textbooks게인
A multiplying factor between signals; for a sinusoid, the ratio of output to input amplitude. Controller gains such as Kp are the values you tune.
Korean usageLectures write 이득 about a thousand times; labs, journal notes and lab descriptions wrote 게인. All now say 이득 (P 이득, 이득 K).
- gain crossover frequency이득 교차 주파수
The frequency where the loop gain |L(jω)| equals 1 (0 dB). The phase margin is read there, and it roughly sets the closed-loop speed.
- gain marginGM이득여유
Also in Korean textbooks이득 여유, 게인 마진
How many times the loop gain could be multiplied before the closed loop becomes unstable, read at the frequency where the phase is −180°.
Korean usage이득여유 (about 80 uses) and 이득 여유 (9) were mixed; the solid form wins, like 위상여유.
H
- hysteresis히스테리시스
A switching rule with two thresholds: an on-off controller turns on below one value and off above another. The gap trades ripple for fewer switches.
I
- imaginary-axis crossing허수축 교차
The gain and frequency at which a root-locus branch crosses into the right half-plane, found from the Routh array or by putting s = jω into the characteristic equation.
- impulse response임펄스 응답
Also in Korean textbooks충격응답, 충격 응답
The output of a system at rest to a unit impulse. For an LTI system it is the inverse Laplace transform of the transfer function and determines every other response.
Korean usageThe lectures wrote 충격응답 while search titles wrote 임펄스 응답. Korean control and signals textbooks mostly use 임펄스 응답, so the site follows them.
- initial conditions초기조건
Also in Korean textbooks초기 조건, 초기값
The values a system stores at the moment its response starts, such as positions, velocities, currents and temperatures. Together with the input from then on they fix the whole response of a causal system, and a transfer function assumes they are all zero.
- initial value theorem초기값 정리
Also in Korean textbooks초깃값 정리, 초기치 정리
A shortcut for y(0⁺) from Y(s), useful to check the first instant of a response before inverting a transform.
- integral action적분 동작
Also in Korean textbooks적분 제어, I 동작
The PID term proportional to the accumulated error. It keeps pushing while any constant error remains, which removes steady-state error under a constant load.
- integral servo적분 서보
State feedback augmented with the integral of the tracking error as an extra state, so a step reference is followed with zero steady-state error despite model error.
- integrator적분기
A block 1/s whose output is the running integral of its input. The number of integrators in a loop is its system type.
- integrator windup적분 와인드업
Also in Korean textbooks적분기 와인드업, 와인드업
The build-up of the integral term while the actuator is saturated, which later shows up as a large overshoot and a slow recovery.
Korean usageSearch titles said 적분 와인드업 while lecture 31, the lab and the journal described the effect as 적분 누적. The term is now 적분 와인드업, shortened to 와인드업 once it has been introduced.
- inverse Laplace transform라플라스 역변환
Also in Korean textbooks역라플라스 변환
Recovering a time function from its Laplace transform, usually by partial-fraction expansion and a table of standard pairs.
- inverted pendulum (cart-pole)역진자
Also in Korean textbooks도립 진자, 역진자 카트
A pole balanced upright on a moving cart. Unstable at the top, it is the standard example for linearization, LQR and state estimation.
K
- Kalman filter칼만 필터
A recursive estimator that blends a model prediction with noisy measurements, weighting each by its uncertainty. For linear models with Gaussian noise its estimate has the smallest error variance.
L
- lag compensator지상 보상기
Also in Korean textbooks래그 보상기
A compensator whose pole lies closer to the origin than its zero. It raises low-frequency gain to cut steady-state error while changing the phase little near crossover.
Korean usageThe newest lectures said 지상 and the labs and older lectures said 래그. The site now writes 지상 보상기, adding (래그) once in titles and in the lab's controller name.
- Laplace transform라플라스 변환
An integral transform that turns linear differential equations with initial conditions into algebraic equations in s.
- lead compensator진상 보상기
Also in Korean textbooks리드 보상기
A compensator whose zero lies closer to the origin than its pole. It adds positive phase near crossover, raising phase margin and speed at the cost of more high-frequency gain.
Korean usageThe newest lectures said 진상 and the labs and older lectures said 리드. The site now writes 진상 보상기, adding (리드) once in titles and in the lab's controller name.
- lead–lag compensator진상-지상 보상기
Also in Korean textbooks리드-래그 보상기
A lead section and a lag section in series, used when a design needs both more phase near crossover and more low-frequency gain.
- left half-planeLHP좌반평면
The part of the complex s-plane with negative real part. A continuous-time LTI system is stable when all its poles lie there.
- linear time-invariant systemLTI선형 시불변
Also in Korean textbooks선형 시불변 시스템, LTI 시스템
A system that obeys superposition and whose behaviour does not change with time, so one impulse response or transfer function describes it completely.
- linear-quadratic regulatorLQR선형 이차 조절기
Also in Korean textbooks최적 레귤레이터
The state-feedback gain that minimizes a quadratic cost on state and input. The weights Q and R say how much you care about error compared with effort.
- linear-quadratic-Gaussian controlLQGLQG 제어
Also in Korean textbooks선형 이차 가우시안 제어
LQR state feedback applied to a Kalman filter's estimate. The separation principle lets the two be designed one at a time.
- linearization선형화
Replacing a nonlinear model by its first-order Taylor approximation about an operating point. The result is valid only for small deviations from that point.
- loop gain루프 이득
The gain picked up once around a feedback loop, |L(jω)| at a given frequency. Where it is large, feedback suppresses disturbances and model error.
- loop transfer function루프 전달함수
Also in Korean textbooks개루프 전달함수, 일순 전달함수
The product of every transfer function around the loop, L(s) = C(s)G(s)H(s), with the loop broken at one point. Root-locus, Bode and Nyquist analyses all work on it.
- Luenberger observer루엔버거 관측기
A copy of the plant model driven by the same input and corrected by the output error through a gain L, chosen to place the poles of the estimation error.
Korean usageLecture 38's title and several pages kept the Latin name inside Korean sentences; the Korean text now writes 루엔버거 like the other people's names.
M
- magnitude condition크기 조건
At a point that satisfies the angle condition, the gain that places a closed-loop pole there is the one that makes |K G(s)H(s)| equal to 1.
- marginal stability임계 안정
Also in Korean textbooks한계 안정
The boundary case with no poles in the right half-plane and simple poles on the imaginary axis. Free responses neither decay nor grow, and some bounded inputs make the output grow.
- Mason's gain formula메이슨 이득 공식
Also in Korean textbooks메이슨 공식
A rule that writes the transfer function of a signal-flow graph directly from its forward paths, its loops and the sets of loops that do not touch, without reducing the graph step by step.
- mass–spring–damper질량–스프링–댐퍼
Also in Korean textbooks스프링-질량-감쇠기 계
A mass on a spring with a viscous damper, the standard mechanical second-order system; mass, stiffness and damping set ωn and ζ.
- mathematical model수학적 모델
A set of equations that predicts how a system responds to its inputs, kept only as detailed as the question requires.
- maximum phase lead최대 진상각
Also in Korean textbooks최대 위상 앞섬, 최대 위상
The largest phase a lead compensator can add, reached at the geometric mean of its zero and pole frequencies; it depends only on their ratio.
- measurement noise측정 잡음
Also in Korean textbooks센서 잡음, 노이즈
Random error added by the sensor. Feedback passes it to the output through T(s), so high-frequency gain, derivative action especially, has to be limited.
- minimum-order observer최소차수 관측기
Also in Korean textbooks축소 차수 관측기
An observer that estimates only the states not measured directly, so its order is the number of states minus the number of independent measurements.
- minimum-phase system최소위상 시스템
A stable system with no zeros and no delay in the right half-plane. Its phase follows from its magnitude curve, so the Bode magnitude plot tells the whole story.
- mode모드
Also in Korean textbooks고유 모드
One exponential or oscillating component e^{λt} of a free response, belonging to one eigenvalue λ of the system.
- model error (model uncertainty)모델 오차
Also in Korean textbooks모델 불확실성, 모델링 오차
The difference between the model used for design and the real system, such as a changed inertia or an unmodelled delay. A robust design keeps working despite it.
- model validation모델 검증
Checking a fitted model against data it was not fitted to. A model that only reproduces its own fitting run has not been validated.
N
- natural frequency고유진동수
Also in Korean textbooks비감쇠 고유진동수, 고유 주파수
The undamped natural frequency ωn: the frequency a second-order system would oscillate at without damping, and the distance of its poles from the origin.
- negative feedback음의 피드백
Also in Korean textbooks부귀환, 부궤환
Feedback that subtracts the measurement from the reference, so the controller pushes against any deviation.
- Nyquist frequency나이퀴스트 주파수
Half the sampling frequency. Only sinusoids below it are represented unambiguously by their samples.
- Nyquist plot나이퀴스트 선도
The curve that L(jω) traces in the complex plane as ω runs over all frequencies. Its position relative to −1 shows stability and margins.
- Nyquist stability criterion나이퀴스트 안정도 판별법
Also in Korean textbooks나이퀴스트 판별법
A test that counts the encirclements of −1 by the Nyquist plot and adds the number of unstable open-loop poles to find the number of unstable closed-loop poles.
Korean usageStability as a property is 안정성; this criterion keeps its established name with 안정도.
O
- observability관측가능성
Also in Korean textbooks가관측성
The property that the initial state can be worked out from the input and output over a finite time. For an LTI model the observability matrix must have full rank.
Korean usageAs with 제어가능성, the lectures' 관측가능성 is the name; 가관측성 stays in parentheses in the search title, and 가관측 표준형 keeps its textbook name.
- observer (state observer)관측기
Also in Korean textbooks상태 관측기, 옵저버
An algorithm that runs a model alongside the plant and uses the measured outputs to estimate the states that are not measured.
- on-off controlON-OFF 제어
Also in Korean textbooks온오프 제어, 2위치 제어
Control whose actuator is either fully on or fully off. It is simple and robust, but the output always cycles, usually within a hysteresis band.
- open-loop control개루프 제어
Also in Korean textbooks개회로 제어, 열린 루프 제어
Control that applies a precomputed input without measuring the output, so it cannot correct disturbances or model error.
- operating point작동점
Also in Korean textbooks동작점, 운전점
The steady condition, a state and an input, about which a nonlinear system is linearized and operated.
- overdamped과감쇠
The second-order case ζ > 1: two distinct real poles and a slow response without oscillation, dominated by the slower pole.
- overshoot%OS오버슈트
Also in Korean textbooks최대 오버슈트, 백분율 오버슈트, 최대 초과량
How far a step response rises past its final value, usually as a percentage. For the standard second-order system it depends only on ζ.
P
- partial-fraction expansion부분분수 전개
Also in Korean textbooks부분분수 분해
Writing a rational function as a sum of simple terms, one for each pole, so that each term can be inverted from a transform table.
- peak time피크시간
Also in Korean textbooks최대 시간, 첨두 시간
The time at which an underdamped step response reaches its first and highest peak.
- phase crossover frequency위상 교차 주파수
The frequency at which the loop phase is −180°; the gain margin is read there.
- phase marginPM위상여유
Also in Korean textbooks위상 여유, 위상 마진
How much extra phase lag the loop could tolerate at the gain crossover frequency before the closed loop becomes unstable. It also predicts damping and overshoot.
Korean usage위상여유 (about 135 uses) and 위상 여유 (35) were mixed, sometimes on one page; the solid form wins.
- PI controlPI 제어
Proportional plus integral control. The integral term removes steady-state error under a constant load and the proportional term sets the speed.
- PID controlPID 제어
Control that adds three terms proportional to the error, its integral and its derivative; the controller used most widely in practice.
- PID tuningPID 동조
Also in Korean textbooksPID 튜닝
Choosing Kp, Ki and Kd from an experiment or a model, typically starting from a rule such as Ziegler–Nichols and then refining for margins and actuator limits.
- plant제어 대상
Also in Korean textbooks플랜트, 공정
The physical system being controlled, such as a motor, a tank or a heated body, together with its actuator and sensor dynamics when they matter.
- pole극점
Also in Korean textbooks극
A value of s at which a transfer function becomes infinite, a root of its denominator. Each pole contributes a mode e^{pt} to the response.
- pole placement극점 배치
Also in Korean textbooks극배치
Choosing a state-feedback gain so that the closed-loop poles land where you want them; any locations are possible when the system is controllable.
- pole–zero cancellation극점·영점 상쇄
Placing a zero on a pole so the pair drops out of a transfer function. It is safe only for well-known, well-damped stable poles, because the cancelled mode still responds to disturbances.
- proportional control비례 제어
Also in Korean textbooks비례 동작, P 제어
Control whose input is the error times a constant gain Kp. Raising Kp makes the loop faster and stiffer but increases overshoot and input demand, and a type-0 loop keeps some steady-state error.
R
- ramp input램프 입력
Also in Korean textbooks경사 입력
A reference that increases at a constant rate. Following it with zero steady-state error takes a type-2 loop; a type-1 loop lags by a constant 1/Kv.
- reaction curve반응곡선
Also in Korean textbooks공정 반응곡선
The open-loop step response of a process, read as a delay L and a time constant T. The first Ziegler–Nichols rule takes its PID gains from them.
- reference (setpoint)목표값
Also in Korean textbooks기준 입력, 설정값
The value the output is asked to follow; a feedback loop compares it with the measurement.
Korean usageMost pages said 목표값; the PID tuning lecture said 설정값 and some state-feedback notes said 기준값 (기준값 이득). All now say 목표값: 목표값 응답, 목표값 이득. 기준값 remains only for a baseline run or a nominal value, which is a different idea.
- resonant frequency공진 주파수
The frequency at which a lightly damped system's frequency-response magnitude peaks, a little below ωn.
- resonant peak공진 피크
Also in Korean textbooks공진 첨두값
The maximum of the frequency-response magnitude of a lightly damped system. It appears when ζ < 1/√2 and grows sharply as damping falls.
- right half-planeRHP우반평면
Also in Korean textbooks오른쪽 반평면, 우반면
The part of the s-plane with positive real part. One closed-loop pole there is enough to make the loop unstable.
- right-half-plane zero우반평면 영점
Also in Korean textbooks비최소위상 영점
A zero with positive real part. It makes a step response start in the wrong direction and limits how fast any controller can make the loop.
- rise time상승시간
Also in Korean textbooks상승 시간
How long a step response takes to climb between two levels, usually from 10% to 90% of its final value.
- RLC circuitRLC 회로
A circuit of a resistor, an inductor and a capacitor. The series RLC circuit is the electrical twin of the mass–spring–damper.
- robustness강인성
Also in Korean textbooks견실성, 강건성
The ability of a control design to keep stability and acceptable performance when the real plant differs from the model.
- root locus근궤적
The paths traced by the closed-loop poles in the s-plane as one gain varies from zero to infinity.
- Routh array라우스 표
Also in Korean textbooks라우스 배열, Routh 표
The table built from a characteristic polynomial's coefficients. The number of sign changes in its first column is the number of roots in the right half-plane.
Korean usage라우스 표 (36 uses) and 라우스 배열 (9) both appeared, plus Routh and 루스 in a few figures; the site now writes 라우스 표.
- Routh–Hurwitz criterion라우스–후르비츠 판별법
Also in Korean textbooks라우스–후르비츠 안정도 판별법, 라우스 판별법
A test that decides from the coefficients alone whether a polynomial has roots in the right half-plane, without solving for them.
S
- sampling샘플링
Also in Korean textbooks표본화
Reading a continuous signal at discrete instants, usually every T seconds. Whatever happens between samples is invisible to the controller.
Korean usageThe digital-control lectures and labs said 샘플링 (about 130 uses) while some exercises said 표본화 and 표본. The site now writes 샘플링, and 샘플 for one sample.
- sampling period샘플링 주기
Also in Korean textbooks표본화 주기, 샘플 주기
The time T between samples; its inverse is the sampling frequency. A longer period adds delay and phase lag to a digital loop.
- saturation포화
A hard limit on an actuator's output, such as ±12 V. Once it is reached, extra controller demand has no effect and the loop is temporarily open.
- second-order system2차 시스템
A system with two poles. Its standard form is fixed by ωn and ζ and models a mass–spring–damper, an RLC circuit and many closed loops.
- sensitivity function민감도 함수
Also in Korean textbooks감도 함수
S = 1/(1 + L): the factor by which feedback reduces the effect of output disturbances and of small plant changes. Feedback helps at frequencies where |S| < 1.
- sensor센서
Also in Korean textbooks검출기
The device that measures the output for the controller. Its noise, delay and range limit what feedback can achieve.
- separation principle분리 원리
For linear systems the state-feedback gain and the observer or Kalman gain can be designed independently; the closed-loop poles are the two sets together.
- settling time정착시간
Also in Korean textbooks정정시간, 정착 시간
The time after which a step response stays within a band around its final value, usually ±2%.
- signal-flow graph신호흐름선도
Also in Korean textbooks신호 흐름 그래프
A graph whose nodes are signals and whose directed branches carry gains. It holds the same information as a block diagram and feeds Mason's formula directly.
- sinusoid (sine wave)정현파
Also in Korean textbooks사인파
A signal A sin(ωt + φ). An LTI system turns a sinusoid into another of the same frequency, which is why a frequency response describes it.
Korean usage정현파 and 사인파 were used equally often (43 each); circuit and control textbooks say 정현파, so the site does too.
- stability안정성
Also in Korean textbooks안정도
For an LTI system, the property that every free response decays to zero; equivalently, every pole lies in the open left half-plane, or inside the unit circle in discrete time.
Korean usage안정성 appears about 130 times and 안정도 a dozen; 안정도 now survives only in the name 나이퀴스트 안정도 판별법.
- state상태
The smallest set of variables that, together with future inputs, determines all future behaviour, such as the position and velocity of a mass.
- state equation상태방정식
Also in Korean textbooks상태 방정식
The first-order matrix differential equation for the state. Paired with the output equation it forms a state-space model.
- state feedback상태 피드백
A control law that uses every state, u = −Kx plus a reference term. With the full state available it can place all the closed-loop poles.
- state variable상태변수
Also in Korean textbooks상태 변수
One component of the state vector, often a quantity tied to stored energy such as a position, a velocity, a current or a temperature.
Learn it03States and the cart-pole model- state-space representation상태공간 표현
Also in Korean textbooks상태공간 모델
A model written as first-order equations for a state vector rather than as one transfer function. It handles several inputs and outputs and shows internal behaviour.
- state-transition matrix상태천이행렬
Also in Korean textbooks행렬 지수함수, 상태 전이 행렬
The matrix e^{At} that carries the state of an LTI system without input from time 0 to time t.
- static error constants오차 상수
Also in Korean textbooks위치·속도·가속도 오차 상수, 편차 상수
The limits Kp, Kv and Ka of the loop transfer function that give the steady-state error for step, ramp and parabolic references.
- steady-state error정상상태 오차
Also in Korean textbooks정상 편차, 정상오차
The error that remains after the transient has died away. For a given kind of reference, the loop's type and gain set its size.
Korean usageThe lectures write 정상상태 오차; exercises and lab text also had 정상 오차, 정상오차 and 잔류 오차, which now read 정상상태 오차.
- step response계단응답
Also in Korean textbooks계단 응답, 스텝 응답
The output when the input jumps from zero to a constant: the most common test of a loop's speed, overshoot and accuracy.
- superposition중첩의 원리
For a linear system the response to a sum of inputs is the sum of the responses, and scaling an input scales its response.
- system identification시스템 식별
Also in Korean textbooks시스템 동정
Building or fitting a model from measured input–output data rather than from physical laws alone.
- system type시스템 형
Also in Korean textbooks형 번호
The number of pure integrators, poles at s = 0, in the loop transfer function. It decides which references a loop can follow with zero steady-state error.
Korean usageKorean text writes the type as 0형, 1형, 2형; the lab buttons and a journal note that said Type 0–2 now match.
T
- time constant시정수
Also in Korean textbooks시상수, 시간 상수
The time τ a first-order response takes to cover 63% of its final change; after about 4τ it is within 2%.
- tracking (reference following)목표 추종
Also in Korean textbooks추종
How closely the output follows a changing reference. Feedback improves it at frequencies where the loop gain is high.
- transfer function전달함수
Also in Korean textbooks전달 함수
The ratio of the Laplace transforms of output and input of an LTI system with zero initial conditions.
- transient response과도응답
Also in Korean textbooks과도 응답
The part of a response that dies away before the output settles. The poles set its speed and shape.
- two-degree-of-freedom control2-DOF2자유도 제어
A structure with separate paths for the reference and for the measurement, so tracking can be shaped without changing the feedback's disturbance response and robustness.
U
- ultimate gain임계이득
Also in Korean textbooks한계 이득
The proportional gain at which a loop just sustains a steady oscillation. The second Ziegler–Nichols rule uses it with the period of that oscillation.
- ultimate period임계주기
Also in Korean textbooks한계 주기
The period of the sustained oscillation at the ultimate gain.
- underdamped부족감쇠
Also in Korean textbooks저감쇠, 미흡감쇠
The second-order case 0 < ζ < 1: complex poles, and a step response that overshoots and rings before settling.
- unit circle단위원
The circle |z| = 1 in the z-plane: the stability boundary for discrete-time systems, as the imaginary axis is for continuous time.
- unit impulse단위 임펄스
Also in Korean textbooks디랙 델타 함수, 단위 충격 함수
An idealized pulse of zero width and unit area, δ(t). Its Laplace transform is 1.
- unit step단위 계단
Also in Korean textbooks단위 계단 함수, 단위계단
The signal that is 0 before t = 0 and 1 afterwards. Its Laplace transform is 1/s.
W
- weights Q and R가중치
Also in Korean textbooks가중 행렬, 가중치 행렬
The matrices in the LQR cost that price state error (Q) and input use (R). Only their relative size changes the gain.
Z
- z-transformz변환
Also in Korean textbooksZ 변환
The discrete-time counterpart of the Laplace transform, which turns difference equations into algebra in z.
- zero영점
A value of s at which a transfer function becomes zero, a root of its numerator. Zeros do not decide stability but change the size and shape of each mode in the response.
- zero-order holdZOH영차 홀드
Also in Korean textbooks영차 유지기
The converter that keeps each computed output constant until the next sample. It adds roughly half a sampling period of delay.
- zero-phase filter영위상 필터
A filter run over a recorded signal forwards and then backwards, so that features keep their timing instead of lagging. Each output uses samples from after its own time, so the filter is noncausal and works only on recorded data or with a delay.
- Ziegler–Nichols tuning지글러–니콜스 동조법
Also in Korean textbooks지글러–니콜스 방법
Two classic rules that set PID gains from either the open-loop reaction curve or the ultimate gain and period. They give a quick but aggressive starting point.
No term matches. Try the English or the Korean name, or fewer letters.
How the Korean names were chosen
- One name per idea. Where lectures, labs and calculators used two Korean names for one term, the site now uses the one that most lectures and Korean university textbooks use. Other names you may meet in a textbook are listed under each entry.
- Lead and lag compensators are 진상 보상기 and 지상 보상기. The loanwords 리드 and 래그 appear only in parentheses, once, in titles and in the lab's controller names, because many students first meet the compensators under those names.
- A loanword stays where textbooks mostly use it (피드백, 오버슈트, 샘플링, 앨리어싱, 임펄스 응답) and gives way where a Korean term dominates (이득 rather than 게인, 구동기 rather than 액추에이터).
- People's names are written in Hangul in Korean text: 라우스, 후르비츠, 나이퀴스트, 루엔버거, 지글러–니콜스. A search title may add the original spelling in parentheses.
- Established compound terms are written solid (전달함수, 계단응답, 정착시간, 위상여유, 이득여유, 지배극점); terms that textbooks space keep their space (보드 선도, 주파수 응답, 정상상태 오차, 단위 계단).
- Stability as a property is 안정성; 안정도 remains only in the established name 나이퀴스트 안정도 판별법.
Definitions are written for this site to match how each term is used in the lectures. Korean alternatives are the names most often found in Korean university textbooks; they are listed so you can recognise them, not because the site uses them.