• 대한전기학회
Mobile QR Code QR CODE : The Transactions of the Korean Institute of Electrical Engineers
  • COPE
  • kcse
  • 한국과학기술단체총연합회
  • 한국학술지인용색인
  • Scopus
  • crossref
  • orcid

  1. (Department of Electrical Engineering, Hanbat National University, Korea)



Takagi-Sugeno (T-S) fuzzy model, quadrotor, Lyapunov stability, guaranteed cost control

1. 서 론

무인 항공기의 한 종류인 쿼드로터는 수직 이착륙과 호버링이 가능하다는 장점을 가진다(1). 이러한 장점을 가진 쿼드로터는 군사, 산업, 민간 분야에서 다양한 목적으로 점차 사용이 증대되고 있다.

최근 쿼드로터에 대한 다양한 제어 기법이 연구되었다(2). (2)에서는 PID 또는 LQR의 선형 제어기법을 적용한 추종 제어를 제안하였으며 실험을 통해 결과를 도출하였다. (3)는 궤환 선형화와 슬라이딩 모드 제어 기법을 적용한 위치 및 자세 제어를 제시하였다. (5)은 적응 제어 기법을 기반으로 옥토로터와 외란에 강인한 자세제어를 연구하였다. (2)에 제시된 선형 제어 기법은 선형화 오차로 인해 전역 안정도를 보장하지 못한다는 단점이 존재한다. 이에 반해 (3)에 제안된 비선형 제어 기법은 전역 안정도를 보장하지만 제어 신호 구현을 위해서는 많은 계산량이 요구되므로 실제 플랜트에 적용할 경우 시간 응답성이 지연될 가능성이 높다.

한편, 퍼지 모델 기반 제어 기법은 부채꼴 비선형(sector nonlinearity) 기법을 이용하여 제어 대상의 동역학을 효과적으로 모델링 할 수 있으며 국소적으로 모델링 오차가 발생하지 않는다. 또한, 비선형 시스템의 제어기 구현 시 계산량이 적은 장점을 가진다(7,11-12). 이러한 장점을 가진 퍼지 모델에 기반한 쿼드로터의 자세, 고도 및 위치에 대한 제어 기법이 제시되었다(1,13-15). (1)에서는 극점 배치 방법을 이용한 선형 행렬 부등식 형태의 제어기 설계 조건을 제시하였으며 (13)에서는 보장 비용 제어 기법을 이용하여 배터리의 에너지 소비를 최소화하는 방법을 제안하였다. (14)에서는 상호 결합된 추종 오차 동역학에 대한 제어기 설계 시 위치, 고도 및 자세의 제어입력이 독립적으로 설계 됨을 수학적으로 증명하였으며 (15)에서는 불완전한 전반부 정합 하에서의 $H_∞$ 제어 기법에 대해 논하였다. 또한, (16)에서는 매개변수의 불확실성을 고려한 보장 비용 제어 기법에 대한 연구를 진행하였으나 외란을 고려하지 않았다.

본 논문에서는 (16)의 연구를 확장하여 쿼드로터의 퍼지 모델 기반 $H_∞$ 보장 비용 제어 기법을 제안한다. 외란을 고려한 쿼드로터 동역학은 부채꼴 비선형 기법을 이용하여 퍼지 모델로 표현된다. 쿼드로터의 위치 및 자세제어를 위한 점근적 안정화 조건은 리아프노브 안정도 기준 함수를 이용하여 선형 행렬 부등식 형태로 유도된다. 마지막으로 시뮬레이션을 통해 제안된 기법의 우수성을 검증한다.

2. 시스템

2.1 T-S 퍼지 모델

다음은 T-S 퍼지 모델의 정의이다 :

Rule $i:$ IF $z_{1}$ is $M_{i1}$ and $\ldots$ and $z_{p}$ is $M_{ip},$

THEN $\dot{\chi}=A_{i} \chi+B_{i} u+B_{w} w$

여기서 $\chi \in \mathbb{R}^{n}$은 상태변수; $u \in \mathbb{R}^{n}$은 제어입력; $w \in \mathbb{R}^{q}$는 외란; $\mathbb{I}_{r}=\{1,2, \ldots, r\}$이며 $i \in \mathbb{I}_{r}$은 $i$번째 퍼지 규칙; $\mathbb{I}_{p}=\{1,2, \ldots, p\}$이며 $z_h, h \in \mathbb{I}_{p}$는 번째 전반부 변수; $M_{i h}, \quad(i, h) \in \mathbb{I}_{r} \times \mathbb{I}_{p}$는 $i$번째 퍼지 규칙에서 $z_h$의 퍼지 집합이다. 싱글톤 퍼지화, 곱셈 추론 및 중심값-평균 비퍼지화를 이용한 전체 퍼지 시스템은

(1)
$\dot{\chi}=\sum_{i=1}^{r} \theta_{i}(z)\left(A_{i} \chi+\tilde{g} u+B_{w} w\right)$

로 나타낼 수 있다(7), 여기서

$\theta_{i}(z)=\frac{w_{i}(z)}{\sum_{i=1}^{r} w_{i}(z)}, w_{i}(z)=\prod_{j=1}^{p} M_{i j}\left(z_{j}\right)$

함수 $\theta_{i}(z)$는 , $0 \leq \theta_{i}(z) \leq 1, \sum_{i=1}^{r} \theta_{i}(z)=1$을 만족한다.

정의 1 (9) : 상수 $\gamma \in \mathbb{R}_{>0}$에 대하여, 퍼지 시스템은 다음의 조건을 만족할 경우 $H_∞$ 제어 성능을 만족한다.

$i. w=0$일 때, 퍼지 시스템은 점근적으로 안정하다.

$ii. $초기 조건 $\chi(0)=0$에 대하여. 퍼지 시스템은

$\int_{0}^{t_{f}} \chi^{T} H \chi d t \leq \gamma^{2} \int_{0}^{t_{f}} w^{T} w d t$

을 만족한다, 여기서 $H=H^T>0$이다.

그림 1 선체 좌표계와 관성 좌표계의 쿼드로터의 구조

Fig. 1. The schematic of quadrotor of body frame and inertial frame

../../Resources/kiee/KIEE.2019.68.11.1396/fig1.png

2.2 쿼드로터 동역학

오일러-라그랑지안 접근법을 이용한 자세제어에 대한 쿼드로터 동역학(1)은 다음과 같다(그림 1 참고):

(2)
$$\ddot{x}=(\cos \phi \sin \theta \cos \psi+\sin \phi \sin \psi) \frac{T}{m}+w_{x}$$ $$\ddot{y}=(\cos \phi \sin \theta \cos \psi-\sin \phi \sin \psi) \frac{T}{m}+w_{y}$$ $$\ddot{z}=-g+\cos \phi \cos \theta \frac{T}{m}$$ $$\ddot{\phi}=\dot{\theta} \dot{\psi} \frac{\left(I_{y y}-I_{z z}\right)}{I_{x x}}+\frac{\tau_{\phi}}{I_{x x}}$$ $$\ddot{\theta}=\dot{\psi} \dot{\phi} \frac{\left(I_{z z}-I_{x x}\right)}{I_{y y}}+\frac{\tau_{\theta}}{I_{y y}}$$ $$\ddot{\psi}=\dot{\phi} \dot{\theta} \frac{\left(I_{x x}-I_{y y}\right)}{I_{z z}}+\frac{\tau_{\psi}}{I_{z z}}$$

여기서 $m$과 $g$는 쿼드로터의 총 무게와 중력이며 $\phi, \theta$와 $\psi$는 각각 횡동요 운동각(rolling angle), 종동요 운동각(pitching angle)과 선수동요 운동각(yawing angle)이며 $I_{x x}, I_{y y}$와 $I_{z z}$는 각각 $x, y$와 $z$축에 대한 전체 관성 모멘트이다. $T$는 병진 운동에 대한 힘이며 $\tau_{\phi}, \tau_{\theta}$와 $\tau_{\psi}$는 회전 운동에 대한 토크이다. 경로점 추종문제를 안정화 문제로 변환하기 위한 새로운 변수 $e _{x} =x-x _{d,} ~e _{y} =y-y _{d}$와 를 정의하고 $x _{d} ,~y _{d}$와 $z _{d}$는 상수로 가정한다. 이때, 상태변수 $\chi:=\left[e_{x} \dot{e}_{x} e_{y} \dot{e}_{y} e_{z} \dot{e}_{z} \phi \dot{\phi} \theta \dot{\theta} \psi \dot{\psi}\right]^{T}, \quad T:=\frac{m\left(q+u_{1}\right)}{\cos \theta \cos \phi}$으로 치환한 제어입력 $u:=\left[\begin{array}{llll}{u_{1}} & {\tau_{\phi}} & {\tau_{\theta}} & {\tau_{\psi}}\end{array}\right]^{T}$를 이용한다. 쿼드로터 동역학 (2)는 ${\ddot{x}}$와 ${\ddot{y}}$에 대한 동역학의 부분선형화를 통해 다음과 같이 표현된다 :

(3)
$$\dot{\chi}=f(\chi)+\tilde{g} u+w$$ $$f=\left[f_{1} f_{2} f_{3} f_{4} f_{5} f_{6} f_{7} f_{8} f_{9} f_{10} f_{10} f_{11} f_{12}\right]^{T}$$ $$ \tilde{g}=\left[\begin{array}{ccccc}{0000010} & {0} & {0} & {0} & {0} & {0} \\ {0000000} & {\frac{1}{I_{x x}}} & {0} & {0} & {0} \\ {0000000} & {0} & {0} & {\frac{1}{J_{y y}}} & {0} & {0} \\ {0000000} & {0} & {0} & {0} & {0} & {\frac{1}{I_{z z}}}\end{array}\right] $$ $$ \begin{array}{l}{f_{1}=\chi_{2}, f_{2}=g \chi_{9}, f_{3}=\chi_{4}, f_{4}=-g \chi_{7}, f_{5}=\chi_{6}, f_{6}=0, f_{7}=\chi_{8}} \\ {f_{8}=\frac{I_{y y}-I_{z z}}{I_{x x}} \chi_{10} \chi_{12}, f_{9}=\chi_{10}, f_{10}=\frac{I_{z z}-I_{x x}}{I_{y y}} \chi_{8} \chi_{12}, f_{11}=\chi_{12}} \\ {f_{12}=\frac{I_{x x}-I_{y y}}{I_{z z}} \chi_{8} \chi_{10}, w=\left[w_{x} w_{y} 00000000000\right]^{T}}\end{array} $$

3. 주요 결과

문제 1 외란을 고려한 비선형 쿼드로터 동역학 (2)를 점근적으로 안정화시키는 를 설계하는 것이다.

명제 1 ((13)) 일 때, (3)은 부채꼴 비선형 기법(10)을 이용하여 모델링 오차 없이 4개의 규칙을 갖는 퍼지 모델로 표현된다 :

(4)
$\dot{\chi}=\sum_{i=1}^{4} \theta_{i}\left(\chi_{10}, \chi_{12}\right)\left(A_{i} \chi+\tilde{g} u+B_{w} w\right)$

여기서

$$ A_{i}=\left[\begin{array}{ll}{\alpha_{11}} & {\alpha_{12}} \\ {0_{6} \times 6} & {\alpha_{22 i}}\end{array}\right], B_{w}=[101000000000]^{T}, $$ $$\alpha_{11}=\left[\begin{array}{lllll}{0} & {1} & {0} & {0} & {0} & {0} \\ {0} & {0} & {0} & {0} & {0}& {0} \\ {0} & {0} & {0} & {1} & {0} & {0} \\ {0} & {0} & {0} & {0} & {0} & {0} \\ {0} & {0} & {0} & {0}& {0} & {1} \\ {0} & {0} & {0} & {0}& {0} & {0}\end{array}\right], \alpha_{12}=\left[\begin{array}{ccccc}{0} & {0} & {0} & {0} & {0} \\ {0} & {0} & {g } & {0}& {0} \\ {0} & {0} & {0}& {0} & {0} \\ {-g} & {0} & {0}& {0} & {0} \\ {0} & {0} & {0} & {0} & {0} \\ {0} & {0} & {0}& {0} & {0}\end{array}\right], $$ $$\alpha_{22 i}=\left[\begin{array}{cccc} {0} & {1} & {000} & {0} \\ {0} & {0} & {000}&{ \frac{I_{y y}-I_{z z}}{I_{x x}} \mu_{1 i}} \\ {0} & {0} & {000} & {0} \\ {0} & {0} & {000}&{ \frac{I_{z z}-I_{x x}}{I_{y y}} \mu_{2 i}} \\ {0} & {0} & {000} & {1} \\ {0}&{ \frac{I_{x x}-I_{y y}}{I_{z z}} \mu_{1 i} }&{000} & {0} \end{array}\right]$$ $$ \begin{array}{l}{\mu_{11}=\mu_{12}=\mathrm{sup}_{\chi \in \mathbb{B}_{\mathrm{r}}} \chi_{10}, \quad \mu_{13}=\mu_{14}=\mathrm{inf}_{\chi \in \mathbb{B}_{\mathrm{r}}} \chi_{10}} \\ {\mu_{21}=\mu_{23}=\mathrm{sup}_{\chi \in \mathbb{B}_{\mathrm{x}}} \chi_{12}, \quad \mu_{22}=\mu_{24}=\mathrm{inf}_{\chi \in \mathbb{B}_{\mathrm{x}}} \chi_{12}} \\ {\theta_{1}\left(\chi_{10}, \chi_{12}\right)=\Gamma_{11} \Gamma_{21}, \quad \theta_{2}\left(\chi_{10}, \chi_{12}\right)=\Gamma_{11} \Gamma_{22}} \\ {\theta_{3}\left(\chi_{10}, \chi_{12}\right)=\Gamma_{12} \Gamma_{21}, \quad \theta_{4}\left(\chi_{10}, \chi_{12}\right)=\Gamma_{12} \Gamma_{22}} \\ {\Gamma_{11}=\frac{\chi_{10}-\mu_{13}}{\mu_{11}-\mu_{13}}, \quad \Gamma_{12}=1-\Gamma_{11}} \\ {\Gamma_{21}=\frac{\chi_{12}-\mu_{22}}{\mu_{21}-\mu_{22}}, \quad \Gamma_{22}=1-\Gamma_{21}}\end{array} $$

비선형 쿼드로터 (3)과 퍼지 제어기

(5)
$u=-\sum_{i=1}^{4} \theta_{i}\left(\chi_{10}, \chi_{12}\right) K_{i} \chi$

를 고려하자. 특정 성능 함수를 최소화하기 위해 이차 성능 함수

(6)
$J=\int_{0}^{\infty}\left\{y^{T} W y+u^{T} R u\right\} d t$

를 도입하자. 여기서 $W=W^{T}>0, \quad R=R^{T}>0$이다. 이 때, 다음의 변수

$\tilde{y}=\left[\begin{array}{c}{y} \\ {u}\end{array}\right]=\sum_{i=1}^{4} \theta_{i}\left(\chi_{10}, \chi_{12}\right)\left[\begin{array}{c}{C} \\ {-K_{i}}\end{array}\right] \chi$

를 정의하여 (6)은

(7)
$J=\int_{0}^{\infty} \tilde{y}^{T}\left[\begin{array}{ll}{W} & {0} \\ {0} & {R}\end{array}\right] \tilde{y} d t$

으로 표현된다.

$H_∞$성능을 만족하며 보장 비용 제어기를 이용한 설계 유도 과정은 참고문헌 (11)에 의거하여 도출된다. 리아프노브 함수

$V=\chi^{T} P \chi$

를 고려하자, 여기서 $P=P^{T}>0$이다. 폐루프 시스템 (4), (5), 이차 성능 함수 (7)과 정의 1에 의거한 $chi$에 따른 $V$의 미분계수(derivative)는

$\left.\dot{V}\right|_{(4),(5)} \leq \dot{\chi}^{T} P \chi+\chi^{T} P \dot{\chi}+\tilde{y}^{T}\left[\begin{array}{c}{W 0} \\ {0} & {R}\end{array}\right] \tilde{y}$$=\sum_{i=1}^{4} \theta_{i}\left(\chi_{10}, \chi_{12}\right)\left[\begin{array}{c}{\chi} \\ {w}\end{array}\right]^{T}\left(\left[\begin{array}{cc}{\Omega_{i}} & {P B_{w}} \\ {B_{w}^{T} P}&{-\gamma^{2} I}\end{array}\right]\right)\left[\begin{array}{c}{\chi} \\ {w}\end{array}\right]-\chi^{T} H \chi+\gamma^{2} w^{T} w$

이다, 여기서

$\Omega_{i}=A_{i}^{T} P-\Pi_{i}^{T} \tilde{g}^{T} P+P A_{i}-P \tilde{g} K_{i}^{-}+H+\left[C^{T}-\Pi_{i}^{-T}\right]\left[\begin{array}{c}{W 0} \\ {0 R}\end{array}\right]\left[\begin{array}{c}{C} \\ {-K_{i}}\end{array}\right]$이다. $\theta_{i}\left(\chi_{10}, \chi_{12}\right) \in \mathbb{R}_{(0,1]}$ 특성을 고려하면 $\chi \in \mathbb{R}^{n} /\{0\}$ 에 대해서

(8)
$\left[\begin{array}{cc}{\Omega_{i}+\left[C^{T}-K_{i}^{T}\right]\left[\begin{array}{c}{W}&{ 0} \\ {0} & {R}\end{array}\right]\left[\begin{array}{c}{C} \\ {-K_{i}}\end{array}\right]}&{ P B_{w}} \\ {B_{w}^{T} P} & {-\gamma^{2} I}\end{array}\right] <0$

만족한다면

(9)
$\left.\dot{V}\right|_{(4),(5)} \leq-\chi^{T} H \chi+\gamma^{2} w^{T} w$

성립함을 알 수 있다. 또한, 식(9)를 부터 $t _{f}$까지 적분하면

$\left.V\right|_{(4),(5)}\left(t_{f}\right)-\left.V\right|_{(4),(5)}(0) \leq-\int_{0}^{t_{f}} \chi^{T} H \chi d t+\gamma^{2} \int_{0}^{t_{f}} w^{T} w d t$

이다. $\left.V\right|_{(4),(5)}\left(t_{f}\right)>0$과 $\chi(0)=0$을 이용하면

$\int_{0}^{t_{f}} \chi^{T} H \chi d t \leq \gamma^{2} \int_{0}^{t_{f}} w^{T} w d t$

이며 정의 1에 제시된 $w=0$과 $H>0$의 조건을 통해

$\left.\dot{V}\right|_{(4),(5)^{-}} \chi^{T} H \chi <0$

$H_∞$성능이 보장됨을 알 수 있다. 식(8)의 선형행렬부등식을 구하기 위해, $\operatorname{diag}\left[P^{-1} I\right]$기반 합동변환(congruence transformation)과 변수치환 $P^{-1}:=X, \quad K_{i} X:=M_{i}$ 와 $XHX:=S$및 슈어 보상(Schur complement)을 이용하여 다음의 선형 행렬 부등식

(10)
$\left[\begin{array}{cccc}{\bar{\Omega}_{i}} & {X C^{T}} & {-M_{i}^{T}} & {B_{w}} \\ {C X} & {-W^{-1}} & {0} & {0} \\ {-M_{i}} & {0} & {-R^{-1}} & {0} \\ {B_{w}^{T}} & {0} & {0} & {-\gamma^{2} I}\end{array}\right]<0 \quad \forall i \in \mathbb{I}_{4}$

을 유도 할 수 있다. 여기서 $\widetilde{\Omega}_{i}=\left(A_{i} X-\widetilde{g} M_{i}\right)^{T}+\left(A_{i} X-\widetilde{g} M_{i}\right)+S.$따라서 선형 행렬 부등식(10)을 만족시키는 $X=X^{T}>0$과 $M_i$가 존재하면 폐루프 시스템 (4), (5)는 전역적으로 점근적 안정하다. 이 때, 제어 이득 행렬 $K_{i}=M_{i} X^{-1}$이다.

4. 시뮬레이션

제안된 설계 기법의 검증을 위해 다음의 변수값(8)

$m=1.336 \mathrm{kg}, g=9.8$ $I_{x x}=0.0259 \mathrm{kgm}^{2}, \quad I_{y y}=0.0260 \mathrm{kgm}^{2}, \quad I_{z z}=0.0397 \mathrm{kgm}^{2}$

을 가진 쿼드로터(1)을 고려하자. 설계 목표는 외란을 고려한 쿼드로터 시스템을 안정화시키는 제어기를 설계하는 것이다. $\left|\chi_{10}\right| \leq \frac{\pi}{4}, \quad\left|\chi_{12}\right| \leq \frac{\pi}{4}$와 외란 신호 $w_{x}=w_{y}=\sin (10 t)$이며 10초에서 15초 동안 발생한다고 가정하자. 선형 행렬 부등식(10)의 가능해가 존재하며 양한정 행렬 $P$와 제어 이득 행렬 $K_{i}, \quad i \in I_{4}$을 얻을 수 있다 :

$P=\left[\begin{array}{c}{P_{11} P_{12}} \\ {P_{12}^{T} P_{22}}\end{array}\right], K_{1}=\left[K_{11} K_{12}\right], K_{2}=\left[K_{21} K_{22}\right]$ $K_{3}=\left[K_{31} K_{32}\right], K_{4}=\left[K_{41} K_{42}\right]$

여기서

$$P_{11}=\left[\begin{array}{l}{0.1389}&{ 0.0996 }&{0 .1323}&{0 .0943}&{0 .0000}&{0 .0000} \\ {0.0996}&{0 .0717}&{0 .0944}&{0 .0671}&{0 .0000}&{0 .0000} \\ {0.1323}&{0 .0944}&{0 .1385}&{0 .0992}&{0 .0000}&{0 .0000} \\ {0.0943}&{0 .0671}&{0 .0992}&{0 .0714}&{0 .0000}&{0 .0000} \\ {0.0000}&{0 .0000}&{0 .0000}&{0 .0000}&{0 .0217}&{0 .0600} \\ {0.0000}&{0 .0000}&{0 .0000}&{0 .0000}&{0 .0600}&{0 .2168}\end{array}\right],$$ $$ P_{12}=\left[\begin{array}{l}{-0.3106}&{-0.3106}&{0 .3306}&{0 .0529}&{0 .0000}&{0 .0000} \\ {-0.2208}&{-0.0351}&{0 .2387}&{0 .0383}&{0 .0000}&{0 .0000} \\ {-0.3291}&{-0.0527}&{0 .3111}&{0 .0495}&{0 .0000}&{0 .0000} \\ {-0.2374}&{-0.0381}&{0 .2210}&{0 .0351}&{0 .0000}&{0 .0000} \\ {0.0000 }&{ 0.0000}&{0 .0000}&{0 .0000}&{0 .0000}&{0 .0000} \\ {0.0000 }&{ 0.0000}&{0 .0000}&{0 .0000}&{0 .0000}&{0 .0000}\end{array}\right], $$ $$ K_{11}=\left[\begin{array}{ccc}{0.0000} & {0.0000} & {0.0000} & {0.0000} & {0.6000} & {2 .1679} \\ {-19.0710} & {-13.5365} & {-20.3282} & {-14.7021}&{0 .0000}&{0 .0000} \\ {20.3606} & {14.7367} & {19.0472} & {13.5094} & {0.0000}&{0 .0000} \\ {0.0002} & {0.0002} & {0.0002}& {0.0002} & {0.0000}&{0 .0000} \end{array}\right], $$ $$ K_{12}=\left[\begin{array}{ccccc}{0.0000} & {0.0000} & {0.0000} & {0.0000} & {0.0000} & {0.0000} \\ {49.2611} & {8.0232} & {-44.3884}&{-7.0184}&{-0.0000} & {-0.0000} \\ {-44.2586} & {-6.9915} & {49.4227} & {8.0571} & {0.00000} & {0.0000} \\ {-0.0006} & {-0.0001} & {0.0006} & {0.0001} & {0.7199} & {0.5263}\end{array}\right], $$ $$ K_{21}=\left[\begin{array}{cccc}{0.0000} & {0.0000} & {0.0000} & {0.0000} & {0.6000} & {2 .1679} \\ {-19.0716} & {-13.5369} & {-20.3287} & {-14.7025} & {0 .0000} & {0 .0000} \\ {20.3612} & {14.7371} & {19.0478} & {13.5097} & {0.0000} & {0 .0000} \\ {0.0000} & {0.0000} & {-0.0000} & {0.0000} & {0 .0000}& {0 .0000}\end{array}\right], $$ $$ K_{22}=\left[\begin{array}{cccccc}{0.0000} & {0.0000} & {0.0000} & {0.0000}& {0.0000} & {0.0000} \\ {49.2624} & {8.0234} & {-44.3898} & {-7.0187} & {-0.0000} & {-0.0000} \\ {-44.2599} & {-6.9917} & {49.4240} & {8.0573} & {-0.0000} & {-0.0000} \\ {0.0000} & {0.0000} & {0.0000} & {0.0000} & {0.7199} & {0.5263}\end{array}\right], $$ $$ K_{31}=\left[\begin{array}{ccc}{0.0000} & {0.0000}& {0.0000} & {0.0000} & {0.6000} & {2 .1679} \\ {-19.0716} & {-13.5369} & {-20.3287} & {-14.7025} & {0 .0000} & {0 .0000} \\ {20.3612} & {14.7371} & {19.0478} & {13.5097} & {0.0000} & {0.0000}\\ {-0.0000} & {-0.0000} & {0.0000}& {-0.0000} & {0.0000}& {0.0000}\end{array}\right], $$ $$ K_{32}=\left[\begin{array}{cccc}{0.0000} & {0.0000} & {0.0000} & {0.0000} & {0.0000} & {0 .0000} \\ {49.2624} & {8.0234} & {-44.3898} & {-7.0187} & {0 .0000} & {0 .0000} \\ {-44.2599} & {-6.9917} & {49.4240} & {8.0573} & {0.0000} & {0 .0000} \\ {-0.0000} & {-0.0000} & {-0.0000} & {-0.0000} & {-0.7199} & {0 .5263}\end{array}\right], $$ $$ K_{41}=\left[\begin{array}{cccc}{0.0000} & {0.0000} & {0.0000} & {0.0000} & {0.6000} & {2 .1679} \\ {-19.0710} & {-13.5365} & {-20.3282} & {-14.7021} & {0 .0000} & {0 .0000} \\ {20.3606} & {14.7367} & {19.0472} & {13.5094} & {0.0000} & {0 .0000} \\ {-0.0002} & {-0.0002} & {-0.0002} & {-0.0002}& {0.0000} & {0 .0000}\end{array}\right], $$ $$ K_{42}=\left[\begin{array}{cccccc}{0.0000} & {0.0000}& {0.0000} & {0.0000} & {0.0000} & {0.0000} \\ {49.2611} & {8.0232} & {-44.3884} & {-7.0184} & {0.0000} & {0.0000} \\ {-44.2586} & {-6.9915} & {49.4227} & {8.0571} & {-0.0000} & {-0.0000} \\ {0.0006} & {0.0001} & {-0.0006} & {-0.0001} & {0.7199} & {0.5263}\end{array}\right]. $$

그림 2-11은 명제 1에 제시된 퍼지 모델 (4)를 이용하여 설계된 제어기 (5)에 대한 실제 비선형 쿼드로터 (1)의 시뮬레이션 결과를 보여준다. 그림 1-6은 쿼드로터의 횡방향 운동($x$), 종방향 운동($y$), 고도($z$), 종동요 운동각($\pi$), 횡동요 운동각($theta$)와 선수동요 운동각($psi$)의 시간응답을 보여준다. 외란을 고려한 제어기 (5)에 의해 비선형 쿼드로터 (1)의 궤적은 시간이 지남에 따라 $x_{d}=5[\mathrm{m}], y_{d}=5[\mathrm{m}]$와 ${z_{d}}=5[\mathrm{m}]$으로 수렴함을 알 수 있다. 결론적으로, 참고문헌 (12)의 제어기 설계 조건을 적용하여 실제 쿼드로터 (1)의 점근적 안정도가 보장됨을 시뮬레이션 결과를 통해 증명하였다.

그림 2 $x$의 시간응답

Fig. 2 Time response of $x$

../../Resources/kiee/KIEE.2019.68.11.1396/fig2.png

그림 $y$3 의 시간응답

Fig. 3 Time response of $y$

../../Resources/kiee/KIEE.2019.68.11.1396/fig3.png

그림 4 $z$의 시간응답

Fig. 4 Time response of $z$

../../Resources/kiee/KIEE.2019.68.11.1396/fig4.png

그림 5 $\pi$의 시간응답

Fig. 5 Time response of $\pi$

../../Resources/kiee/KIEE.2019.68.11.1396/fig5.png

그림 6 $\theta$의 시간응답

Fig. 6 Time response of $\theta$

../../Resources/kiee/KIEE.2019.68.11.1396/fig6.png

그림 7 $\psi$의 시간응답

Fig. 7 Time response of $\psi$

../../Resources/kiee/KIEE.2019.68.11.1396/fig7.png

그림 8 제어입력 $u_1$

Fig. 8 Control input $u_1$

../../Resources/kiee/KIEE.2019.68.11.1396/fig8.png

그림 9 제어입력 $u_2$

Fig. 9 Control input $u_2$

../../Resources/kiee/KIEE.2019.68.11.1396/fig9.png

그림 10 제어입력 $u_3$

Fig. 10 Control input $u_3$

../../Resources/kiee/KIEE.2019.68.11.1396/fig10.png

그림 11 제어입력 $u_4$

Fig. 11 Control input $u_4$

../../Resources/kiee/KIEE.2019.68.11.1396/fig11.png

Acknowledgements

이 논문은 2017년도 한밭대학교 교내학술연구비의 지원을 받았음.

This paper was supported by Hanbat National University Research Fund in 2017.

References

1 
H. Lee, H. J. Kim, 2014, Robust control of a quadrotor using Takagi-Sugeno fuzzy model and an LMI approach, in Proc. of International Conference on Control, Vol. automation and systems, No. , pp. 370-374DOI
2 
C. Liu, J. Pan, Y. Chang, 2016, PID and LQR Trajectory Tracking Control for An Unmanned Quadrotor Helicopter: Experimental Studies, in Proc. of the Chinese Control Conference, Vol. , No. , pp. 5566-5571DOI
3 
A. Mokhtari, A. Benallegue, and Y. Orlov, 2006, Exact Linearization and Sliding-Mode Observer for A Quadrotor Unmanned Aerial Vehicle, International Journal Robotics and Automation, Vol. 21, No. 1, pp. 39-49DOI
4 
D. Lee, H. J. Kim, S. Sastry, Jun 2009, Feedback linearization vs. Adaptive Sliding-Mode Control for A Quadrotor Helicopter, International Journal of Control Automation Systems, Vol. 7, No. 3, pp. 419-428DOI
5 
X. J. Chen, D. Li, Y. Bai, 2011, Modelling and Nero-Fuzzy Adaptive Control for Eight-Rotor MAV, International Journal of Control Automation Systems, Vol. 9, No. 6, pp. 1154-1163Google Search
6 
T. Lee, 2013, Robust Adaptive Attitude Tracking on SO(3) With an Application to a Quadrotor UAV, IEEE Transactions on control systems technology, Vol. 21, No. 5, pp. 1924-1930DOI
7 
D. W. Kim, H. J. Lee, M. Tomizuka, 2010, Fuzzy Stabilization of Nonlinear Systems under Sampled-Data Feedback: An Exact Discrete-Time Model Approach, IEEE Transactions on Fuzzy Systems, Vol. 18, No. 2, pp. 251-260DOI
8 
Y. C. Choi, H. S. Ahn, 2015, Nonlinear Control of Quadrotor Control for Point Tracking : Actual Implementation and Experimental Tests, IEEE/ASME Transactions on Mechatronics, Vol. 20, No. 3, pp. 1179-1192DOI
9 
X. Liu, Q. Zhang, 2003, Approaches to Quadratic Stability Conditions and $H_∞$ Control Designs for T-S Fuzzy Systems, IEEE Transactions on Fuzzy Systems, Vol. 11, No. 6, pp. 830-839DOI
10 
K. Tanaka, H. O. Wang, 2004, Fuzzy control systems design and analysis: a linear matrix inequality approach, John Wiley & SonsDOI
11 
H. N. Wu, S. Feng, Z. Y. Liu, et al., 2016, Disturbance Observer Based Robust Mixed $H_2$/$H_∞$ Fuzzy Tracking Control for Hypersonic Vehicles, Fuzzy Sets and Systems, Vol. 30, No. 6, pp. 118-136DOI
12 
J. Dong, Y. Wang, G. Yang, 2009, Control Synthesis of Continuous-Time T-S Fuzzy Systems with Local Nonlinear Models, IEEE Transactions on Systems, Man, and Cybernetics-Part B : Cybernetics, Vol. 39, No. 5, pp. 1245-1258DOI
13 
Y. M. Choi, D. W. Kim, 2017, A T-S Fuzzy Model Approach to Guaranteed Cost Control of Quadrotor, Journal of Institute of Control, Robotics and Systems, Vol. 23, No. 7, pp. 559-565Google Search
14 
D. W. Kim, 2019, Fuzzy model-based control of a quadrotor, Fuzzy Sets and Systems, Vol. 371, No. 15, pp. 136-147Google Search
15 
H. S. Kim, J. B. Park, Y. H. Joo, 2016, A systematic approach to fuzzy-model-based robust $H_∞$ control design for a quadrotor UAV under imperfect premise matching, International Journal of Fuzzy Systems, Vol. 19, No. 4DOI
16 
Z. Xu, X. Nian, H. Wang, Y. Chen, 2017, Robust guaranteed cost tracking control of quadrotor UAV with uncertainties, ISA Transactions, Vol. 69, pp. 157-165DOI

저자소개

이경학 (Gyeong Hak Lee)
../../Resources/kiee/KIEE.2019.68.11.1396/au1.png

Gyeong Hak Lee received B.S. and M.S. degrees in Electrical Engineering from Hanbat National University, South Korea, in 2016 and 2019, respectively.

He is currently working toward the Ph.D. degree in Electrical Engineering from Hanbat National University, South Korea.

His research interests include T-S fuzzy model, fuzzy modeling and underwater glider.

E-mail : mi101183@hanbat.ac.kr

김도완 (Do Wan Kim)
../../Resources/kiee/KIEE.2019.68.11.1396/au2.png

Do Wan Kim received the B.S., M.S., and Ph.D. degrees from the Department of Electrical and Electronic Engineering, Yonsei University, Seoul, Korea, in 2002, 2004, and 2007, respectively.

He was with the Engineering Research Institute, Yonsei University, and a Visiting Scholar with the Department of Mechanical Engineering, University of California, Berkeley.

Since 2009, he has been a Research Professor with the Department of Electrical and Electronic Engineering, Yonsei University.

Since 2010, he has been with the School of Electrical Engineering, Hanbat National Unversity, South Korea.

His current research interests include discrete-time, sampled-data, and digital nonlinear control systems, linear and nonlinear systems with nonlinear perturbations, fuzzy systems, and digital redesign.

E-mail : dowankim@hanbat.ac.kr.kr