IEEE/ASME TRANSACTIONS ON MECHATRONICS, VOL. 19, NO.

6, DECEMBER 2014

1737

Experimental Investigation of Robust Motion

Tracking Control for a 2-DOF
Flexure-Based Mechanism
Umesh Bhagat, Student Member, IEEE, Bijan Shirinzadeh, Member, IEEE, Leon Clark, Student Member, IEEE,
Yanding Qin, Yanling Tian, and Dawei Zhang

AbstractThe design, parameter identification and robust motion tracking control of a two degree of freedom (2-DOF) flexurebased micro/nanomechanism are presented in this paper. In the
presented compliant mechanism, the cross-axis coupling ratio is
below 1% indicating excellent decoupling performance. Despite
this, during motion tracking the cross coupling effect cannot be ignored. To enhance the accuracy of micro/nanomanipulation, a laser
interferometry-based sensing and measurement system is established. Nonlinearities such as creep/drift and hysteresis are present
in this system, which are compensated with closed-loop control.
Open-loop tracking results for a 1-DOF trajectory, with and without cross-axis coupling compensation are also presented. Robust
motion tracking control is extended to support 2-DOF motion trajectories. This controller is implemented to track the desired trajectories over one and two axes of motion. Robust motion control
demonstrates high precision and accurate motion tracking of the 2DOF flexure-based mechanism. The cross-axis coupling is treated
as a known disturbance and the performance of tracking 1-DOF
trajectory, with and without cross-axis coupling compensation, is
presented. Circular motion trajectories with radii of 10 m, 1 m,
and 250 nm are also tracked. The experimental results presented
in this paper demonstrate effective compensation of the cross-axis
coupling with high precision motion tracking. The resultant 2-DOF
closed-loop position tracking error in the X and Y axes are within
20 nm during dynamic motion, and 8 nm in the steady state.
Index TermsCross-axis coupling, laser interferometry,
micro/nanomechanism, robust motion control.

Manuscript received August 21, 2013; revised November 27, 2013; accepted
December 20, 2013. Date of publication January 31, 2014; date of current version June 13, 2014. Recommended by Technical Editor S. Fatikow. This work
was supported in part by an ARC LIEF Grant (LE0347024 and LE0775692), in
part by ARC Discovery Projects (DP110104970, DP0666366, and DP0986814),
and in part by an ARC Discovery Grant (DP0450944).
U. Bhagat is with the Department of Mechanical and Aerospace Engineering, Monash University, Clayton, Vic. 3800, Australia (e-mail: umesh.
[email protected]).
B. Shirinzadeh and L. Clark are with the Robotics and Mechatronics
Research Laboratory, Department of Mechanical and Aerospace Engineering,
Monash University, Clayton, Vic. 3800, Australia (e-mail: bijan.shirinzadeh@
monash.edu; [email protected]).
Y. Qin is with the Tianjin Key Laboratory of Intelligent Robotics, Nankai
University, Tianjin 300071, China (e-mail: [email protected]).
Y. Tian and D. Zhang are with the School of Mechanical Engineering, Tianjin University, Tianjin 300072, China (e-mail: [email protected];
[email protected]).
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/TMECH.2014.2300481

I. INTRODUCTION
ECENT advancement and rapidly increasing requirements
in the field of engineering, biotechnology, and medical sciences have created need for high-speed and ultraprecise
micro/nanopositioning systems [1][7]. Flexure-based mechanisms driven by piezoelectric actuators have been recognized as
appropriate platforms for micro/nanomanipulation. The combination of flexure-based mechanisms with piezoelectric actuation
can produce smooth, continuous, highly precise, frictionless,
and fast motions. Research in the area of design and development of compliant mechanisms has increased significantly
in the last decade [8][16]. The nonlinearities associated with
the piezoelectric actuation, such as hysteresis, creep, and drift
cannot guarantee positioning accuracy and precise motion tracking of the flexure-based mechanism. To compensate for these
nonlinearities and to achieve desired performance of the flexurebased systems, various closed-loop control methodologies have
been proposed and established [17][21].
Many multiaxis flexure-based micro/nanomanipulator mechanisms use parallel kinematic chains in their design. Parallel
structure-based mechanisms offer higher stiffness and accuracy
as compared with serial structures [4], [22][24]. Cross-axis
coupling generally exists in most multiaxis compliant mechanisms designed for micro/nanopositioning and manipulation
[25]. The presence of cross-axis coupling increases the complexity of modeling and controlling such micro/nanomanipulators.
Large cross-axis coupling can damage the piezoelectric actuators. One solution to eliminate cross-axis coupling is to build
serial mechanisms by stacking or nesting multiple 1-degree of
freedom (DOF) structures [26]. However, such strategies result
in nonuniform performance across the workspace.
In recent years, the use of parallel kinematic chains has gained
attention because of its distinct advantages such as low inertia,
high stiffness, high accuracy, and identical behaviors in each
axis. Parallelogram-based structures are widely utilized to build
decoupled parallel mechanisms. However, considerations of decoupling capability, structural complexity, and dynamic performance demand novel and accurate design. In a piezo-driven
compliant system, in addition to the hysteresis effect, discrepancies in the desired performance can also result from modeling
inaccuracies, the contact interface between the piezoelectric actuator and the mechanism manufacturing tolerances, and the
varying properties of the piezoelectric actuator under different
circumstances.

1083-4435 2014 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.

1738

IEEE/ASME TRANSACTIONS ON MECHATRONICS, VOL. 19, NO. 6, DECEMBER 2014

parameter dynamic model of the 2-DOF flexure-based mechanism, driven by piezoelectric actuators is given as follows:

ml p x x
m
ml p y ym

bl p x x m
bl p y y m

k l p x xm
kl p y ym

vh x
vh y

vi n x

vi n y

(1)
blpx
klpx
mlpx
), (
), and(
) are mass, damping, and
where (
mlpy
blpy
klpy
x
stiffness of the system in the X and Y axes and, ( m ) are
ym
vin x
the actuator displacement in the X and Y axes, (
) are the
vin y
v
applied (input) voltage, and ( hx ) are the hysteresis voltages
vhy
added into the equation to account for the nonlinear effects
of hysteresis, drift, and external disturbances. By using the
lumped parameter dynamic model given in (1), the motion
equation for the 2-DOF flexure-based micro/nanomanipulator
is extended to account for all the nonlinear effects and external
disturbances,
Fig. 1.

Schematic diagram of the 2-DOF flexure-based mechanism.

This paper presents the structure and modeling of a 2-DOF

flexure-based planar mechanism in Section II. In Section III, the
modeling of uncertainties and considerations for robust motion
tracking control methodology are presented. The experimental
setup with laser interferometry-based sensing and measurement
methodology, together with cross-axis coupling estimation are
presented in Section IV. The experimental study of tracking
desired trajectories in the X, Y and XY axes are described
and analyzed in Section V. Conclusions and future work are
presented in Section VI.
II. MODEL OF 2-DOF FLEXURE-BASED
MICRO/NANOMANIPULATOR
A schematic diagram of the 2-DOF flexure-based mechanism
is shown in Fig. 1. Identical kinematic chains in the X and Y
axes guarantee the same stiffness and bandwidth in both directions, resulting in uniform performance across the workspace.
Development and design optimization of this decoupled 2-DOF
flexure-based mechanism is presented in our previous research
[27]. The material for the mechanism is chosen as Aluminum
7075-T6, which is light weight, strong, and has good fatigue
strength as compared with other materials. This aluminum alloy has a density of 2700 kg/m3 , Youngs modulus of 70 GPa,
and a yield strength of 434 MPa. The overall dimension of the
presented 2-DOF mechanism is 187 mm 187 mm 20 mm.
This 2-DOF stage provides sufficient space for the PEAs installment and for measurement components on the central
platform.
A lumped parameter dynamic model which combines the 2DOF flexure-based mechanism and the piezoelectric actuators
are prepared for the purpose of high-precision motion control.
This is realized by extending the model of a piezoelectric
actuator presented in previous studies [28], [29]. The lumped

vhx
vhy


=

vn x
vn y


+

vdx
vdy


(2)

v
vn x
) and ( dx ) represent the nonlinear effects and
vn y
vdy
external disturbances. With the model described in (1), an advanced control methodology is established to effectively control
the piezo-driven 2-DOF flexure-based micro/nanomanipulator
as described further.
where (

III. UNCERTAINTY ANALYSIS AND ROBUST MOTION

TRACKING CONTROL METHODOLOGY
Robust motion tracking methodology has been designed and
utilized in our previous research study for single axis flexurebased mechanism control [26]. The robust motion tracking control methodology is employed to track the desired motion trajectories of 2-DOF parallel compliant mechanism in the X, Y , and
XY axes. The control methodology is designed to adapt the unknown system parameters, nonlinearities associated with piezoelectric actuation, cross-axis coupling motion, and external disturbances in the 2-DOF flexure-based micro/nanomanipulation
system. Under the proposed control approach, the physical parameters of the system are assumed to be unknown or uncertain. However, there exist bounded nonlinear effects and external disturbances within closed-loop system. With the estimated system parameters, and their associated bounds, the robust motion tracking control approach is developed for the multiDOF mechanism. The bounds of the system parameters and
associated uncertainties of the motion system are expressed as
follows:



 
 

 mlpx
 mlpx 
m
lpx 
mlpx





(3a)
 =


m
lpy 
 mlpy
 mlpy 
mlpy

BHAGAT et al.: EXPERIMENTAL INVESTIGATION OF ROBUST MOTION TRACKING CONTROL FOR A 2-DOF FLEXURE-BASED MECHANISM







 b 
blpx  
 blpx 
lpx
blpx








 =
 blpy 
blpy 
blpy
 blpy

 




 k 
 klpx 
klpx 
klpx
 lpx




 = 
klpy
 klpy 
klpy 
 klpy

1739

(3b)

(3c)

b
klpx
mlpx
), ( lpx ), and (
) represent the paramlpy
blpy
klpy
b
k
m

metric errors, ( lpx ), ( lpx ), and ( lpx ) represent the estim

lpy
blpy
klpy
blpx
k
mlpx
), (
), and ( lpx ) denote
mated parameters, and (
mlpy
blpy
klpy
the bounds of the system parameters for the X and Y axes,
respectively. In establishing the robust control methodology, a
switching function is specified as
where (


=

e px

e py


+


(4)

e px
) are velocity errors in the X and Y axes, respece py

tively, and ( x ) are the state of a dynamic compensator used

y
to shape the position tracking errors in the X and Y axes. The
dynamic compensator is chosen as described in the following
equation:

 
 

 
x x
kpx epx
kv x e px
x
=
+
+
y y
kpy epy
kv y e py
y

(5)

where ( px ) are position errors in X and Y, ( x ) are constant

epy
y
k
k
vectors greater than zero, and ( px ) and ( v x ) are the control
kpy
kv y
gains which are related to the specified target performance where
the cumulative tracking error converge to zero. It is assumed
that the exact values of the system parameters including the
mass, damping, and stiffness in (1) are unknown. However, the
estimated values and corresponding bounds are known. With
this assumption, gain values are chosen as
kpx


=

kpy


kv x
kv y



m1
dy


m1
dx

m1
dx bdx
m1
dy bdy

kdx

kdy
 

Schematic of the experimental setup.

boundary layer is obtained as

epssx
kdx

kdy
epssy
ssx

=
ssy
mdx
x
mdy
y

(7)

epssx
) are the steady-state position errors in the X and
epssy
Y axes. By comparing the system with the standard secondorder characteristic equation, the estimated system parameters
are obtained as

   2 

  

md x
bd x
x nx
kd x
n x
1
=
,
=2
, and
=
1
md y
bd y
y ny
kd y
n2 y
(8)
x
n x
where ( ) are damping ratios and (
) are the undamped
y
n y
natural frequencies.
where (

where (

Fig. 2.

(6a)
x
y


.

(6b)

To keep the accuracy of the switching function guaranteed

to stay within the boundary layer, a saturation function is introduced in the control rule [28]. The robust motion tracking
control law is formulated to drive the system to reach the sliding

mode. The steady-state switching function ( ssx ) within the

ssy

IV. EXPERIMENTAL SETUP AND CROSS-AXIS COUPLING

ESTIMATION
A. Experimental Setup
An experimental research facility to track desired motion trajectories in the X, Y and XY directions is established utilizing a
laser interferometry-based sensing and measurement technique.
Real-time capability of the control system is enhanced with
VMEbus system to perform the laser interferometry-based digital position and velocity feedback.
The system architecture is shown in Fig. 2. Two flat mirrors are mounted on the central platform of the 2-DOF flexurebased mechanism. Two piezoelectric actuators from Physik Instrumente (model P-843.30) are installed and precompression
forces are applied to keep the actuators tip and the mechanism in contact. These piezoelectric actuators are multilayer
PZT stacked ceramic translators capable of 45 m displacement corresponding to a range of operating voltage from 0 to
100 V. They have nominal open-loop resolution of 0.9 nm and
closed-loop resolution of 0.45 nm. The amplifier module from
Physike Instrumente has a fixed output gain of 10, providing a
voltage range from 20 to +120 V. A control PC, with 16-bit
analogue-to-digital (A/D) and digital-to-analogue (D/A) modules perform the control action and data acquisition tasks. A
control PC is also equipped with an operating system capable of

1740

IEEE/ASME TRANSACTIONS ON MECHATRONICS, VOL. 19, NO. 6, DECEMBER 2014

TABLE I
FIRST NATURAL FREQUENCY OF THE 2-DOF MECHANISM

TABLE II
LUMPED PARAMETERS OF FLEXURE-BASED 2-DOF MICRO/NANOMECHANISM

Fig. 3.

Fig. 4.

Schematic of high stability plain mirror interferometry.

Photograph of the experimental facility.

hard real-time control and a PCI/VME interface: a PCI board

(PCI-MXI-2 board) and a 6U VME bus board (VME-MXI2 module). The laser interferometry-based sensing and measurement technique is developed with optical components from
Zygo, with high stability plain mirror interferometer (HSPMI)
configuration. The schematic of HSPMI configuration is shown
in Fig. 3. Zygo laser head generates HeliumNeon frequency
stabilized laser beam with two frequency components (f1 and
f2 ) that are orthogonally polarized, and offset by 20 MHz.
The overall system is installed on an optical table with four
laminar flow isolators to obtain seismic isolation. The experiments are performed in a stable environment where variations of
room temperature and humidity were kept under control, covering the entire experimental setup with a box. The photograph
of the system setup is provided in Fig. 4.
B. Parameter Estimation
A computational approach has been used to estimate the
mechanisms performance and is validated with experimentation, details of which are presented in our previous study.
Modal analysis is performed to investigate and establish the
first natural frequency, modes, and modal shapes with and
without actuators installed. The frequency response of the 2DOF mechanism is experimentally verified to evaluate its dynamic performance within 5 kHz. A modal hammer (MEGGITT ENDEVCO 2302-10) has been used to provide excitation
to the 2-DOF mechanism, and the response is measured by
two accelerometers (Bruel & Kjr 4507B) mounted on central

platform of the mechanism. The computational and experimental values of first natural frequency are presented in Table I.
With the first natural frequency, which is considered as a
lowest structural resonance (r m ), undamped natural frequency
(n ) is estimated using a rule of thumb, that undamped natural
frequency is confined to n 12 r m . Systems estimated values
of mass, damping, and stiffness are calculated with (8). The estimated values of the system parameters and their corresponding
bounds are presented in Table II.

C. Estimation of Cross-Axis Coupling

Cross-axis coupling is inevitable in the parallel compliant
mechanisms, and estimation of which helps in establishing
effective control. Cross-axis coupling is estimated to utilize
single-input-single-output control strategies for tracking. In order to estimate the cross-axis coupling effect in the presented
2-DOF flexure-based mechanism, nonnegative voltages of up
to 9.9 V are applied to both X and Y piezo drivers. The displacement of the central platform is measured with the laser
interferometry-based sensing and measurement technique. The
measured workspace is approximately a square with the maximum displacements of 84 and 83 m in the X and Y axes, respectively. The slight discrepancy in the X and Y displacement
is attributed to possible manufacturing tolerances, measurement
uncertainties, and the unavoidable variations in precompression
operations. The cross-axis coupling is measured by activating
the piezoelectric actuator in the X or Y axis, while a constant
zero voltage is applied on the other axis. Fig. 5(a) shows measured workspace of the 2-DOF flexure-based mechanism. The
experimental result with the X-axis activated and the central
platforms displacement in the Y -axis, representing the coupling
motion, is shown in Fig. 5(b). The resulting (xcoupling /yprim ary )

BHAGAT et al.: EXPERIMENTAL INVESTIGATION OF ROBUST MOTION TRACKING CONTROL FOR A 2-DOF FLEXURE-BASED MECHANISM

Fig. 6.

1741

Jerk-free desired 1-DOF trajectory.

A. One-DOF Desired Motion Trajectory

Under the proposed control methodology, the closed-loop
system is required to follow the jerk-free desired motion trajectory. This desired motion trajectory is formed by segments of
higher order polynomials with zero acceleration at the beginning
and the end, as shown in Fig. 6.
B. X and Y Axes Decoupled Open-Loop Tracking

Fig. 5. (a) Workspace of the 2-DOF mechanism. (b) Cross-axis coupling and
hysteresis behavior.

and (ycoupling /xprim ary ) coupling ratios are found to be 0.91%

and 0.89%, respectively. The hysteresis behavior of the system
in the X-axis is presented in Fig. 5(b).

V. EXPERIMENTAL RESULTS AND DISCUSSION

In order to investigate the tracking performance of a 2-DOF
mechanism, 1-DOF and 2-DOF smooth trajectories are formulated. Open-loop and robust closed-loop control are implemented to track the desired motion trajectories with and without implementing cross-axis coupling compensation. The robust
motion control requires only knowledge of the estimated parameters and their bounds. Gain values are obtained from the desired
parameters. Dynamic compensator and switching function employed are achieved with gain values and system parameters.
In the experimental regime, the 1-DOF desired trajectory is a
jerk-free oscillatory motion defined by polynomials of fourth to
seventh order, with a maximum 20 m displacement. Two-DOF
trajectories consisted of circles of 10 m, 1 m, and 250 nm
radii. The results of tracking these trajectories are discussed at
length in this section.

In order to study the tracking performance and to compensate

for the cross-axis coupling, an open-loop controller is established as follows:
 
 
 


m
blpx x m
klpx xm
mlpx x
vin x
+
+
=
(9)
mlpy ym
blpy y m
klpy ym
vin y
where mlp , blp , and klp are estimated parameters. This openloop controller is implemented only for comparison purposes.
With this open-loop controller, the X-axis tracking error is
found to be less than 2.0 m and the Y -axis cross-coupling
is found to be 225 nm. Open-loop position tracking error is
mostly dominated by the hysteresis of the system and uncertain
system parameters, which are used to construct this controller.
Cross-axis coupling is prominently seen in pure open-loop
tracking. To compensate for the cross-axis coupling, the openloop controller defined in (9) is extended with the addition of
cross-axis coupling ratios to take the crosstalk effect into account
 



Cxx Cxy
vin x
Vin ex
=
(10)
vin ey
Cy x Cy y
vin y
where (cxx = cy y = 1), cxy , and cy x are the experimentally estimated cross-coupling ratios and their corresponding values,
which are found to be 0.91% and 0.89%. The experimental
open-loop tracking with and without cross-axis coupling compensation is shown in Fig. 7.

1742

IEEE/ASME TRANSACTIONS ON MECHATRONICS, VOL. 19, NO. 6, DECEMBER 2014

Fig. 8. Closed-loop tracking of 1-DOF trajectory in X -axis without cross-axis

coupling compensation.

Fig. 7.

Open-loop tracking with cross-coupling compensation.

TABLE III
SUMMARY OF ERRORS FOR OPEN-LOOP TRACKING

Overall open-loop tracking errors are found to be below

2.0 m. With the extended open-loop controller, cross-axis coupling is reduced by almost 70%, i.e., from 225 to 65 nm. The
error summary of open-loop tracking with and without crossaxis coupling compensation is presented in Table III.
C. X and Y Axes Decoupled Robust Motion Closed-Loop
Tracking Without Cross-Axis Coupling Compensation
In the second step, robust motion closed-loop control is established to track the 1-DOF trajectory in the X-axis. At the
same time, Y -axis is kept free from any input and trajectory,
i.e. no cross-axis compensation is applied. Thus, the cross-axis
coupling of Y due to the X-primary motion can be observed.
Robust motion closed-loop control demonstrates high precision
tracking of the 1-DOF trajectory. At same time, Y cross-axis
coupling can be seen in Fig. 8.Closed-loop tracking error is be-

Fig. 9. Closed-loop tracking of 1-DOF trajectory with cross-axis coupling

compensation.

low 50 nm and cross-axis coupling is found to be less than

225 nm.
D. X and Y Axes Decoupled Robust Motion Closed-Loop
Tracking With Cross-Axis Coupling Compensation
As shown in Fig. 8, the cross-axis coupling needs to be
compensated for the effective tracking of a 2-DOF mechanism. Hence, same robust motion control scheme is applied
to the other axis; in this case, Y -axis, to keep it at zero

BHAGAT et al.: EXPERIMENTAL INVESTIGATION OF ROBUST MOTION TRACKING CONTROL FOR A 2-DOF FLEXURE-BASED MECHANISM

1743

TABLE IV
ERROR SUMMARY OF THE ROBUST MOTION CONTROL 1-DOF TRACKING

Fig. 11. Closed-loop tracking of 250-nm radius circle and resulting tracking
errors in X and Y axes.

A summary of the tracking 1-DOF trajectories with and without cross-axis coupling compensation for the X and Y axes is
presented in Table IV.
E. Two-DOF Trajectory Tracking
Fig. 10.

Closed-loop tracking of circle with the radius of 10 and 1 m.

all the time during tracking of X-axis motion. The crossaxis coupling is considered as a known internal disturbance
and compensated with implementing a robust motion control
scheme.
The resulting performance of the 1-DOF tracking with crossaxis coupling compensation is presented in Fig. 9. With the
implementation of the closed-loop control on the Y -axis, the
cross-axis coupling error is reduced to a minimal value of
10 nm, demonstrating effective compensation of the cross-axis
coupling.

The flexure-based mechanism utilized is capable of providing motions in 2-DOF. Hence, to evaluate the effectiveness of
the closed-loop control and to investigate the performance of
the presented compliant mechanism at micro/nanolevel, 2-DOF
motion trajectories are formulated. In this study, only circular
motion trajectory in the X and Y axes, with the radii of 10 m,
1 m, and 250 nm are tracked with the robust motion control.
The results of the tracking are shown in Figs. 10 and 11.
Table V shows the comparative tracking error summary for
2-DOF circular trajectories. Figs. 10 and 11 validate the effective tracking of a 2-DOF compliant mechanism with laser
interferometry-based robust motion control. Experiments to
track 1-DOF and 2-DOF trajectories are repeated for several

1744

IEEE/ASME TRANSACTIONS ON MECHATRONICS, VOL. 19, NO. 6, DECEMBER 2014

TABLE V
ERROR SUMMARY OF THE 2-DOF TRACKING

times. The resulting tracking errors during dynamic motion in

the X and Y axes are always found to be below 20 nm and
steady-state error of 8 nm. The cross-axis coupling motion is
compensated effectively with robust motion control to achieve
high precision tracking of 2-DOF trajectories. The robust motion
closed-loop control effectively tracks the 2-DOF trajectories by
compensating cross-axis coupling.
The overall accuracy of laser interferometry-based system
is affected by a number of factors, which can be divided into
three main categories, such as geometric errors, instrumental errors, and environment errors [30]. Environmental errors are the
largest contributor of the measurement error; hence, controlling and monitoring the environmental conditions will reduce
the measurement errors. Variation in ambient conditions, mechanical vibrations, acoustic noise, thermal fluctuations, manufacturing imperfections, and improper precompression forces
applied at the contact point are some of the major sources causing measurement error. The measurement results can be further
improved with investigating error sources, their contribution to
the overall measurement error, and by implementing effective
counter measures.
VI. CONCLUSION
An advanced research facility utilizing a laser interferometrybased sensing and measurement technique is established for
micro/nanomanipulation. A laser interferometry-based robust
motion closed-loop control is implemented for the effective
tracking of 2-DOF flexure-based mechanism. The workspace
envelope and cross-axis coupling of the presented 2-DOF
mechanism are investigated experimentally. Open-loop performance demonstrates cross-axis coupling and hysteresis behavior of the 2-DOF manipulation system. In the presented 2DOF flexure-based mechanism, the resulting coupling ratios
(xcoupling /yprim ary ) and (ycoupling /xprim ary ) are found to be
0.91% and 0.89%, respectively. Cross-axis coupling presented
in the system is compensated effectively to achieve precise
and accurate tracking. In the robust motion control scheme,
the resulting cross-axis coupling is considered as an internal disturbance. The experimental study demonstrates highprecision planar motion tracking of a 2-DOF flexure-based micro/nanomanipulator. Future work on hysteresis compensation,
experimental system identification, and laser interferometrybased closed-loop controller design to track smooth and nonsmooth motion trajectories is under investigation.

REFERENCES
[1] M. N. M. Zubir, B. Shirinzadeh, and Y. Tian, Development of a novel
flexure-based microgripper for high precision micro-object manipulation,
Sensors Actuators A, Phys., vol. 150, pp. 257266, 2009.
[2] Y. Tian, D. Zhang, and B. Shirinzadeh, Dynamic modelling of a flexurebased mechanism for ultra-precision grinding operation, Precis. Eng.,
vol. 35, pp. 554565, 2011.
[3] S. Polit and J. Dong, Development of a high-bandwidth XY nanopositioning stage for high-rate micro-/nanomanufacturing, IEEE /ASME
Trans. Mechatronics, vol. 16, no. 4, pp. 724733, Aug. 2011.
[4] Y. Li and Q. Xu, A totally decoupled piezo-driven XYZ flexure parallel
micropositioning stage for micro/nanomanipulation, IEEE Trans. Autom.
Sci. Eng., vol. 8, no. 2, pp. 265279, Apr. 2011.
[5] Z. Yan Liang, Z. Yong, R. Changhai, B. K. Chen, and S. Yu, A load-lockcompatible nanomanipulation system for scanning electron microscope,
IEEE/ASME Trans. Mechatronics, vol. 18, no. 1, pp. 230237, Feb. 2013.
[6] M. Zareinejad, S. M. Rezaei, A. Abdullah, and S. S. Ghidary, Development of a piezo-actuated micro-teleoperation system for cell manipulation, Int. J. Med. Robot. Comput. Assisted Surg., vol. 5, pp. 6676,
2009.
[7] J. C. Wang, W. B. Rong, X. X. Li, L. N. Sun, and L. Ma, Integrating
positioning X-Y stage based on bulk micromachining, Guangxue Jingmi
Gongcheng/Opt. Precision Eng., vol. 17, pp. 333340, 2009.
[8] Y. Qin, B. Shirinzadeh, D. Zhang, and Y. Tian, Compliance modeling and
analysis of statically indeterminate symmetric flexure structures, Precis.
Eng., vol. 37, pp. 415424, 2013.
[9] Y. Tian, B. Shirinzadeh, D. Zhang, and Y. Zhong, Modelling and analysis
of a three-revolute parallel micro-positioning mechanism, Proc. Institution Mech. Eng., J. Mech. Eng. Sci., vol. 225, pp. 12731286, 2011.
[10] Y. Qin, B. Shirinzadeh, Y. Tian, and D. Zhang, Design issues in a decoupled XY stage: Static and dynamics modeling, hysteresis compensation,
and tracking control, Sens. Actuators A, Phys., vol. 194, pp. 95105,
2013.
[11] J. X. Zhang and B. Y. Sun, Design and analysis of 2-DOF nanopositioning stage based on dual flexure hinges, Yadian Yu Shengguang/
Piezoelect. Acoustoopt., vol. 28, pp. 624626, 2006.
[12] W. Wen, Z. Zhu, and C. Zichen, Design and analysis of two dimensional
X-Y micropositioning stage, Proc. SPIE, vol. 7544, art. no. 75441I, 7
pp., 2010.
[13] D. Kang, K. Kim, Y. Choi, D. Gweon, S. Lee, and M. Lee, Design
and control of flexure based XYzstage, J. Mech. Sci. Technol., vol. 19,
pp. 21572164, 2005.
[14] W. Rong, L. Ma, L. Sun, H. Xie, and L. Chen, Analysis and optimum design of 2-DOF micro-positioning stage, Jixie Gongcheng Xuebao/Chin.
J. Mech. Eng., vol. 42, pp. 2630, 2006.
[15] Q. Xu, Y. Li, and N. Xi, Design, fabrication, and visual servo control
of an xy parallel micromanipulator with piezo-actuation, IEEE Trans.
Autom. Sci. Eng., vol. 6, no. 4, pp. 710719, Oct. 2009.
[16] L. Yangmin and X. Qingsong, A novel piezoactuated XY stage with parallel, decoupled, and stacked flexure structure for micro-/nanopositioning,
IEEE Trans. Ind. Electron., vol. 58, no. 8, pp. 36013615, Aug. 2011.
[17] H. C. Liaw and B. Shirinzadeh, Constrained motion tracking control of
piezo-actuated flexure-based four-bar mechanisms for micro/nano manipulation, IEEE Trans. Autom. Sci. Eng., vol. 7, no. 3, pp. 699705, Jul.
2010.
[18] Q. Xu and Y. Li, Precise tracking control of a piezoactuated micropositioning stage based on modified Prandtl-Ishlinskii hysteresis model, in
Proc. IEEE Conf. Autom. Sci. Eng., 2010, pp. 692697.
[19] H. C. Liaw, B. Shirinzadeh, and J. Smith, Enhanced sliding mode motion
tracking control of piezoelectric actuators, Sens. Actuators A, Phys.,
vol. 138, pp. 194202, 2007.
[20] Y. Chen, J. Qiu, H. Ji, and K. Zhu, Tracking control of piezoelectric
actuator system using inverse hysteresis model, Int. J. Appl. Electromagn.
Mech., vol. 33, pp. 15551564, 2010.
[21] Y. Qin, Y. Tian, D. Zhang, B. Shirinzadeh, and S. Fatikow, A novel direct
inverse modeling approach for hysteresis compensation of piezoelectric
actuator in feedforward applications, IEEE/ASME Trans. Mechatronics,
vol. 18, no. 3, pp. 981989, Jun. 2013.
[22] Q. Yao, J. Dong, and P. M. Ferreira, Design, analysis, fabrication and
testing of a parallel-kinematic micropositioning XY stage, Int. J. Mach.
Tools Manuf., vol. 47, pp. 946961, 2007.
[23] S. Awtar, J. Ustick, and S. Sen, An XYZ parallel-kinematic flexure mechanism with geometrically decoupled degrees of freedom, J. Mechanisms
Robot., vol. 5, pp. 015001015001, 2012.

BHAGAT et al.: EXPERIMENTAL INVESTIGATION OF ROBUST MOTION TRACKING CONTROL FOR A 2-DOF FLEXURE-BASED MECHANISM

[24] T. Xueyan, I. M. Chen, and Q. Li, Design and nonlinear modeling of

a large-displacement XYZ flexure parallel mechanism with decoupled
kinematic structure, Rev. Sci. Instrum., vol. 77, p. 115101, 2006.
[25] H. Wang and X. Zhang, Input coupling analysis and optimal design of
a 3-DOF compliant micro-positioning stage, Mechanism Mach. Theory,
vol. 43, pp. 400410, 2008.
[26] B. J. Kenton and K. K. Leang, Design and control of a three-axis serialkinematic high-bandwidth nanopositioner, IEEE/ASME Trans. Mechatronics, vol. 17, no. 2, pp. 356369, Apr. 2012.
[27] Y. Qin, B. Shirinzadeh, Y. Tian, D. Zhang, and U. Bhagat,Design
and computational optimization of a decoupled 2-DOF monolithic
mechanism, IEEE/ASME Trans. Mechatronics, [Online]. Available:
http://ieeexplore.ieee.org. DOI: 10.1109/TMECH.2013.2262801
[28] H. C. Liaw, B. Shirinzadeh, and J. Smith, Robust motion tracking control
of piezo-driven flexure-based four-bar mechanism for micro/nano manipulation, Mechatronics, vol. 18, pp. 111120, 2008.
[29] U. Bhagat, B. Shirinzadeh, Y. Tian, and D. Zhang, Experimental analysis
of laser interferometry based robust motion tracking control of a flexurebased mechanism, IEEE Trans. Autom. Sci. Eng., vol. 10, no. 2, pp. 267
275, Apr. 2013.
[30] U. Bhagat, B. Shirinzadeh, and Y. Tian, Experimental study of laser
interferometry based motion tracking of a flexure-based mechanism, Int.
J. Intell. Mechatron. Robot., vol. 1, no. 3, pp. 3145, 2011.

Umesh Bhagat (S13) received the B.Eng. degree

from the University of Pune, Maharashtra, India, in
2003, and the Master of Computer Science and IT
degree from the University of New South Wales, Sydney, Australia, in 2008. He is currently working toward the Ph.D. degree in the Department of Mechanical and Aerospace Engineering, Monash University,
Melbourne, Australia.
He possesses ten years experience in industrial automation and control system design. He has worked
for major automotive giants such as TATA Motors
and General Motors Corporation as a Technical Lead in the area of robotics
and plant floor systems. His research interests include micro/nanomanipulation,
laser-based measurements and sensory-based controls, robotics, and automated
flexible manufacturing.

Bijan Shirinzadeh (M06) received the B.E. and

M.S.E. degrees in mechanical and aerospace engineering from the University of Michigan, Ann Arbor,
MI, USA, and the Ph.D. degree in mechanical engineering from the University of Western Australia,
Perth, Australia, in 1990.
He has held various positions in academia and industry. He is currently a Professor and the Director of
the Robotics and Mechatronics Research Laboratory,
Department of Mechanical and Aerospace Engineering, Monash University, Melbourne, Australia. His
current research interests include laser-based measurements and sensory-based
control, micro/nanomanipulation mechanisms and control, systems kinematics
and dynamics, haptics, robotic assisted surgery and microsurgery, and advanced
manufacturing and automated sciences.

Leon Clark (S13) received the B.Eng. (Hons.) and

B.Sc. degrees from Monash University, Melbourne,
Australia, in 2011. Currently, he is working toward
the Ph.D. degree in the Robotics and Mechatronics
Research Laboratory, Department of Mechanical
and Aerospace Engineering, Monash University.
His research interests include micro/
nanomanipulation
mechanisms,
design
and
analysis of compliant structures, high-precision
measurement systems, and real-time control.

1745

Yanding Qin received the B.Eng. and M.Sc. degrees

in industrial design and the Ph.D. degree in mechanical engineering from Tianjin University, Tianjin, China, in 2005, 2007, and 2012, respectively.
From September 2009 to August 2010, he was a
Visiting Scholar in the School of Mechanical Engineering, Purdue University, West Lafayette, IN, USA.
From May 2012 to June 2013, he was a Postdoctoral Research Officer in the Robotics and Mechatronics Research Laboratory, Department of Mechanical and Aerospace Engineering, Monash University,
Melbourne, Australia. He is currently a Lecturer with the Institute of Robotics
and Automatic Information System, Nankai University, Tianjin, China. His research interests include flexure-based mechanisms, micro/nanomanipulation,
hysteresis modeling and compensation, laser-based measurement, mechanical
dynamics, and real-time control methodology.

Yanling Tian received the B.Eng. degree in mechanical engineering from Northwest Institute of Light
Industry, Xianyang, China, in 1997, and the M.Sc.
and Ph.D. degrees in mechanical engineering from
Tianjin University, Tianjin, China, in 2002 and 2005,
respectively.
From 2005 to 2006, he was a Postdoctoral Research Fellow, and currently holds the position of
Associate Professor at Tianjin University. From 2007
to 2009, he was a Postdoctoral Research Fellow in the
Robotics and Mechatronics Research Laboratory, Department of Mechanical and Aerospace Engineering, Monash University, Australia. He was also a Visiting Scholar at Hong Kong University of Science and
Technology, China, and the University of Warwick, U.K., in 2001 and 2006, respectively. He was a Visiting Professor at Tohoku University, Japan, in 2010. He
obtained the prestigious Alexander von Humboldt Fellowship for experienced
researchers in 2010. His research interests include micro/nanomanipulation,
mechanical dynamics, finite-element method, and surface metrology and characterization. He was awarded and completed a number of government and
industry-based projects, and has published more than 50 peer-reviewed papers.

Dawei Zhang received the B.Eng. degree in mechanical engineering from Shenyang University of Technology, Shenyang, China, in 1984, and the M.Sc.
and Ph.D. degrees in mechanical engineering from
Tianjin University, Tianjin, China, in 1990 and 1995,
respectively.
From 1984 to 1987, he was an Associate Engineer at Tianjin Institute of Power Source, China. He
is currently a Professor of the School of Mechanical
Engineering, Tianjin University. His current research
interests include micro/nanopositioning techniques,
structural dynamics, signal processing, high-speed machining methodologies,
and design and dynamics of machine tools.