Nussbaum Function Based Control Design for A Class of
Uncertain Stochastic High-order Nonlinear Systems
ZHANG Jian
1
,MUXiaowu
1
, LIU Yungang
2
1. School of Mathematics and Statistics, Zhengzhou University, Zhengzhou 450001, P. R. China
E-mail: zhangjian@zzu.edu.cn, muxiaowu@zzu.edu.cn
2. School of Control Science and Engineering, Shandong University, Jinan 250061, P. R. China
E-mail: lygfr@sdu.edu.cn
Abstract: In this paper, the global adaptive stabilization is investigated for a class of uncertain stochastic high-order nonlinear
systems in which there exist serious nonlinearities and unknowns. For details, on one hand, the considered systems allow the
higher system nonlinearities (i.e. the power of the function) and include the strict-feedback systems as its special case. On
the other hand, the unknown parameterized system nonlinear terms and the unknown control direction (i.e. the sign of the
control gain) are both considered in the control design, without a doubt, those unknowns make the system appreciably general
and practical. By skillfully combining the adaptive design method and Nussbaum function based design technique, a feasible
adaptive stabilizing control scheme is successfully presented, by which the obtained controller can effectively guarantee that all
the closed-loop system states are bounded almost surely, and especially the original system states converge to the origin with
probability one.
Key W ords: Stochastic High-order Nonlinear Systems, Unknown Control Direction, Nussbaum Function, Global Adaptive
Stabilization.
1 Introduction
Obviously, the stability analysis and stabilizing control
design of systems is the elementary problem of the control
theory. Stochastic system is a typical system which is wor-
thy of intensively studying, this is because of the universality
and practicality of the system models. As a matter of fact,
the actual systems are likely influenced by some stochastic
disturbances or changes, and hence the systems inevitably
will be modeled by various types of stochastic systems, such
as the Itˆo type [1] and the Markov type[2]. Consequently,
the research on stochastic stability analysis and control de-
sign has been prosperously developed in the past decades,
e.g. see [3–7] and the references therein. Moreover, the es-
sential nonlinearities usually can not be facilely neglected in
the modeling, this directly give birth to the control problems
of stochastic nonlinear systems which are much complicat-
ed, and hence the targeted methods should be proposed and
much progress has been made [8–16].
Recently, the control problem of the stochastic high-order
nonlinear systems have attracted much intensive attention,
e.g. see [17–20]. This paper mainly considers the global
state-feedback stabilization of the uncertain stochastic high-
order nonlinear systems as follows:
⎧
⎪
⎪
⎪
⎨
⎪
⎪
⎪
⎩
dx
1
= x
p
2
dt + f
1
(t, x)dt + g
1
(t, x)
T
dw,
···
dx
n−1
= x
p
n
dt + f
n−1
(t, x)dt + g
n−1
(t, x)
T
dw,
dx
n
= du
p
dt + f
n
(t, x)dt + g
n
(t, x)
T
dw,
(1)
where x =[x
1
,...,x
n
]
T
∈ R
n
is the measurable system
state with the initial value x(t
0
)=x
0
; u ∈ R is the sys-
tem control input; p ≥ 1 is an odd integers; d is an un-
known nonzero constant, called the control gain;thedrift
terms f
i
: R
+
× R
n
→ R ,i =1,...,n and the diffusion
This work was supported by the National Natural Science Foundations
of China (61403348, 11571322, 61325016), the Key Science and Technol-
ogy Research Projects of the Education Department of Henan Province of
China (14B120004).
terms g
i
: R
+
× R
n
→ R
m
,i =1,...,n are unknown lo-
cally Lipschitz nonlinear functions, collectively referred to
as the system nonlinearities; and w ∈ R
m
is an independent
standard Wiener process defined on a complete probability s-
pace (Ω, F ,P) with Ω being a sample space, F a σ-algebra
on Ω,andP a probability measure. For simplicity of expres-
sion, we denote [x
1
,...,x
k
]
T
by x
[k]
.
When p =1, system (1) accord with the standard from
of stochastic strict-feedback nonlinear systems which have
been wildly investigated and are of plentiful significant re-
sults [5, 8–14]. However, it is necessary to mention that
those existing results are inapplicable to system (1) for the
case p>1, Thus, by noting the general form of system (1),
which also show the higher degree of the system nonlinear-
ities (i.e. the power of the function), the control problem of
system (1) is certainly meaningful and worthy of investiga-
tion. Moreover, it is clear that the more unknowns in the
system, the more challenging of achieving the desired con-
trol objective. Hence, the unknowns in the system nonlinear-
ities usually can not be neglected during the researching, d-
ifferent from the simple restriction that the uncertain system
nonlinear terms could be bounded by the known functions,
the more complicated case is that one have to suffer from the
upper bounds with unknown parameters, i.e. the so-called
unknown parameterized nonlinearities[19].
On the other hand, the unknowns may also be imposed on
the control gain of (1), whose familiar supposition is that it
may be bounded by known or unknown constants, but espe-
cially its sign must be known[17–20]. In fact, the sign of the
control gain, also called the control direction, is an importan-
t system information for constructing the desired controller,
without it, the corresponding stabilization will be more dif-
ficult. To author’s knowledge, under the unknown control
direction, there have already been some works on the non-
stochastic systems (e.g. [21–23]) and the stochastic strict-
feedback systems (e.g. [24–26]), but for the considered sys-
tem (1), there is only work [27] investigates the related sta-
Proceedings of the 36th Chinese Control Conference
Jul
26-28, 2017, Dalian, China
499