Abstract—
Due to the strong nonlinearity, coupling and
unknown external disturbances of the robot system, the
problem of high precision trajectory tracking control of robot is
always difficult to solve. Therefore, a robust adaptive trajectory
tracking method based on extended state observer (ESO) is
proposed in this paper. Extended state observer can estimate the
sum of the internal and external disturbances, that is, the total
disturbances of the system, and compensate it to achieve high
precision tracking of the robot trajectory. Robust adaptive
algorithm has good self-adjustability to the structural
parameters of the robot system, so the robustness of the
controller can be guaranteed. Simulation results show that the
proposed method has good tracking performance and
anti-interference performance.
I. INTRODUCTION
With the development of technology, more and more
modern factories use robots instead of human beings in a
large number of repetitive and high-risk jobs. This not only
saves a lot of human capital, but also significantly improves
work efficiency. Robot manipulator is a typical system of
multi-input-multi-output, nonlinear, strong coupling [1-2].
There are a lot of uncertainties in the system (internal friction,
system structural parameters, load changes, etc.). Some
control methods based on precise mathematical model of
controlled object cannot achieve satisfactory results.
Therefore, it is difficult to realize high-speed and
high-precision trajectory tracking control of manipulators.
In recent years, many representative control methods have
been proposed to achieve better control performance of robot
manipulators. If the robot's gravity and external disturbance
are ignored, the independent PD control can satisfy the
requirement of the robot's fixed-point control. However, such
a robot system does not exist at all, and independent PD
control can only be used as the basis of analysis. When the
gravity is considered, the fixed-point control of the robot can
be satisfied by adopting the PD control based on gravity
compensation. Since the estimation of gravity moment cannot
be very accurate, an adaptive algorithm for on-line estimation
of gravity is designed in [3], which realizes PD control of
robot based on on-line gravity compensation. In [4], the
authors adopt the fixed gravity term calculated beforehand as
compensation and adopt the method of increasing feedback
gain to reduce the steady-state error. Although the above
methods achieve fixed-point control of the robot, most of
*Research supported by supported by National Natural Science
Foundation (NNSF) of China under Grant 61873022, 61573052, Beijing
Natural Science Foundation under Grant 4182045 and the Fundamental
Research Funds for the Central Universities under Grant XK1802-4.
Dazi Li and Jiang Wang are with the Department of Automation, Beijing
University of Chemical Technology, Beijing, 100029, P. R. China (Email:
lidz@ mail.buct.edu.cn).
these algorithms require a lot of calculation and require the
actuator to provide a large initial torque at the initial moment.
In [5], the authors put forward an adaptive robust PD control
strategy, which avoids excessive initial torque output of the
actuator and guarantees good dynamic performance of the
system. But this method has not given more consideration to
anti-interference of the controller.
Many methods based on modern control theory and
intelligent control theory are used to realize trajectory control
of manipulator, such as robust fuzzy adaptive method [6-7],
neural network control method [8-10], Iterative learning
control method [11] and so on. Although these methods can
realize the trajectory tracking task of manipulator, due to
complex structure, poor reliability and many parameters to be
adjusted, these methods cannot be well applied in practice.
As a kind of disturbance observer, the extended state
observer (ESO) has attracted much attention in recent years.
The idea of total disturbance (that is, the sum of internal
disturbance and external disturbance) can help the system
effectively reject the influence of disturbance and reduce the
error caused by disturbance in the process of system operation.
Robust adaptive algorithm has strong adaptability to the
change of system structure parameters, so it is widely used in
the process control of the system. Since the structural
parameters of the manipulator with multiple degrees of
freedom will change slightly in the process of motion, the
application of the robust adaptive control method to the
trajectory tracking problem of the manipulator will produce
better control effect.
In this paper, we propose a practical robust adaptive
method based on ESO. The control structure can adjust the
control law adaptively according to the change of the system
structure parameters or load to adapt to the dynamic change
of the system. In addition, the ESO compensation structure is
introduced in the control structure, which can deal with the
effect of disturbance (internal disturbance and external
disturbance) for the control performance very well. Therefore,
the trajectory of the manipulator can be controlled accurately,
and it has good anti-interference and robustness.
The remainder of this paper is organized as follows. In
Section II, we briefly introduce the robot kinematics and
dynamics scheme. In Section III, we discuss a robust adaptive
method for robot manipulators with known disturbance upper
bounds. At the same time, a practical robust adaptive method
based on extended state observer is proposed. In Section IV,
simulations are carried out with two-link robot manipulators
and In Section V, we conclude with a brief summary of this
paper.
A Robust adaptive method based on ESO for Trajectory Tracking of
Robot Manipulator
Dazi Li, Jiang Wang
2019 IEEE 15th International Conference on Control and Automation (ICCA)
Edinburgh, Scotland, July 16-19, 2019
978-1-7281-1163-6/19/$31.00 ©2019 IEEE 506