IEEE TRANSACTIONS ON SYSTEMS, MAN, AND CYBERNETICS: SYSTEMS, VOL. 46, NO. 1, JANUARY 2016 27
Observer-Based Adaptive Fuzzy Control for a
Class of Nonlinear Delayed Systems
Bing Chen, Chong Lin, Senior Member, IEEE, Xiaoping Liu, and Kefu Liu
Abstract—This paper considers the problem of observer-based
adaptive fuzzy control for a class of nonlinear time-delay systems
in nonstrict-feedback form, which includes the nonlinear strict-
feedback systems as a special case. An adaptive fuzzy output
feedback backstepping approach is first proposed for nonlinear
systems in nonstrict-feedback form. Fuzzy logic systems are used
to approximate the unknown nonlinear functions. Adaptive tech-
nique and backstepping are utilized to construct a controller.
The proposed adaptive fuzzy output feedback controller guaran-
tees that all the signals in the adaptive closed-loop system are
semi-globally uniformly ultimately bounded. Simulation results
are provided to demonstrate the effectiveness of the presented
approach.
Index Terms—Adaptive control, backstepping, fuzzy control,
nonlinear systems, time delay.
I. INTRODUCTION
A
DAPTIVE backstepping, which was first used for nonlin-
ear control in [1], leads to the discovery of a structurally
strict-feedback condition, under which the systematic con-
struction of a robust control scheme is always possible. So
far, the backstepping-based adaptive control technique has
become one of the most popular design methods for a large
class of nonlinear systems (see [2]–[7] and the references
therein). Earlier classical adaptive backstepping is mainly used
for the robust control design of nonlinear systems with para-
metric uncertainties. However, a limitation of these works
is that they cannot be applied to nonlinear systems where
exact knowledge of the structure of system functions is
unavailable. To overcome this problem, adaptive fuzzy/neural
backstepping control strategies have been proposed [8]–[11],
respectively. In those reported adaptive fuzzy/neural control
schemes, neural networks or fuzzy logic systems are used
to approximate the completely unknown nonlinear functions,
a classical adaptive technique is used to estimate the ideal
weighting vector of neural networks or fuzzy logic systems
and then the backstepping approach is employed to construct
Manuscript received October 21, 2014; revised January 15, 2015; accepted
February 28, 2015. Date of publication April 24, 2015; date of current
version December 14, 2015. This work was supported by the National
Natural Science Foundation of China under Grant 61473160, Grant 61174033,
and Grant 61034005. This paper was recommended by Associate Editor
S. Tong.
B. Chen and C. Lin are with the School of Automation
Engineering, Qingdao University, Qingdao 266071, China (e-mail:
chenbing1958@126.com).
X. Liu and K. Liu are with the Faculty of Engineering, Lakehead University,
Thunder Bay, ON P7B 5E1, Canada.
Color versions of one or more of the figures in this paper are available
online at http://ieeexplore.ieee.org.
Digital Object Identifier 10.1109/TSMC.2015.2420543
the desired controllers. With the use of neural networks or
fuzzy logic systems, the need for exact knowledge of sys-
tem nonlinearities is removed. In [12]–[15], an adaptive neural
backstepping control approach was used to design robust
controllers for single-input single-output nonlinear strict-
feedback systems with completely unknown nonlinearities.
In [16] and [17], decentralized adaptive neural backstepping
control strategies were proposed for a class of large-scale non-
linear systems. Simultaneously, backstepping-based adaptive
fuzzy control schemes were also investigated in [18]–[21].
In recent years, the works in [22]–[24] further investigated
adaptive fuzzy/neural control for nonlinear strict-feedback
systems with unknown input nonlinearities. The research
in [25] developed a novel fuzzy predictive adaptive control
scheme. Subsequent works extended the problem to more com-
plex nonlinear systems cases, for instance, see [26]–[33]for
multi-input multioutput systems (MIMO), [34]–[38] for time-
delay systems, and [24], [39]–[42] for nonlinear stochastic
systems.
Notice that all the above-mentioned works are essentially
based on state feedback control methodology. The assump-
tion of state variables being available must be imposed on
the controlled systems, which limits the applicability of these
control schemes in practical engineering. Therefore, state esti-
mation and observer-based stabilization become significant
issues in [43]–[45]. In an effort to develop output feedback
adaptive fuzzy/nerual control strategies, the work in [46] stud-
ied the problem of observer-based adaptive neural control for a
class of simple nonlinear strict-feedback systems and obtained
a regionally stable result. This result was subsequently
extended to nonlinear strict-feedback time-delay systems
in [36], [47], and [48] and was further extended to nonlin-
ear interconnected systems in [49]. Hereafter, more results on
adaptive neural/fuzzy output feedback control were reported.
In [50], the problem of observer-based adaptive fuzzy output
feedback control was considered for nonlinear delayed systems
with unknown control direction. By using Nussbaum-type
function, an adaptive fuzzy output feedback controller was
proposed, which guaranteed the uniformly ultimate bounded-
ness of all the adaptive closed-loop signals. In [39] and [40],
adaptive neural output feedback control schemes were devel-
oped for stochastic nonlinear systems, with the assumption
that the systems are in strict-feedback form and the unknown
nonlinearities are the functions of the system output variables.
In [51], a direct output feedback adaptive fuzzy controller
was constructed, and the result was recently extended to time-
delay systems in [52]. An indirect output feedback adaptive
2168-2216
c
2015 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.
See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.