Received July 14, 2020, accepted July 26, 2020, date of publication August 10, 2020, date of current

version August 21, 2020.

Digital Object Identifier 10.1109/ACCESS.2020.3015589

Stability Analysis and T-S Fuzzy Dynamic

Positioning Controller Design for Autonomous
Surface Vehicles Based on Sampled-Data Control
MINJIE ZHENG , SHENHUA YANG, AND LINA LI
Navigation College, Jimei University, Xiamen 361021, China
Corresponding author: Shenhua Yang ([email protected])
This work was supported in part by the National Natural Science Foundation of China under Grant 51579114 and Grant 51879119; and in
part by the Natural Science Foundation of Fujian Province under Grant 2018J01485, Grant 2018J01536, and Grant 2018J01484.

ABSTRACT This article investigates the stability analysis and the fuzzy dynamic positioning (DP) controller
design for an autonomous surface vehicle (ASV) based on sampled-data. Firstly, the Takagi-Sugeno (T-
S) fuzzy model for the ASV with dynamic positioning system (DPS) is established. Then the criteria
for asymptotically stability analysis and controller synthesis are provided by means of linear matrix
inequalities. And less conservative results can be obtained by introducing convex reciprocal inequalities.
Finally, simulation result is shown that the fuzzy sampled-data controller is effective to guarantee that the
states of the ASV are stable and have good DP performance under the external disturbance.

INDEX TERMS Autonomous surface vehicles, dynamic positioning system, T-S fuzzy model, sampled-data
control.

I. INTRODUCTION underactuated ASV with ocean disturbance and uncertainties

Autonomous surface vehicle (ASV) is a flexible surface is discussed by using the control method of dynamic surface.
robot carrier platform with independent navigation. Gen- In [10], the method about dynamic output feedback controller
erally speaking, ASV has a rigid structure, and its shape for an ASV based on network model is discussed, which
is very small such as wave piercing boats, sliding boats, considers the packet dropouts between sampler and control
hydrofoil boats, etc. Due to its independent navigation abil- station. In [11], based on the finite time control (HFC) frame-
ity, flexibility, and high efficiency in accomplishing various work, the trajectory tracking control for ASV with unknown
complex tasks, ASV has very important application value variables is studied.
in civil, military and coastal construction, such as marine Fuzzy model is an effective method for realizing the
resource exploration, coastal defense, hydrographic research, modeling and control of nonlinear systems, such as T-S
etc., especially in some fields like high risk and beyond the fuzzy model, and sector nonlinear and local approximation
physiological limit of human body, the ASV has irreplaceable is usually used in the fuzzy model [12], [13]. So far, many
value. Recently, ASV has become a hot research area, and a scholars have focused on the fuzzy ASV system [14], [15].
large number of literature has been reported for ASV, which In [16], based on the fuzzy theory, a fuzzy learning controller
includes heading control [1], [2]; roll stabilization [4]; moor- is proposed for ASV, where the proposed controller has
ing control [3]; fault tolerated [5]; and path following [6]; obvious advantages than the traditional controller in reducing
tracking control [7]; neural network control [8] and so on. For energy and external interference; in [17], combined with
instance, [8] deal with the issue about neural network control the adaptive control and modern theory, a fuzzy adaptive
for the ASV in leader–follower formation with unknown local controller is proposed, which can overcome the model error
and leader dynamics. In [9], the formation control issue for and suppress the external interference at the same time.
In [18], based on ontology knowledge, a fuzzy support agent
The associate editor coordinating the review of this manuscript and is proposed to provide control model for ASV system. In [19],
approving it for publication was Shun-Feng Su . a T-S fuzzy DP controller is designed for an ASV in network

This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
VOLUME 8, 2020 148193
 M. Zheng et al.: Stability Analysis and T-S Fuzzy DP Controller Design for ASVs Based on Sampled-Data Control

environments which considered the network-induced delay.

In [20], by constructing a fuzzy control law, the problem
about distributed maneuvering for multiple ASVS with state
constraint and unknown kinetics is discussed. In [7], by com-
bining with Nussbaum method and the adaptive fuzzy back-
stepping technology, an adaptive fuzzy trajectory tracking
control scheme for ASV is proposed.
Dynamic positioning ship (DPS) is a marine vessel with
a computer-control system, which used the dynamic posi-
tioning (DP) system to maintain their horizontal position and
heading, and it has been used for different kinds of vessels like
offshore support vessel, cruise vessel, shuttle tankers and so
on. (see [21]). Compared with the conventional vessel which
use anchor moored to keep position, the DPS has higher FIGURE 1. Body-fixed coordinate systems.
positioning accuracy and flexibility. Thus, the control prob-
lem for DPS has attracted considerable attention recently. the designed method of sampled-data controller is introduced
In [22], the robust controller for the vessels with DPS is to provide good DP performance under the external distur-
designed based on mixed sensitivity and uncertain model. bances. Finally, a simulation of an ASV is conducted to
The issue about optimal H∞ control for a DPS is focused illustrate the effectiveness of the proposed methods.
on by [23] with T-S fuzzy models. In [24], the issue about This note is organized as follow. The problem formulation
input saturation and robust nonlinear control for the DPS is introduced in section 2. Section 3 introduces the stability
under extern unknown disturbance is discussed. In [25], the analysis and controller design for the system. A simulation
problem of thrust distribution for DPS with cycloidal pro- result is given in section 4 to validate the effectiveness of the
peller is discussed, and the sequential quadratic programming proposed algorithm.
algorithm is introduced for the thrust distribution. In [26],
a new adaptive constraint control algorithm is introduced for II. PROBLEM FORMULATION
DPS with input amplitude and rate saturation. By adding mass The motion equations of the ASV, which is equipped with
terms, a new robust controller for DP ships has been designed thruster and dynamic positioning system, are considered as
in [27]. follow.
Recently, sampled-data system has become an impor-
tant topic, because modern control systems widely used M v̇(t) + Dυ(t) + Gη(t) = u(t) + w(t) (1)
the digital computers to control continuous-time systems
(see [28]–[30]). Until now, the system has attracted many where
attention of scholars, and considerable results have been m11

0 0
 
n11 0 0

reported for that system (see [31]–[39]). The ASV with
M = 0 m22 m23  , D =  0 n22 n23  ,
DPS is also a sampled-data system, where a variety of sen-
0 m32 m33 0 n32 n33
sors is used by digital computers to obtain several kinds of
sensor information, such as higher accuracy position, head- 
g11 0 0

ing, velocity and so on. However, few literatures have been G= 0 g22 0 ,
reported for nonlinear sampled-data ASV with DPS based 0 0 g33
on T-S fuzzy model. Since the sampled-data ASV with DPS
is a complex nonlinear system. So, how to propose a T-S where η(t) = xa (t) ya (t) ψ(t)
 T
represents the positions
fuzzy model and control approach to improve the DP perfor- xp (t), yp (t) and yaw angle ψ(t) which is based on the earth-
mance is a challenging work. Besides, how to improve the T
fixed frame. υ(t) = p(t) q(t) r(t) represents the body-

Lyapunov-Krasovskii functional (LKF) by adding novel item
fixed velocity, where p(t), q(t), r(t) represents surge velocity,
or introducing novel approach such as reciprocally convex
sway velocity and the yaw velocity, respectively.
combination approach to get less conservative results, which
u(t) represents the control input; w (t) represents the
are the motivations of the paper.
exogenous disturbances like wave, wind, ocean currents
In this brief, the issue about sampled-data control for
which belongs to L2 [0, ∞); M , D represent the inertia and
fuzzy ASV with DPS is investigated. The T-S fuzzy model
linear damping matrix respectively, G represent mooring
for the system is established firstly. Then the lower bound
forces, and
and upper bound of the network-induced delays between
ASV and control center is both taken into consideration. η̇(t) = J (ψ)υ(t) (2)
In terms of LMI approach, Lyapunov theorems is involved
for the stability analysis. And less conservative result can be where J (ψ) represents the transformation matrix between
obtained by introducing convex reciprocal inequalities. Then, the two coordinates mentioned above, which is presented

148194 VOLUME 8, 2020

M. Zheng et al.: Stability Analysis and T-S Fuzzy DP Controller Design for ASVs Based on Sampled-Data Control

in Fig. 1, and Model Rule 2:

π
IF ψ(t) is about 2, THEN
cos (ψ(t)) − sin (ψ(t))
 
0
J (ψ(t)) =  sin (ψ(t)) cos (ψ(t)) 0 ẋ(t) = A2 x(t) + B2 u(t) + Bw2 w(t),
0 0 1 y(t) = C2 x(t). (7)
Define Model Rule 3:
IF ψ(t) is about − π2 , THEN
T
x(t) = η(t) υ(t)

T
ψ(t)

= xp (t) yp (t) p(t) q(t) r(t) ẋ(t) = A3 x(t) + B3 u(t) + Bw3 w(t),
Let y(t) = C3 x(t). (8)
where
 
a11 0 0
−M −1 D =  0 a23  ,
 
a22 0 0 0 1 −α 0
0 a32 a33  0
 0 0 α 1 0 

d11 0 0
  0 0 0 0 0 1 
A1 =  ,
M −1 =  0 d22 d23  ,  b11
 0 0 a11 0 0 
0 d32 d33  0 b22 b23 0 a22 a23 

b11 0 0
 0 b32 b33 0 a32 a33
β
 
−M −1 G =  0 b22 b23  . (3) 0 0 0 −1 0
0 b32 b33  0
 0 0 1 β 0 
 0 0 0 0 0 1 
Substituting (2) and (3) into (1) yields A2 =  ,
 b11
 0 0 a11 0 0 
ẋ(t) = Ax(t) + Bu(t) + Bw w(t), (4)  0 b22 b23 0 a22 a23 
0 b32 b33 0 a32 a33
where
β
 
0 0 0 1 0
0 0 0 cos (ψ(t)) − sin (ψ(t)) 0 β
 
 0 0 0 −1 0 
sin (ψ(t)) cos (ψ(t))
 
 0 0 0 0  0 0 0 0 0 1 .


 0 0 0 0 0 1
 A3 = 
,
  b11 0 0 a11 0 0 
A=
 b11 0 0 a11 0 0

 0



 b22 b23 0 a22 a23 
 0 b22 b23 0 a22 a23  0 b32 b33 0 a32 a33
 
0 b32 b33 0 a32 a33 0 0 0
   0 0 0 
0 0 0 
 0

0 0 ,
 0 0 0  Bi = Bwi = i = 1, 2, 3.

 0
  d11 0 0 
0 0 .

 0

B = Bw =  d22 d23 
 d11 0 0 

 0
 0 d32 d33
d22 d23 
0 d32 d33 The overall fuzzy model inferred from (6)-(8) is repre-
sented as:
Then the T-S fuzzy model for system (4) can be described
3
by following rules. X
ẋ(t) = µi (z(t)) [Ai x(t) + Bi u(t) + Bwi w(t)],
Mode Rule i: IF z1 (t) is fi1 , · · · zn (t) is fin , THEN
i=1
ẋ(t) = Ai x(t) + Bi u(t) + Bwi w(t), i = 1, 2, . . . , n X3

y(t) = Ci x(t), (5) y(t) = µi (z(t))Ci x(t), (9)

i=1
where fij denote the fuzzy set, n denotes the number of rule, where
z1 (t), z2 (t), . . . , zn (t) are premise variables. y(t) is controlled
ωi (z(t))
output, and Ci are known matrices. µi (z(t)) = ≥ 0, i = 1, 2, 3,
3
that yaw angle ψ(t) is varying between −π 2

Assumed P
ωi (z(t))
and π 2, so the exact T-S fuzzy model is obtained by the

i=1
n
following three rules. Y
ωi (z(t)) = fij (zj (t)),
Model Rule 1:
j=1
IF ψ(t) is about 0, THEN 3
X
ẋ(t) = A1 x(t) + B1 u(t) + Bw1 w(t), µi (z(t)) = 1,
i=1
y(t) = C1 x(t). (6) z(t) = [z1 (t), z2 (t), . . . , zn (t)] ,

VOLUME 8, 2020 148195

 M. Zheng et al.: Stability Analysis and T-S Fuzzy DP Controller Design for ASVs Based on Sampled-Data Control

where τk represents the sampler-to-controller network-

induced delay. K1 , K2 , K3 are state-feedback gain matrices,
then the overall fuzzy controllers are described as
3
X
u(t) = µj (z(t))Kj x(tk + τk ),
j=1
tk ≤ t < tk+1 , k = 0, 1, 2, . . . (14)
Substituting (14) into Eq. (9), we obtain
3 X
X 3
ẋ(t) = µi (z(t))µj (z(t)) [Ai x(t)
i=1 j=1
FIGURE 2. Membership function of ψ(t ). + Bi Kj x(tk + τk ) + Bwi w(t)

(15)
Define
τ (t) = t − tk − τk , (16)
where the time-varying delay τ (t) is piecewise-linear satisfy-
ing τ (t) ∈ [τm , τM ).Thus, the sampled-data system in (15) is
converted to a system with time-varying delay as follow:
3 X 3

X
µi (z(t))µj (z(t)) [Ai x(t)




 ẋ(t) =


 i=1 j=1
+ Bi Kj x(t − τ (t)) + Bwi w(t) , t > 0

(17)
3 X 3


 X
µi (z(t))µj (z(t))Ci x(t)

 y(t) =



i=1 j=1
The paper’s object is finding sampled-data controllers
FIGURE 3. The framework of sampled-data ASV scheme.
K1 , K2 , K3 satisfying:
1) The system (17) with w(t) = 0 is asymptotically stable;
with fij (zj (t)) representing the membership grade of zj (t). The 2) To reject the varying environment disturbances such
membership function of ψ(t) is shown in Fig.2. as wave, current and wind, it is required that ||y(t)||2 ≤
The state variables of ASV are assumed to be measured γ ||w(t)||2 for all non-zero w(t) ∈ L2 [0, ∞) at zero initial
in 0 = t0 < t1 < t2 < . . . < tk < . . ., where tk , k = condition, where γ > 0.
1, 2, . . . is sampling instant. The framework of sampled-data Remark 1: Compared with the references which use
DPS scheme is descripted in Fig.3. continuous-time control method for ASV, the sampled-data
The sampling period is assumed that control and network-induced delays are both included in the
controller (14), which has more significance. If we ignore the
τm ≤ tk+1 − tk ≤ τM , ∀k ≥ 0, d > 0, (10) sampled-data control and network-induced delays, the con-
It can be seen from Fig.3. that the AVS is connected with troller is similar with the one in [40]. So, the controller in
the control center by communication networks. Thus, there [40] is a special case of (14).
exist network-induced delays between AVS and the control Remark 2: It is noted that the lower and upper bounds of
center. So, we consider the fuzzy sampled-data control law the network-induced delays are both considered in the paper,
as follow, and it has more significance than [41], where the lower bound
Controller Rule 1: of the time-varying delay is assumed to be zero. Actually, the
IF ψ(t) is about 0, THEN delay is often varying in a range, so the lower bound is not
restricted to be zero and it will lead to conservatism.
u(t) = K1 x(tk + τk ). (11) To get the main results, the following lemma is shown.
Controller Rule 2: Lemma1 ([42]) For given matrix M ∈ Rn×n > 0, scalars
IF ψ(t) is about π r1 and r2 , and a differentiable function ω ∈ Rn , then the
2, THEN
following inequality holds:
u(t) = K2 x(tk + τk ). (12) Z r2  Z r2 
ω (s)ds M
T
ω(s)ds
Controller Rule 3: r1 r1
IF ψ(t) is about − π2 , THEN
Z r2
≤ (r2 − r1 ) ωT (s)M ω(s)ds (18)
u(t) = K3 x(tk + τk ). (13) r1

148196 VOLUME 8, 2020

M. Zheng et al.: Stability Analysis and T-S Fuzzy DP Controller Design for ASVs Based on Sampled-Data Control

III. MAIN RESULTS V̇4 (t) = (τM − τm )2 ẋ(t)T R3 ẋ(t)

Z t−τm
In this section, we firstly exhibit the sufficient stability criteria
of network-based ASV with DPS by establishing Lyapunov- − (τM − τm ) ẋ T (s)R3 ẋ(s)ds (22)
t−τM
Krasovskii functional. Then the sampled-data controller is
designed to analyse the stability criteria. From Lemma1, we have
Theorem 1: For constant delay τM ≥ τm > 0,, the system
(17) is asymptotically stable with H∞ norm bound γ , if there
exist positive matrices P, S, Q1 , Q2 , R1 , R2 , R3 such that V̇3 (t) ≤ τm2 ẋ(t)T Z1 x(t) + τM
2
ẋ(t)T R2 ẋ(t)

R3 S
 − (x(t) − x(t − τm ))T R1 (x(t) − x(t − τm ))
> 0, (19) − (x(t) − x(t − τM ))T R2 (x(t) − x(t − τM ))
∗ R3
 ij
011 PT Bi Kj R1 R2 PT Bwi ATi
 V̇4 (t) ≤ (τM − τm )2 ẋ(t)T R3 ẋ(t)
Z t−τm
 ∗
 022 023 024 0 KjT BTi 
 − (τM − τm ) ẋ T (s)R3 ẋ(s)ds
9ij = 
 ∗ ∗ 033 T 0 0   t−τ (t)
 ∗ ∗ ∗ 044 0 0  t−τ (t)
Z 
+ ẋ T (s)R3 ẋ(s)ds
 
 ∗ ∗ ∗ ∗ −γ 2 I BTwi  t−τM
∗ ∗ ∗ ∗ ∗ 066
≤ (τM − τm )2 ẋ(t)T R3 ẋ(t)
<0 (20) 
τM − τm
− (x(t − τ (t)) − x(t − τM ))T
where τM − τ (t)
ij
011 = PT Ai + ATi P + Q1 + Q2 − R1 − R2 + CiT Ci × R3 (x(t − τ (t)) − x(t − τM ))
τM − τm
022 = −2R3 + S + S T + (x(t − τm ) − x(t − τ (t))T
τ (t) − τm
023 = R3 − S × R3 (x(t − τm ) − x(t − τ (t))
024 = R3 − S = (τM − τm )2 ẋ(t)T R3 ẋ(t)
033 = −Q1 − R1 − R3
T
d2 − τ (t)
 r
(x(t − τm ) − x(t − τ (t))) 
044 = −Q2 − R2 − R3 
rτ (t) − τm
−
τ (t) − τm

066 = −R−1 (x(t − τ (t)) − x(t − τM ))
 
−
R = τm2 R1 + τM
2
R2 + (τM − τm )2 R3 τM − τ (t)
 
R3 S
Proof. The Lyapunov-krasovskii function is chosen as ×
∗ R3
4
τM − τ (t)
 r 
τ τ
X
V (t) = Vi (t), t ∈ [tk , tk+1 ) x(t − m ) − x(t − (t))
r τ (t) − τm
 
−
i=1
τ (t) − τm

(x(t − τ (t)) − x(t − τM ))
 
V1 (t) = x(t)T Px(t) −
Z t Z t τM − τ (t)
T T 
x(s)T Q2 x(s)ds x(t − τm ) − x(t − τ (t))
 
V2 (t) = x(s) Q1 x(s)ds + R3 S
−
t−τm t−τM
x(t − τ (t)) − x(t − τM ) ∗ R3
Z 0 Z t
V3 (t) = τm x(t − τm ) − x(t − τ (t))
 
ẋ T (s)R1 ẋ(s)dsdθ
−
−τm t+θ x(t − τ (t)) − x(t − τM )
Z 0 Z t
+ τM ẋ T (s)R2 ẋ(s)dsdθ ≤ (τM − τm )2 ẋ(t)T R3 ẋ(t)
T 
−τM t+θ
x(t − τm ) − x(t − τ (t))
 
Z −τm Z t R3 S
−
V4 (t) = (τM − τm ) ẋ T (s)R3 ẋ(s)dsdθ (21) x(t − τ (t)) − x(t − τM ) ∗ R3
x(t − τm ) − x(t − τ (t))
 
−τM t+θ
− (23)
Calculating the derivative of V (t), it can be obtained that: x(t − τ (t)) − x(t − τM )
V̇1 (t) = 2x(t)T PT ẋ(t)
V̇2 (t) = x(t)T Q1 x(t) − x(t − τm )T Q1 x(t − τm ) + Substituting (23) into (22) that
+ x(t)T Q2 x(t) − x(t − τM )T Q2 x(t − τM ) 3 X
3
V̇3 (t) = τm2 ẋ(t)T R1 x(t) + τM 2
ẋ(t)T R2 ẋ(t)
X
V̇ (t) ≤ µi (z(t))µj (z(t))ς T (t) 4ij
Z t
i=1 j=1
− τm ẋ T (s)R1 ẋ(s)ds T  2
τm R1 + τM2

t−τm
Z t + Ai 0 0
Bi Kj R2
− τM ẋ T (s)R2 ẋ(s)ds

+ (τM − τm )2 R3 Ai Bi Kj 0 0 ς(t)


(24)
t−τM

VOLUME 8, 2020 148197

 M. Zheng et al.: Stability Analysis and T-S Fuzzy DP Controller Design for ASVs Based on Sampled-Data Control

where
ij
0̃11
 
PT Bi Kj R1 R2
 ∗ 022 423 424 
4ij =  
 ∗ ∗ 033 S 
∗ ∗ ∗ 044
ij
0̃11 = PT Ai + ATi P + Q1 + Q2 − R1 − R2
T
ζ (t) = x T (t) x T (t − τ (t)) x T (t − τm ) x T (t − τM )


By Schur complement, (20) guarantee:

T  2
4ij + Ai Bi Kj 0 0 τm R1 + τM
2

R2

+ (τM − τm )2 R3 Ai Bi Kj 0 0 < 0

(25)

Therefore, there exists a small enough scalar σ > 0

to guarantee V̇ (t) < −σ kx(t)k2 when x(t) 6 = 0. Hence,
the system (10) with w(t) = 0 is asymptotically stable. FIGURE 4. Responses of the x, y positions of the ASV without extern
Next, we will prove that the system (17) has robust stability disturbance.
and satisfies the H∞ performance γ under external distur-
bances. Then, under zero initial conditions, for all nonzero
w(t) ∈ L2 [0, ∞), the H∞ performance is considered:
Z ∞h i
Jzw = yT (s)y(s) − γ 2 wT (s)w(s) ds, γ > 0 (26)
0
then
yT (t)y(t) − γ 2 wT (t)w(t) + V̇ (t)
3 X
X 3
≤ µi (z(t))µj (z(t))ζ T (t) 2ij
i=1 j=1
T 
τm2 R1

+ Ai Bi Kj 0 0 Bwi

+ τM
2
R2 + (τM − τm )2 R3

× Ai Bi Kj 0 0 Bwi ζ (t)
 

(27) FIGURE 5. Responses of the yaw angle of the ASV without extern
disturbance.
where
ζ (t) = x T (t) x T (t − τ (t)) x T (t − τm )

T Then, under zero initial conditions, we can obtain V (0) = 0
× x T (t − τM ) wT (t) and V (∞) ≥ 0.According to (29), for all nonzero w(t) ∈
011 PT Bi Kj L2 [0, ∞) , ||y(t)||2 ≤ γ ||w(t)||2 is obtained, which means
 
R1 R2 PBwi
 ∗ 0 22 0 23 024 0  that Jzw ≤ 0◦ The proof is completed.
2ij =  033 Remark 3 Compared with [41], if τm = 0, Q1 = Q2 =
 
 ∗ ∗ S 0 
044

 ∗ ∗ ∗ 0  0, R2 = R3 = 0, then LKF is similar with the one in [41].
∗ ∗ ∗ ∗ −γ 2 I therefore, the LKF in [41] is a special case of (9). So, the result
in Theorem 1 has less conservatism than [41].
By Schur complement, (20) guarantee
Remark 4 Compared with [43] and [44], which
T
2ij + Ai Bi Kj 0 0 Bwi

use Jensen inequalityR tk directly to estimate the integral
  term − (τM − τm ) t−τ ẋ T (s)R ẋ(s)ds, and we use a substep
3
× τm2 R1 + τM
2
R2 + (τM − τm )2 R3 M
to make the results less conservative which called reciprocally
× Ai Bi Kj 0 0 Bwi < 0
 
(28) convex combination inequalities. Therefore, the conserva-
tiveness of stability criterion is greatly reduced.
From (28), we have: Furthermore, the sampled-data controller (14) will be pro-
Z ∞  posed to stabilize the system (17) based on the following
V (∞) + V (0) + ky(t)k22 − γ 2 kw(t)k22 dt ≤ 0 (29)
0
Theorems.

148198 VOLUME 8, 2020

M. Zheng et al.: Stability Analysis and T-S Fuzzy DP Controller Design for ASVs Based on Sampled-Data Control

FIGURE 6. Responses of the surge, sway and yaw angle of velocity for the FIGURE 8. Responses of the yaw angle of the ASV with extern
ASV without extern disturbance. disturbance.

FIGURE 9. Responses of the surge, sway and yaw angle of velocity for the
ASV with extern disturbance.

FIGURE 7. Responses of the x, y positions of the ASV with extern

disturbance. 0̄33 = −Q̄1 − R̄1 − R̄3
0̄44 = −R̄2 − R̄3
Theorem 2: For scales τM ≥ τm > 0, , γ > 0,if there exist 0̄66 = −P̄R̄−1 P̄
positive matrices P̄, S̄, Q̄1 , Q̄2 , R̄1 , R̄2 , R̄3 , such that: R̄ = τm2 R̄1 + τM
2
R̄2 + (τM − τm )2 R̄3
 
R̄3 S̄ Then, the system (17) is asymptotically stable with H∞ per-
∗ R̄3 formance γ . Moreover, the gain of the controller Kj is given
> 0, (30) that
 ij
0̄11 Bi K̄j R̄1 R̄2 P̄ATi P̄CiT

Bwi
 ∗ 0̄22 0̄23 0̄24 Kj = K̄j P̄−1 (32)
 0 K̄jT B̄Ti 0  
 ∗ ∗ 0̄33 S̄ 0 0 0  Proof: By noticing that −P̄R̄−1 P̄ ≤ R̄ − 2P̄, let
9̄ij = 
 
 ∗ ∗ ∗ 0̄44 0 0 0   ξ = diag P−T , P−T , P−T , P−T , I , I , I . Defining P̄ =
 ∗
 ∗ ∗ ∗ −γ 2 I P̄BTwi 0   P−1 , K̄j = Kj P−1 , R̄ = P−T RP−1 , S̄ = P−T SP−1 ,
 ∗ ∗ ∗ ∗ ∗ 0̄
66 0  Q̄1 = P−T Q1 P−1 , Q̄2 = P−T Q2 P−1 , R̄i = P−T Ri P−1 , i =
∗ ∗ ∗ ∗ ∗ ∗ I 1, 2, 3 Pre- and post-multiplying (20) with ξ and ξ T , then
<0 (31) according to Schur complement, we can see that (31) is
0̄11 = Ai P̄ + P̄ATi + Q̄1 + Q̄2 − R̄1 − R̄2 satisfied. This completed the proof.
0̄22 = −Q̄2 − 2Z̄3 + S̄ + S̄ T
IV. NUMERICAL EXAMPLES
0̄23 = R̄3 − S̄ In the section, one example for an ASV is used to verify the
0̄24 = R̄3 − S̄ effectiveness of the given method. Considered the M , D and

VOLUME 8, 2020 148199

 M. Zheng et al.: Stability Analysis and T-S Fuzzy DP Controller Design for ASVs Based on Sampled-Data Control

G in model (1) as follow ([1]): TABLE 1. The maximum upper bound for τM .
 
1.0852 0 0
M = 0 2.0575 −0.4087  ,
0 −0.4087 0.2153
 
0.0865 0 0
D= 0 0.0762 0.1510  .
0 0.0151 0.0031
  sampled-data vessel with DPS in [41]. We choose the same
0.0389 0 0 sampling interval τm = 0, τM = 0.25, then the H∞ per-
G= 0 0.0266 0. formance is obtained that γmin = 5.621.which improves
0 0 0 the corresponding ones in [41] above 135.94%. Besides,
Let α = sin 2◦ and β = cos 88◦ , we have obtained by the the lower and upper bounds of the time-varying delay are
equation as shown at the bottom of the page. both considered in the paper, which has more significance
Let than [41].
   T
1 0 0 1 0 0 Then, the ship’s initial state is xs (t) = 1 1 0.1 0 0 0 .
Ci =  0 1 0 0 1 0  , i = 1, 2, 3. Assuming the sampling interval τm = 0.6, τM = 1.33,
0 0 1 0 0 1 and the network-induced delays τk is assumed to be var-
Firstly, we compare the results in Theorem 2 with the ied stochastically in (τm , τM ), then the H∞ performance is
references which use the traditional T-S fuzzy sampled-data obtained that γmin = 4.032. Then, we can obtain the con-
control method. We choose the same parameters, and then troller gain matrix, is obtained by the equation as shown at
the maximum upper bound of sampling period τM obtained the top of the next page. The responses of the ASV state
by different methods are listed in Table 1, which shows that are shown in Figs.4-6. Fig. 4-5 shows the responses of x,
theorem 2 improves the corresponding ones in [45], [46] and y directions and yaw angle ψ of ASV. Fig.6 shows response
[47] above 112.8%,101.5%, and 17.43%, respectively. Thus, of velocities of ASV.
the sampling period obtained from Theorem 2 are larger than It is easy to know that the ASV can be stabilized by the
those in [44]–[46], which shows that the results in the paper designed controller without extern disturbances. Considered
is less conservative than the references mention above. the extern environment disturbance as follow
Then we compare the results in Theorem 2 with the
results using the same state-derivative feedback for the fuzzy w(t) = 0.1 sin(t) (33)

 
0 0 0 1.0000 −0.0349 0

 0 0 0 0.0349 1.0000 0 

 0 0 0 0 0 1.0000 
,
A1 = 

 −0.0358 0 0 −0.0797 0 0 

 0 −0.0208 0 0 −0.0818 −0.1224 
0 −0.0394 0 0 −0.2254 −0.2468
 
0 0 0 0.0349 −1.0000 0

 0 0 0 1.0000 0.0349 0 

 0 0 0 0 0 1.0000 
,
A2 = 
 −0.0358
 0 0 −0.0797 0 0 

 0 −0.0208 0 0 −0.0818 −0.1224 
0 −0.0394 0 0 −0.2254 −0.2468
 
0 0 0 0.0349 1.0000 0

 0 0 0 −1.0000 0.0349 0 

 0 0 0 0 0 1.0000 
,
A3 = 
 −0.0358
 0 0 −0.0797 0 0 

 0 −0.0208 0 0 −0.0818 −0.1224 
0 −0.0394 0 0 −0.2254 −0.2468
 
0 0 0
 0 0 0 
 
 0 0 0 
Bi = Bwi = , i = 1, 2, 3.
 0.9215 0 0 
 
 0 0.7802 1.4811 
0 1.4811 7.4562
148200 VOLUME 8, 2020
M. Zheng et al.: Stability Analysis and T-S Fuzzy DP Controller Design for ASVs Based on Sampled-Data Control

 
−0.1090 −0.0318 −0.0085 −0.3932 0.0253 0.0002
K1 =  0.0556 −0.1142 0.0338 −0.0321 −0.7675 0.6416  ,
−0.0096 0.0228 −0.0411 0.0035 0.1434 −0.4002
 
−0.1261 −0.0063 −0.0003 −0.4254 0.0088 −0.0012
K2 =  0.0118 −0.1352 0.0281 0.0138 −0.8273 0.6494  ,
−0.0024 0.0264 −0.0399 −0.0031 0.1549 −0.4050
 
−0.1261 −0.0063 −0.0003 −0.4254 −0.0088 −0.0012
K3 =  0.0118 −0.1352 0.0281 0.0138 −0.8273 0.6494  ,
−0.0024 0.0264 −0.0399 −0.0031 0.1549 −0.4050

Then the responses of the ASV state are shown in Fig.7-9, [10] Y.-L. Wang and Q.-L. Han, ‘‘Network-based modelling and dynamic out-
which indicates that the η(t) and velocities υ(t) are stabile in put feedback control for unmanned marine vehicles in network environ-
ments,’’ Automatica, vol. 91, pp. 43–53, May 2018.
a short time. That is, the designed sampled-data controllers [11] N. Wang, S. Lv, M. J. Er, and W.-H. Chen, ‘‘Fast and accurate trajec-
can stabilize the ASV and have good DP performance even if tory tracking control of an autonomous surface vehicle with unmodeled
the external disturbance exists. dynamics and disturbances,’’ IEEE Trans. Intell. Vehicles, vol. 1, no. 3,
pp. 230–243, Sep. 2016.
[12] Y. Wang, H. R. Karimi, H.-K. Lam, and H. Yan, ‘‘Fuzzy output tracking
V. CONCLUSION control and filtering for nonlinear discrete-time descriptor systems under
This article discusses the fuzzy sampled-data control problem unreliable communication links,’’ IEEE Trans. Cybern., vol. 50, no. 6,
pp. 2369–2379, Jun. 2019.
for ASV with DPS. T-S fuzzy model for the ASV has been
[13] W.-J. Chang, H.-Y. Qiao, and C.-C. Ku, ‘‘Sliding mode fuzzy control
established, and both the lower and upper bounds of network- for nonlinear stochastic systems subject to pole assignment and variance
induced delays have been considered in the model. By inte- constraint,’’ Inf. Sci., vol. 432, pp. 133–145, Mar. 2018.
grating the convex reciprocal inequalities and Lyapunov [14] W.-J. Chang, Y.-H. Lin, J. Du, and C.-M. Chang, ‘‘Fuzzy control with
pole assignment and variance constraints for continuous-time perturbed
theorems, adequate stability criterion is derived and less takagi-sugeno fuzzy models: Application to ship steering systems,’’ Int.
conservative result is obtained. The simulation result shows J. Control, Autom. Syst., vol. 17, no. 10, pp. 2677–2692, Oct. 2019.
that the proposed fuzzy sampled-data controller is effective [15] M. Zheng, Z. Yujie, Y. Shenhua, and L. Lina, ‘‘Sampled-data control
in providing good DP performance for the ASV. In the for nonlinear singular systems based on a Takagi–Sugeno fuzzy model,’’
Trans. Inst. Meas. Control, vol. 40, no. 14, pp. 4027–4036, 2018.
future, we will study new inequalities such as Wirtinger-
[16] J. R. Layne and K. M. Passino, ‘‘Fuzzy model reference learning control
based inequalities to improve the results of this article. for cargo ship steering,’’ IEEE Control Syst., vol. 13, no. 6, pp. 23–34,
Dec. 1993.
[17] G. Rigatos and S. Tzafestas, ‘‘Adaptive fuzzy control for the ship steering
REFERENCES
problem,’’ Mechatronics, vol. 16, no. 8, pp. 479–489, Oct. 2006.
[1] N. E. Kahveci and P. A. Ioannou, ‘‘Adaptive steering control for uncer- [18] K.-Y. Seo, G.-K. Park, C.-S. Lee, and M.-H. Wang, ‘‘Ontology-based fuzzy
tain ship dynamics and stability analysis,’’ Automatica, vol. 49, no. 3, support agent for ship steering control,’’ Expert Syst. Appl., vol. 36, no. 1,
pp. 685–697, Mar. 2013. pp. 755–765, Jan. 2009.
[2] Y.-L. Wang and Q.-L. Han, ‘‘Network-based heading control and rudder [19] Y.-L. Wang, Q.-L. Han, M.-R. Fei, and C. Peng, ‘‘Network-based T–S
oscillation reduction for unmanned surface vehicles,’’ IEEE Trans. Control fuzzy dynamic positioning controller design for unmanned marine vehi-
Syst. Technol., vol. 25, no. 5, pp. 1609–1620, Sep. 2017. cles,’’ IEEE Trans. Cybern., vol. 48, no. 9, pp. 2750–2763, Sep. 2018.
[3] M. Chen, S. S. Ge, B. V. E. How, and Y. S. Choo, ‘‘Robust adaptive position
[20] Z. Peng, J. Wang, and D. Wang, ‘‘Distributed maneuvering of autonomous
mooring control for marine vessels,’’ IEEE Trans. Control Syst. Technol.,
surface vehicles based on neurodynamic optimization and fuzzy approxi-
vol. 21, no. 2, pp. 395–409, Mar. 2013.
mation,’’ IEEE Trans. Control Syst. Technol., vol. 26, no. 3, pp. 1083–1090,
[4] R.-Y. Ren, Z.-J. Zou, and X.-G. Wang, ‘‘A two-time scale control law based May 2018.
on singular perturbations used in rudder roll stabilization of ships,’’ Ocean
[21] T. I. M. Fossen, ‘‘Control systems: Guidance, navigation and control of
Eng., vol. 88, pp. 488–498, Sep. 2014.
ships,’’ in Rigs and Underwater Vehicles, 1st ed. Trondheim, Norway:
[5] Y.-L. Wang and Q.-L. Han, ‘‘Network-based fault detection filter and
Marine Cybernetics, 2002.
controller coordinated design for unmanned surface vehicles in network
environments,’’ IEEE Trans. Ind. Informat., vol. 12, no. 5, pp. 1753–1765, [22] Y. Wang, Y. Sui, J. Wu, and J. Jiao, ‘‘Research on nonlinear model pre-
Oct. 2016. dictive control technology for ship dynamic positioning system,’’ in Proc.
IEEE Int. Conf. Autom. Logistics, Aug. 2012, pp. 348–351.
[6] Y. Wang, S. Wang, and M. Tan, ‘‘Path generation of autonomous approach
to a moving ship for unmanned vehicles,’’ IEEE Trans. Ind. Electron., [23] W. Ngongi and J. Du, ‘‘Linear fuzzy controller design for dynamic posi-
vol. 62, no. 9, pp. 5619–5629, Sep. 2015. tioning system of surface ships,’’ Int. J. Intell. Syst. Technol. Appl., vol. 14,
[7] N. Wang, Y. Gao, Z. Sun, and Z. Zheng, ‘‘Nussbaum-based adaptive no. 1, pp. 1–26, 2015.
fuzzy tracking control of unmanned surface vehicles with fully unknown [24] J. Du, X. Hu, M. Krstić, and Y. Sun, ‘‘Robust dynamic positioning of
dynamics and complex input nonlinearities,’’ Int. J. Fuzzy Syst., vol. 20, ships with disturbances under input saturation,’’ Automatica, vol. 73,
no. 1, pp. 259–268, Jan. 2018. pp. 207–214, Nov. 2016.
[8] Z. Peng, X. Hu, and D. Wang, ‘‘Robust adaptive formation control of [25] M. Fu, J. Sun, and D. Wang, ‘‘Research on thrust allocation of dynamic
underactuated autonomous surface vehicles with uncertain dynamics,’’ IET positioning ship with cycloidal propeller,’’ in Proc. 37th Chin. Control
Control Theory Appl., vol. 5, no. 12, pp. 1378–1387, Aug. 2011. Conf. (CCC), Jul. 2018, pp. 620–624.
[9] Z. Peng, D. Wang, Z. Chen, X. Hu, and W. Lan, ‘‘Adaptive dynamic [26] G. Zhang, C. Huang, X. Zhang, and W. Zhang, ‘‘Practical constrained
surface control for formations of autonomous surface vehicles with uncer- dynamic positioning control for uncertain ship through the minimal learn-
tain dynamics,’’ IEEE Trans. Control Syst. Technol., vol. 21, no. 2, ing parameter technique,’’ IET Control Theory Appl., vol. 12, no. 18,
pp. 513–520, Mar. 2013. pp. 2526–2533, Dec. 2018.

VOLUME 8, 2020 148201

 M. Zheng et al.: Stability Analysis and T-S Fuzzy DP Controller Design for ASVs Based on Sampled-Data Control

[27] B. Bidikli, E. Tatlicioglu, and E. Zergeroglu, ‘‘Compensating of added [44] H. Zhang, D. Yang, and T. Y. Chai, ‘‘Guaranteed cost networked control
mass terms in dynamically positioned surface vehicles: A continuous for T–S fuzzy systems with time delays,’’ IEEE Trans. Syst., Man, Cybern.
robust control approach,’’ Ocean Eng., vol. 139, pp. 198–204, Jul. 2017. C, Appl. Rev., vol. 37, no. 2, pp. 160–172, Mar. 2007.
[28] Y. Wang, X. Yang, and H. Yan, ‘‘Reliable fuzzy tracking control of [45] Z.-G. Wu, P. Shi, H. Su, and R. Lu, ‘‘Dissipativity-based sampled-data
near-space hypersonic vehicle using aperiodic measurement information,’’ fuzzy control design and its application to truck-trailer system,’’ IEEE
IEEE Trans. Ind. Electron., vol. 66, no. 12, pp. 9439–9447, Dec. 2019. Trans. Fuzzy Syst., vol. 23, no. 5, pp. 1669–1679, Oct. 2015.
[29] S. M. Abedi, ‘‘An optimal sample-data holds by using a bi-objective [46] C. Peng, Q.-L. Han, D. Yue, and E. Tian, ‘‘Sampled-data robust control for
criterion: Trade-off between the phase delay and the stability robustness,’’ T–S fuzzy systems with time delay and uncertainties,’’ Fuzzy Sets Syst.,
Control Eng. Appl. Informat., vol. 17, no. 2, pp. 32–42, 2015. vol. 179, no. 1, pp. 20–33, Sep. 2011.
[30] Y. Wang, H. R. Karimi, H. K. Lam, and H. Shen, ‘‘An improved result [47] Z. Du, Z. Qin, H. Ren, and Z. Lu, ‘‘Fuzzy robust H∞ sampled-data control
on exponential stabilization of sampled-data fuzzy systems,’’ IEEE Trans. for uncertain nonlinear systems with time-varying delay,’’ Int. J. Fuzzy
Fuzzy Syst., vol. 26, no. 6, pp. 3875–3883, Dec. 2018. Syst., vol. 19, no. 5, pp. 1417–1429, 2017.
[31] F. Chen, H. Garnier, and M. Gilson, ‘‘Robust identification of continuous-
time models with arbitrary time-delay from irregularly sampled data,’’
J. Process Control, vol. 25, pp. 19–27, Jan. 2015.
[32] Y. Wang, C. K. Ahn, H. Yan, and S. Xie, ‘‘Fuzzy control and filtering MINJIE ZHENG received the B.S. and M.S.
for nonlinear singularly perturbed Markov jump systems,’’ IEEE Trans. degrees from Xiamen University, in 2005 and
Cybern., early access, Jul. 22, 2020, doi: 10.1109/TCYB.2020.3004226. 2008, respectively, and the Ph.D. degree in elec-
[33] E. Fridman, ‘‘Robust sampled-data H∞ control of linear singularly per- tronic science and technology from Shanghai Jiao
turbed systems,’’ IEEE Trans. Autom. Control, vol. 51, no. 3, pp. 470–475,
Tong University, Shanghai, China, in 2018. He is
Mar. 2006.
[34] D. W. Kim, H. J. Lee, M. H. Kim, S.-Y. Lee, and T.-Y. Kim, ‘‘Robust
currently an Associate Professor with the Naviga-
sampled-data fuzzy control of nonlinear systems with parametric uncer- tion College, Jimei University, Xiamen, China. His
tainties: Its application to depth control of autonomous underwater vehi- research interests include robust control, sampled-
cles,’’ Int. J. Control, Autom. Syst., vol. 10, no. 6, pp. 1164–1172, data control, and fuzzy control.
Dec. 2012.
[35] M. Liaquat and M. B. Malik, ‘‘Sampled data output regulation of n-link
robotic manipulator using a realizable reconstruction filter,’’ Robotica,
vol. 34, no. 4, pp. 900–912, Apr. 2016.
[36] M. Moarref and L. Rodrigues, ‘‘Sensor allocation with guaranteed expo-
SHENHUA YANG received the B.S. degree from
nential stability for linear multi-rate sampled-data systems,’’ Int. J. Robust
Nonlinear Control, vol. 26, no. 7, pp. 1512–1529, May 2016.
Jimei University, Xiamen, China, in 1998, and
[37] Z. G. Wu, P. Shi, and H. Y. Su, ‘‘Stochastic synchronization of Markovian the Ph.D. degree from Shanghai Maritime Univer-
jump neural networks with time-varying delay using sampled data,’’ IEEE sity, Shanghai, China. He is currently a Professor
Trans. Cybern., vol. 43, no. 6, pp. 796–1806, Jan. 2013. with the Navigation College, Jimei University.
[38] Y. Wang, X. Xie, M. Chadli, S. Xie, and Y. Peng, ‘‘Sliding mode control of His research interests include artificial intelligence
fuzzy singularly perturbed descriptor systems,’’ IEEE Trans. Fuzzy Syst., technology, virtual simulation technology, and
early access, May 29, 2020, doi: 10.1109/TFUZZ.2020.2998519. navigation technology.
[39] Y. Wang, W. Zhou, J. Luo, H. Yan, H. Pu, and Y. Peng, ‘‘Reliable
intelligent path following control for a robotic airship against sensor
faults,’’ IEEE/ASME Trans. Mechatronics, vol. 24, no. 6, pp. 2572–2582,
Dec. 2019, doi: 10.1109/TMECH.2019.2929224.
[40] S. C. Martin and L. L. Whitcomb, ‘‘Nonlinear model-based tracking
LINA LI received the B.S. degree from Dalian
control of underwater vehicles with three Degree-of-freedom fully cou-
Maritime University, Dalian, China, in 1979.
pled dynamical plant models: Theory and experimental evaluation,’’ IEEE
Trans. Control Syst. Technol., vol. 26, no. 2, pp. 404–414, Mar. 2018. She is currently a Professor with the Navigation
[41] M. Zheng, Y. Zhou, and S. Yang, ‘‘Robust fuzzy sampled-data control for College, Jimei University, Xiamen, China. Her
dynamic positioning ships,’’ J. Shanghai Jiaotong Univ. Sci., vol. 23, no. 2, research interests include marine intelligent trans-
pp. 209–217, Apr. 2018. portation technology, intelligent navigation, and
[42] K. Q. Gu, V. L. Kharitonov, and J. Chen, Stability of Time-Delay Systems. automatic control of ships.
Boston, MA, USA: Birkhäuser, 2003.
[43] J. Yoneyama, ‘‘Robust sampled-data stabilization of uncertain fuzzy sys-
tems via input delay approach,’’ Inf. Sci., vol. 198, pp. 169–176, Sep. 2012.

148202 VOLUME 8, 2020