See discussions, stats, and author profiles for this publication at: [Link]

net/publication/235004205

Continuous Time Controller Based on SMC and Disturbance Observer for

Piezoelectric Actuators

Article in International Review of Electrical Engineering (IREE) · November 2007

CITATIONS READS

2 239

2 authors:

Selim Yannier Asif Sabanovic

TOGG (Türkiye'nin Otomobili Girişim Grubu) Sabanci University
23 PUBLICATIONS 30 CITATIONS 320 PUBLICATIONS 4,361 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Asif Sabanovic on 31 May 2014.

The user has requested enhancement of the downloaded file.

 International Review of Electrical Engineering (I.R.E.E.), Vol.03, n. 6 December 2007

Continuous Time Controller Based on SMC and Disturbance Observer

for Piezoelectric Actuators

S. Yannier1, A. Sabanovic2

Abstract – In this work, analog application for the Sliding Mode Control (SMC) to piezoelectric
actuators (PEA) is presented. DSP application of the algorithm suffers from ADC and DAC
conversions and mainly faces limitations in sampling time interval. Moreover piezoelectric
actuators are known to have very large bandwidth close to the DSP operation frequency.
Therefore, with the direct analog application, improvement of the performance and high
frequency operation are expected. Design of an appropriate SMC together with a disturbance
observer is suggested to have continuous control output and related experimental results for
position tracking are presented with comparison of DSP and analog control application.

Keywords: Sliding Mode Control, Disturbance Observer, Piezoelectric Actuator, Analog

Electronics

nondifferential, multivalued, and is usually unknown.

The existence of hysteresis often limits the performance
I. Introduction of the piezoelectric actuator, leads to the severe
Piezoelectric effect [1] is a crystalline effect and inaccuracies (up to 10%-15% of the traveling path) and
therefore piezoelectric actuators do not suffer from causes undesirable oscillation or even instabilities when
“stick slip” effect mainly caused by the friction between the piezo actuator is operated in an open loop manner
elements of a mechanical system. This property [14] Achieving high speed, large-range precision
theoretically offers an unlimited resolution and positioning of piezo actuators is therefore challenging
therefore the use in many applications to provide sub- [15], hence how to design an effective controller for
micrometer resolution; ultrasonic motors, sports dealing with the hysteretic feature becomes a very
materials like skis and bikes [2], aerospace [3], hard important topic.
disk drives [4] etc… Other main application of these Another undesired characteristic of piezoelectric
ceramics is the scanning tunneling microscope (STM) actuators is the “creep effect”, that can be observed
and atomic force microscope (AFM)[5]. However, the when a step input voltage is applied to the input of the
achievable resolution in practice can be limited by a actuator [16, 17]. Figure 1 shows a step response of a
number of other factors such as the piezo control general stack-type piezoelectric actuator. It has been
amplifier (electronic noise), sensor (resolution, noise known that the creep response has a logarithmic shape
and mounting precision) and control electronics (noise over time that can be represented by the following
and sensitivity to EMI). equation [16]:
In naturally occurring piezoelectric materials, such as
quartz, piezoelectric effect is too small to be of practical
use. Modern man-made piezoelectric polycrystalline
ceramics, such as lead zirconate titanate (PZT),
lanthanum modified lead zirconate titanate (PLZT) and
polyvinylidene fluoride (PVDF), are much more
suitable for actuator purposes [1, 6].
Beside their high speed, high bandwidth, high
stiffness, high electrical-mechanical transformation
efficiency and little heat generation properties making
piezoelectric materials agreeable actuators, all
piezoelectric materials are ferroelectric and as all
ferroelectric materials they exhibit fundamental Fig. 1: Open loop step response of the piezoelectric
hysteresis phenomena in the polarization versus the actuator. “d” is the displacement difference due to
hysteresis. Creep effect is magnified in the circular
applied electric field, as well as in all the material view [16].
properties coupled to the polarization [4-12].
Hysteresis yields a rate-independent lag and residual Piezoelectric actuators shows dominant capacitive
displacement near zero input [13]. Hysteresis is behavior and therefore can be controlled with either

Manuscript received January 2007, revised January 2007 Copyright © 2007 Praise Worthy Prize S.r.l. - All rights reserved
 S. Yannier, A. Sabanovic

voltage (voltage control) or current input (charge Woronko et al. used SMC to control additional
control). degree of freedom added to tool position on
Some charge control methods are investigated by conventional CNC turning centers to improve
researchers. In [18], a current source is directly used for machining precisions [27]. Similarly Chiang et al.
charge steering, while in [19] a simple configuration is worked on the control of a piezoelectric actuator
proposed consisting of a voltage source, an operational mounted on the tip of a pneumatic cylinder. That is how
amplifier, an external capacitance, and a high voltage they obtained a large stroke yet precise linear actuation
amplifier. It is known that this simple configuration has system using adaptive discrete variable structure
two apparent drawbacks: both sides of the piezoelectric control.
actuator are floating with respect to ground and the Abidi et al. used SMC in conjunction with the
configuration is very sensitive to op-amp bias current disturbance observer for both position and force
[10]. tracking in piezoelectric actuators [28, 29]. In their
In voltage steering on the other hand, much effort is work, they used the lumped parameter model of the
done to model the hysteresis with invertible functions piezoelectric actuator to estimate the external
that will be used for compensation purposes [5, 10, 12, disturbances via disturbance observer. Feedforward
13, 20, 21 Lining, 2004 #35]. The choice of open loop application of the estimated disturbance improved the
control is generally shown as a “must” since the results.
position change measurement in the orders of The aim of this work is to design appropriate control
micrometers requires expensive devices. However, to drive piezoelectric actuators on “analog control
open-loop techniques have been successful in providing circuit” to improve the tracking performance of
results up to 2% error which is acceptable compare to piezoelectric actuators.
the 15-20% hysteresis error in the nature of the The rest of the paper is organized as follows. Section
piezoelectric actuator. Still due to the difficulties II describes the controller design with disturbance
involved in modeling the actuator precisely, those observer and plat model. The analog controller
techniques seem to approach the boundaries. electronics that is designed is presented in Section III.
In [4], disturbance compensation based on a Section IV presents the experimental results while
hysteresis model is used. However, unmodeled conclusions and areas for future research are presented
H in Section V.
disturbances required the addition of a robust ∞
controller. A similar work is done by Tamer and Dahleh
II. Sliding Mode Control
[22]. They tried both the H ∞ method with estimated
velocity. Experimental results show that although the In this paper we will consider dynamical systems that
error boundaries decreased from 8% to 6%, hysteresis is can be represented as a class of nonlinear systems linear
H with respect of control as described by the following
still a problem even at low frequencies. ∞ control on
equation
the other hand gives better tracking and eliminates the x& = f ( x) + B( x)u + d (1)
high frequency oscillations for the cost of rounding the
corners of the triangular waves and noticeable delay where x T ∈ ℜ n is the state vector, u ∈ ℜ m is the
increasing with frequency. control vector, f ( x) ∈ ℜ n is an unknown, continuous
In order to design a control scheme that will achieve
successful tracking performance without precise and bounded nonlinear function, B( x) ∈ ℜ nxm is a
dynamic modeling, some fuzzy logic and neural known input matrix whose elements are continuous and
network solutions are presented in the literature. bounded and rank (B(x )) ∀x = m , with d ∈ ℜ n being an
However, due to the limited performance, this research
unknown, bounded external disturbance. Both
area did not find much popularity [23].
Originally designed as system motion for dynamic f ( x) ∈ ℜ n and d ∈ ℜ n satisfy the matching conditions
systems whose essential open-loop behavior can be and all their components are bounded f i ( x) ∀x
≤M
sufficiently modeled with ordinary differential
and d i (t ) ≤ N . Fully actuated mechanical systems
equations, Sliding Mode Control (SMC) is one of the ∀t
effective nonlinear robust control approaches that belong to the class of systems described by (1). Such
provides system invariance to uncertainties once the systems can be interpreted as m interconnected sub-
system is in the sliding mode [24, 25]. ( )
systems &x&i = hi (xi ,x& i ) + bi (xi ,t ) ⋅ u i + g i xi ,x j , hi (xi ,x&i )
Bonnail et al. applied SMC on piezoelectric actuated
scanning tunneling microscope to precisely follow the
in general represents Coulomb friction term, g i xi ,x j ( )
sample surface with the feedback of the tunneling represents the interaction term and is regarded as a
current [26]. Compare to the commercial PI controlled disturbance.
motion, their solution shows less oscillating tunneling The aim is to determine the control input
current due to the better tracking of the surface. u = [u1 ,..., u m ]T such that the system states

Copyright © 2007 Praise Worthy Prize S.r.l. - All rights reserved International Review of Electrical Engineering,
Vol. xx, n. x
 S. Yannier, A. Sabanovic

x1 (t ),..., xn (t ) track the desired trajectories discrete time implementation the sliding mode is
xd1 (t ),..., xd n (t ) while control error satisfies selected guarantied with continuous control action. We are
targeting analog application for which controller will be
dynamical constraints. implemented in continuous time domain so in our
σ
application the discontinuous term μ will be
II.1. Controller Design σ Tσ
The controller will be designed in the SMC omitted and we will be determining the control action
framework by firstly selecting a suitable sliding that satisfies conditions (σ& + Dσ ) = 0 but all further
manifold that will ensure desired systems dynamics and analysis can be easily adopted for application of
then selecting control such that the Lyapunov stability expression (6) if the term Dσ is replaced with
conditions are satisfied. Selecting the Lyapunov (Dσ + μ σ σ T σ ) .
function candidate in terms of the sliding function is a For system (1) with sliding mode manifold (2) the
natural way of guaranteeing the sliding mode existence control that satisfies (σ& + Dσ ) = 0 can be determined as
on the selected manifold and thus having desired closed
loop dynamics. Finally, the necessary control input ((
u = −(GB )−1 G f + d − x& di − Dσ ) )
should be selected that will fulfill the requirements of (7)
= u eq + (GB ) Dσ −1
the Lyapunov stability criteria.
Sliding Manifold
where [
xd = xd1 ,....., xd n ] and u eq is so-called

For system (1) the natural selection of the sliding equivalent control obtained as a solution of the equation
manifold is in the following form ( )
σ& = GB u eq − u = 0 . By substituting (7) into (1) the
σ = Get = 0 , (2) equations of motion of system (1) in manifold (2) are
where tracking error vector is defined as obtained as σ = Get = 0 and the approach to this
et = [e1 ,..., en ]T ∈ ℜ n , ei = xdi − xi and the sliding solution is governed by equation (6). This is a result of
the specific structure of the plant (1) in which states is
surface satisfies σ = [σ i ,...,σ m ]T ∈ ℜ m , G ∈ ℜ mxn . selected as the derivatives of the measurable outputs
and each sub-block is represented in the canonical form.
Computing the Necessary Control Input To implement this control input, information about
A Lyapunov Function candidate can be selected as the plant dynamics and external disturbances are
1 needed, which is hard to achieve. Hence, this solution
V = σ Tσ (3)
2 needs the information on the equivalent control thus
where, V ∈ ℜ . This function can also be stated as may be applied for the plants when u eq is known or can
V = (1 / 2) ⋅ σ
2 be estimated with sufficient accuracy. In this paper we
2
, where • 2
indicates Euclidian norm
will be using a fact proven in [32, 33] that the solution
with V (0 ) = 0 . The time derivative of the candidate of the differential equation
Lyapunov function V& should be negative definite. In (
τ z& + z = u − (GB )−1 ⋅ σ& )
(8)
order to use this condition in selection of the control, we with small enough filtering time constant τ is close
may require that the V& satisfies some preselected form. to the equivalent control. In this paper we will be using
Equating the time derivative of this function to a this result in order to avoid direct calculation of the
negative definite function like in (4),
equivalent control from u eq = −(GB )−1 G ( f + d − x& d )
σ
V& = −σ T Dσ − μ T (4)
σ σ but instead to use approximated result u eq = z .
where D is a positive definite symmetric matrix, and Equation (8) can be used together with the control
μ > 0 thus Lyapunov conditions are satisfied. By presented in equation (7) to obtain the simplified
substituting (3) into (4), the following requirement is controller equation;
found. u eq
+ (GB ) Dσ
−1
u=
⎛ σ ⎞ τ ⋅ s +1
σ T ⎜ σ& + Dσ + μ T ⎟ = 0 (5) (9)
⎝ σ σ⎠ u ⎛ sσ ⎞
= + K ⋅ ⎜ Dσ + ⎟
Therefore, for σ ≠ 0 , the control law can be τ ⋅ s +1 ⎝ τ ⋅ s +1⎠
calculated by satisfying the following equation. Do not use Laplace “s” and derivative “d/dt0dot” in
⎛& σ ⎞ the same equation
⎜ σ + Dσ + μ T ⎟ = 0 (6)
⎝ σ σ⎠ where K = (GB )−1 .
and the sliding mode conditions are satisfied. The
discontinuous term can be selected as small in order to
avoid chattering. It had been proven [30, 31] that in the

Copyright © 2007 Praise Worthy Prize S.r.l. - All rights reserved International Review of Electrical Engineering,
Vol. xx, n. x
 S. Yannier, A. Sabanovic

II.2. Piezoelectric Actuator Model Fd = TN ⋅ h + ΔT (h + u ) + Δm ⋅ &x&

(13)
In this work a piezoelectric actuator (PEA) that + Δc ⋅ x& + Δk ⋅ x& + c N x& + k N x
consists of a piezo-drive integrated with a sophisticated is the disturbance on the system.
flexure structure for motion amplification is used. The For simplicity the disturbance observer proposed by
flexure structure is wire-EDM-cut and is designed to Ohnishi et al. can be used [35]. The derivation of the
have zero stiction and friction. observer is as follows. From equation-b the disturbance
In addition to the absence of internal friction, flexure for is
stages exhibit high stiffness and high load capacity. F d = T N ⋅ u − m ⋅ &x& (14)
Flexure stages are also insensitive to shock and
vibration. However, since the piezo-drive exhibits non- Ohnishi et al. proposes that;
linear hysteresis behavior, the overall system will also
exhibit the same behavior.
Fˆd =
g
s+g
(
⋅ TN ⋅ u − m ⋅ s 2 x (15))
The dynamics of the piezo-stage can be represented
where F̂d ⋅ is the observed disturbance force, g is
by the following second-order differential equation
coupled with hysteresis in the presence of external the constant determining the corner frequency of the
forces first order filter. Since x and u are measurable, û , the
meff &x& + c eff x& + k eff x = T ⋅ [u (t ) − h( x, u )] − Fext (10) correction to the control output, can then be calculated
as
where x denotes the displacement of the stage,
Fˆ g ⎛ m 2 ⎞
meff , ceff and k eff denotes the effective mass, uˆ = d = ⋅ ⎜⎜ u − ⋅ s x ⎟⎟ (16)
TN s + g ⎝ TN ⎠
effective damping and effective stiffness of the stage
respecively, T denotes the electromechanical The same remark as earlier
transformation ratio, u denotes the input voltage and
h(x, u ) denotes the unknown, bounded, non-linear III. Analog Circuit Design
hysteresis that has been found to be a function of x and
u , and Fext is the external force acting on the stage
[12, 34]. [Link] Mode Controller
The model represented by (10) shows that the Following the derivation in section II.1 above, the
hysteresis may be perceived as a disturbance force that analog controller should calculate the control u from
satisfies matching conditions. This means that the the two inputs; desired trajectory xd and the actual
sliding mode based control should be able to reject the
trajectory x . Selecting the sliding manifold as;
influence of the hysteresis nonlinearity on the
mechanical motion. At the same time it is obvious that ⎡e ⎤
σ = G ⋅ et = G ⋅ ⎢ 1 ⎥ = C ⋅ e + e& = 0 (17)
the lumped disturbance consisting of the external force ⎣e 2 ⎦
acting on the system and the hysteresis can be
where tracking error vector is defined as
estimated, thus allowing the application of the
disturbance rejection method in the overall system et = [e1 e2 ] ∈ ℜ
T 2
with e1 = e = x d − x and
design. e2 = e&1 = e& = x& d − x& . Accordingly the control u is
calculated as below
II.3. Disturbance Observer ⎛ sσ ⎞
u = u~ + K ⋅ ⎜ Dσ + ⎟ (18)
The structure of the observer is based on (10) under ⎝ τ ⋅ s +1⎠
the assumption that all the plant parameter uncertainties, u
nonlinearities and external disturbances can be where u~ = . Equation (2) can further be
τ ⋅ s +1
represented as a lumped disturbance. To show that integrated to this last equation for simplification.
assume all plan parameters have nominal values
⎛
u = u~ + K ⋅ ⎜ D ⋅ (Ce + e& ) +
(Ce& + e&&) ⎞
denoted with subscript N and uncertainties shown with ⎟ (19)
Δ•. ⎝ τ ⋅ s +1 ⎠
meff = m N + Δm ceff = c N + Δc The same remark as earlier
(11) Only for very small time constants τ , this last
k eff = k N + Δk T = T N + ΔT
equation can further be simplified to obtain,
As obvious, x the displacement of the plant and u u = u~ + K ⋅ (D ⋅ C ⋅ e + (D + C ) ⋅ e& + &e&)
the input to the plant are measurable. Hence, the (20)
= u~ + K ⋅ ϕ
nominal structure of the plant is found as follows
m N &x& = TN ⋅ u − Fd (12) where ϕ = D ⋅ C ⋅ e + (D + C ) ⋅ e& + e&& . From this last
where; equation the algorithm can be deduced immediately.
Fig. 2 depicts this algorithm as functional blocks of the

Copyright © 2007 Praise Worthy Prize S.r.l. - All rights reserved International Review of Electrical Engineering,
Vol. xx, n. x
 S. Yannier, A. Sabanovic

analog controller designed. The algorithm of the observer is also straight forward
as shown if Fig. 4. The observers output is added to the
last summation block before high voltage amplifier
shown in Fig. 2.

Fig. 3: Block-Diagram representation of the

controller and disturbance observer.

This diagram is wrong. The input to the observer is

after summation not before

Fig. 4: Disturbance observer algorithm for analog

application.

IV. Experimental Results

IV.1. Experimental Setup

For experimental purposes, the setup shown in Fig. 5
is constructed; high voltage amplifier is built using
MP108 power operational amplifier from Apex
Fig. 2: Summary of the analog circuit. Microtechnology, PEA is the piezoelectric actuator with
embedded strain gage for position measurement and the
The circuit is based on mathematical operations like strain gage amplifier is the SCM5B38-03 wide band
summation, differentiation and will be realized with op- strain gage amplifier from Dataforth Corporation. Here
amp circuits. Therefore the algorithm also includes the SMC is the designed sliding mode control algorithm
inverting behavior of the op-amps. implemented in DSP (for DSP experiments) or is the
Accordingly the calculation starts with the error analog circuit (for analog controller experiments).
calculation of the error e and then two successive
differentiation blocks to calculate e& and e&& . Then the
calculation of the intermediate variable ϕ is just the
summation of those variables with appropriate
coefficients.
Finally the control is summed up with the filtered
control u~ and disturbance observers output û before
fed to the high voltage amplifier (H.V.).

[Link] Observer
The designed disturbance observer can simply be Fig. 5: Piezoelectric actuator control setup.
integrated to the system to use the actual position x and Both in DSP and analog control experiments, the
control u as inputs and calculate the correction term to structure of the high voltage amplifier assumed to be
the control output; û . The interconnection is depicted limited with a low pass filter.
in Fig. 3. The data is captured by Agilent Technologies

Copyright © 2007 Praise Worthy Prize S.r.l. - All rights reserved International Review of Electrical Engineering,
Vol. xx, n. x
 S. Yannier, A. Sabanovic

54622D digital oscilloscope. The reference and actual error corresponding to only 0.14%. Compare to 2.5%
signals are given without offset to have better feeling on tracking error of the DSP implementation, this result is
the tracking error. The third channel shown in some almost 18 times better.
figures is the error signal. Peak to peak values are given
at the image captions in metric correspondents: Reference & Actual
17.96um (micrometers) corresponds to 1V of the strain 8.98um/Div
gage amplifier reading, or in other words 1um position
deflection results 55.68mV.

IV.2 Position Tracking Experiments Using DSP

For comparison of the results DSP application of the Error
control is realized on dSpace DS1102 platform which 90nm/Div
possesses TMS320C31 DSP chip running at 40 MHz
with 50ns cycle time. The platform does have two 16- Fig. 8: Tracking of 35.69um pp 1Hz sinusoidal
bit ADC (Input) ±10V and four 12-bit DAC (Outputs) reference. The error is 50nm pp (0.14%).
±10V. The algorithm runs at 10kHz. discretization is
made based on Euler’s method. Tracking of a single sinusoidal pulse of period 1
second and peak to peak amplitude 35.60um is shown
on Fig. 9. According to the experiment the peak tp peak
Reference & Actual
error value is 54nm corresponding to 0.15% tracking
900nm/Div
error.
Triangular wave shapes are also studied. As an
example tracking of 35.60um pp 1Hz triangular
reference is presented on Fig. 10. The resulting peak to
peak error is 54nm pp corresponding to 0.15%.
Error Triangular wave shape are in particular importance
90nm/Div
since they constitutes non-continuous references.

Fig. 6: DSP tracking experiment for 4.5um pp 1Hz Reference & Actual
sinusoidal reference.
8.98um/Div
Position tracking of 1Hz sinusoidal inputs is studied.
First 4.5um peak to peak and then 10.8um peak to peak
inputs are tested. Results are shown on Fig. 6 and on
Fig. 7 respectively. The errors for comparison are Error
110nm and 200nm, corresponding to 2.5% and 1.9% 90nm/Div
respectively.

Reference &
Actual Fig. 9: Tracking of 35.60um pp 1Hz sinusoidal
reference. The error is 54nm pp (0.15%).
1800nm/Div

Reference & Actual

8.98um/Div
Error
180nm/Div

Error
Fig. 7: DSP tracking experiment for 10.8um pp 1Hz 90nm/Div
sinusoidal reference.

IV.3. Analog Circuit, Position Tracking Experiments Fig. 10: Tracking of 35.60um pp 1Hz triangular
reference. The error is 54nm pp (0.15%).
Similar experiments are conducted for analog circuit
To present the tracking of an abstract but continuous
realization of the SMC with disturbance observer. The
wave form the tracking of human beat signal is
tracking of a 35.69um peak to peak 1Hz sinusoidal
presented on Fig. 11. For this waveform with peak to
reference is resulted with 50nm peak to peak tracking
peak amplitude of 21.70um the tracking resulted with

Copyright © 2007 Praise Worthy Prize S.r.l. - All rights reserved International Review of Electrical Engineering,
Vol. xx, n. x
 S. Yannier, A. Sabanovic

46nm peak to peak tracking error corresponding to [9] K. Abidi, "Sliding-Mode control for high-precision motion control
systems / Thesis Advisor Asif Sabanovic." Istanbul: Sabanci
0.21%.
University, 2004.
[10] H. J. M. T. S. Adriaens, W. L. De Koning, and R. Banning,
Reference & Actual "Modeling piezoelectric actuators," Mechatronics, IEEE/ASME
Transactions on, vol. 5, pp. 331-341, 2000.
8.98um/Div
[11] R. Banning, W. L. de Koning, H. J. M. T. A. Adriaens, and R.
K. Koops, "State-space analysis and identification for a class of
hysteretic systems," Automatica, vol. 37, pp. 1883-1892, 2001.
[12] M. Goldfarb and N. Celanovic, "Modeling piezoelectric stack
actuators for control of micromanipulation," Control Systems
Error
Magazine, IEEE, vol. 17, pp. 69-79, 1997.
90nm/Div [13] J.-J. Tzen, S.-L. Jeng, and W.-H. Chieng, "Modeling of
piezoelectric actuator for compensation and controller design,"
Precision Engineering, vol. 27, pp. 70-86, 2003.
[14] C.-H. Yang and K.-M. Chang, "Adaptive Neural Network
Control for Piezoelectric Hysteresis Compensation in A Positioning
Fig. 11: Tracking of human hearth beat signal. The System," presented at Industrial Electronics, 2006 IEEE International
peak o peak value is 21.70um. The error is 46nm pp Symposium on, 2006.
(0.21%). [15] Y. Wu and Q. Zou, "Iterative control approach to compensate
for the hysteresis and the vibrational dynamics effects of piezo
actuators," presented at American Control Conference, 2006, 2006.
[16] H. Jung, H. Jung, J. Y. Shim, J. Y. Shim, D. Gweon, and D.
V. Conclusion Gweon, "New Open-loop Actuating Method of Piezoelectric
Actuators for Removing Hysteresis and Creep," Review of Scientific
In this paper we have formulated a SMC with Instruments, vol. 71, 2000.
continuous control output, based on [28, 30, 32, 33], [17] Y.-C. Yu and M.-K. Lee, "A dynamic nonlinearity model for a
applicable in analog electronics for the position tracking piezo-actuated positioning system," presented at Mechatronics, 2005.
of the piezoelectric actuators. This SMC is supported by ICM '05. IEEE International Conference on, 2005.
[18] C. Newcomb and I. Flnn, "Improving the Linearity of
a disturbance observer as introduced in [35].
Piezoelectric Ceramic Actuators," Electronics Letters, vol. 18, pp.
The use of such a controller in systems controlled by 442-444, 1982.
digital controllers, including DSP, PC, microchip and [19] R. H. Comstock, "Charge Control of Piezoelectric Actuators to
FPGA, will help users to save from heavy Reduce Hysteresis Effects." U.S. Patent 4,263,527: The Charles Stark
computational load. Moreover, due to the operation Draper Laboratory, Inc. (Cambridge, MA), 1981.
[20] D. Song and C. J. Li, "Modeling of piezo actuator's nonlinear
frequency limitation in digital systems mainly due to the and frequency dependent dynamics," Mechatronics, vol. 9, pp. 391-
analog to digital conversion, the controller is expected 410, 1999.
to have better performance than the DSP application. [21] K. Uchino, Piezolectric Actuators and Ultrasonic Motors.
Experimental results proved that the analog Boston ; Dordrecht: Kluwer Academic Publishers, 1997.
[22] N. Tamer and M. Dahleh, "Feedback control of piezoelectric
production of the proposed SMC is possible and that the tube scanners," presented at Decision and Control, 1994., Proceedings
controller can track a reference signal with a good of the 33rd IEEE Conference on, 1994.
degree of accuracy. [23] R.-J. Wai and J.-D. Lee, "Intelligent motion control for linear
piezoelectric ceramic motor drive," Systems, Man and Cybernetics,
Part B, IEEE Transactions on, vol. 34, pp. 2100-2111, 2004.
[24] K. D. Young, V. I. Utkin, and U. Ozguner, "A control engineer's
References guide to sliding mode control," Control Systems Technology, IEEE
[1] A. Suleman and A. P. Costa, "Adaptive control of an aeroelastic Transactions on, vol. 7, pp. 328-342, 1999.
flight vehicle using piezoelectric actuators," Computers & Structures, [25] K. D. Young, V. I. Utkin, and U. Ozguner, "A control engineer's
vol. 82, pp. 1303-1314, 2004. guide to sliding mode control," presented at Variable Structure
[2] S. Ashley, "Digital dampers," in Mechanical Engineering, vol. Systems, 1996. VSS '96. Proceedings., 1996 IEEE International
120: American Society of Mechanical Engineers, 1998, pp. 130. Workshop on, 1996.
[3] "Shape shifting aircraft," in Design Engineering: Centaur [26] N. Bonnail, D. Tonneau, F. Jandard, G.-A. Capolino, and H.
Communications, 2001, pp. 5. Dallaporta, "Variable Structure Control of a Piezoelectric Actuator for
[4] B. M. Chen, T. H. Lee, C.-C. Hang, Y. Guo, and S. Weerasooriya, a Scanning Tunneling Microscope," in IEEE Transactions on
"An Hinf almost disturbance decoupling robust controller design for a Industrial Electronics, vol. 51, 2004, pp. 354-363.
piezoelectric bimorph actuator with hysteresis," Control Systems [27] A. Woronko, J. Huang, and Y. Altintas, "Piezoelectric tool
Technology, IEEE Transactions on, vol. 7, pp. 160-174, 1999. actuator for precision machining on conventional CNC turning
[5] H. Richter, E. A. Misawa, D. A. Lucca, and H. Lu, "Modeling centers," in Precision Engineering, vol. 27, 2003, pp. 335.
nonlinear behavior in a piezoelectric actuator," Precision Engineering, [28] K. Abidi, A. Sabanovic, and S. Yesilyurt, "Sliding mode control
vol. 25, pp. 128-137, 2001. based disturbance compensation and external force estimation for a
[6] PI, "Physik Instrumente (PI) GmbH & Co. KG: piezoelectric actuator," presented at Advanced Motion Control, 2004.
[Link] 2005. AMC '04. The 8th IEEE International Workshop on, 2004.
[7] S. Bauer, "Piezo-, pyro- and ferroelectrets: soft transducer [29] K. Abidi, A. Sabanovic, and S. Yesilyurt, "Sliding-mode based
materials for electromechanical energy conversion," Dielectrics and force control of a piezoelectric actuator," presented at Mechatronics,
Electrical Insulation, IEEE Transactions on [see also Electrical 2004. ICM '04. Proceedings of the IEEE International Conference on,
Insulation, IEEE Transactions on], vol. 13, pp. 953-962, 2006. 2004.
[8] H.-J. Shieh, F.-J. Lin, P.-K. Huang, and L.-T. Teng, "Adaptive [30] S. Drakunov and V. Utkin, "A semigroup approach to discrete-
tracking control solely using displacement feedback for a piezo- time sliding modes," presented at American Control Conference,
positioning mechanism," Control Theory and Applications, IEE 1995. Proceedings of the, 1995.
Proceedings-, vol. 151, pp. 653-660, 2004.

Copyright © 2007 Praise Worthy Prize S.r.l. - All rights reserved International Review of Electrical Engineering,
Vol. xx, n. x
 S. Yannier, A. Sabanovic

[31] W.-C. Su, S. V. Drakunov, and U. Ozguner, "An O(T^2)

boundary layer in sliding mode for sampled-data systems," Automatic
Control, IEEE Transactions on, vol. 45, pp. 482-485, 2000.
[32] V. Utkin, "Variable structure systems with sliding modes,"
Automatic Control, IEEE Transactions on, vol. 22, pp. 212-222, 1977.
[33] V. Utkin, J. Guldner, and J. Shi, Sliding Mode Control in
Electro-mechanical Systems. London; Philadelphia, PA: Taylor &
Francis, 1999.
[34] M. Goldfarb and N. Celanovic, "Behavioral implications of
piezoelectric stack actuators for control of micromanipulation,"
presented at Robotics and Automation, 1996. Proceedings., 1996
IEEE International Conference on, 1996.
[35] K. Ohnishi, M. Shibata, and T. Murakami, "Motion control for
advanced mechatronics," Mechatronics, IEEE/ASME Transactions on,
vol. 1, pp. 56-67, 1996.

Authors’ information
1
Sabanci University, Faculty of Engineering and Natural Sciences,
Mechatronics Program.
2
Sabanci University, Faculty of Engineering and Natural Sciences,
Mechatronics Program.

Selim Yannier was born in Istanbul,

Türkiye on 1978. He received his B.S.
degree in Physics with honors from Orta
Doğu Teknik Üniversitesi (Middle East
Technical University), Ankara, Türkiye
in Fall 2001. He then began M.S. studies
at Mechatronics Engineering Program,
Sabanci University, Istanbul, Türkiye in
Spring 2001 and completed at Spring
2002. He is currently Ph.D. student at the same institution.
Since January 2001 he is working as a Research and
Teaching Assistant in Sabanci University, Istanbul, Türkiye.
His research interests are motion control, mobile robotics and
control for nanotechnologies.

Asif Sabanovic SM IEEE is a full

professor of mechatronics at Sabanci
University, Istanbul. is a Full Professor
at Sabanci University, Istanbul. He
received B.S. degree in electrical
engineering (in 1970) , M.S. degree and
Dr. Eng. degree from University of
[Link] fields of interest include
nonlinear control, robust control,
robotics, electric drives, power electronics and mechatronics.

and the other authors may include biographies at the end of

regular papers. The first paragraph may contain a place
and/or date of birth (list place, then date). Next, the author’s
educational background is listed: type of degree in what field,
which institution, city, state or country, and year degree was
earned. The author’s major field of study should be lower-
cased.

The second paragraph uses the pronoun of the person (he or

she) and not the author’s last name. Information concerning
previous publications may be included. Current and previous
research interests ends the paragraph.

The third paragraph begins with the author’s title and last
name (e.g., Dr. Smith, Prof. Jones, Mr. Kajor, Ms. Hunter).
List any memberships in professional societies. The
photograph is placed at the top left of the biography. Personal
hobbies will be deleted from the biography.

Copyright © 2007 Praise Worthy Prize S.r.l. - All rights reserved International Review of Electrical Engineering,
Vol. xx, n. x

View publication stats