Proceedings of the 18th World Congress

The International Federation of Automatic Control

Milano (Italy) August 28 - September 2, 2011

System Dynamics and Adaptive Control of MEMS Gyroscope Sensor

Juntao Fei, Weili Dai, Mingang Hua, Yuncan Xue

Jiangsu Key Laboratory of Power Transmission and Distribution Equipment Technology
College of Computer and Information, Hohai University
Changzhou, 213022, P.R.China

Abstract: This paper presents an adaptive control approach for Micro-Electro-Mechanical Systems
(MEMS) z-axis gyroscope sensor. The dynamical model of MEMS gyroscope sensor is developed and
established . The proposed adaptive control approaches can estimate the angular velocity and the
damping and stiffness coefficients including the coupling terms due to the fabrication imperfection. The
stability of the closed-loop systems are established with the proposed adaptive control strategies.
Numerical simulations are investigated to verify the effectiveness of the proposed control schemes.
Keywords: Adaptive control, sliding mode control, MEMS gyroscope.

control with application to active suspension system. Some

1. INTRODUCTION
Gyroscopes are commonly used sensors for measuring
angular velocity in many areas of applications such as
navigation, homing, and control stabilization. Vibratory
gyroscopes are the devices that transfer energy from one axis
to other axis through Coriolis forces. The conventional mode
of operation drives one of the modes of the gyroscope into a
known oscillatory motion and then detects the Coriolis
acceleration coupling along the sense mode of vibration,
which is orthogonal to the driven mode. The response of the
sense mode provides information about the applied angular
velocity. Fabrication imperfections result in some cross
stiffness and cross damping effects that may hinder the
measurement of angular velocity of MEMS gyroscope.
Therefore the angular velocity measurement and
minimization of the cross coupling between two axes are
challenging problems in gyroscopes that need to be solved
using advanced control methods.
Adaptive control is an effective approach to handle
parameter variations. In the presence of model uncertainties
and external disturbances, sliding mode control is necessary
to be incorporated into the adaptive control to improve the
robust performance of control system. Sliding mode control
is a robust control technique which has many attractive
features such as robustness to parameter variations and
insensitivity to disturbances. Adaptive sliding mode control
has the advantages of combining the robustness of variable
structure methods with the tracking capability of adaptive
control. In the last few years, many applications have been
developed using sliding mode control and adaptive control.
Utkin (1977)showed that variable structure control is
insensitive to parameters perturbations and external
disturbances. Ioannou et al. (1996) described the model
reference adaptive control. Chou et al. (2003) proposed an
integral sliding surface and derived an adaptive law to
estimate the upper bound of uncertainties. Sam et al. (2004)
presented a class of proportional and integral sliding mode
control algorithms have been proposed to control the MEMS
gyroscope. Batur et.al. (2006) developed a sliding mode
control for a MEMS gyroscope system. Leland (2006)
presented an adaptive force balanced controller for tuning the
natural frequency of the drive axis of a vibratory gyroscope.
Novel robust adaptive controllers are proposed by Fei et al.
(2009 a, b) to control the vibration of MEMS gyroscope. Sun
et al. (2009) developed a phase-domain design approach to
study the mode-matched control of MEMS vibratory
gyroscope. Zheng et al. (2009) described an active
disturbance rejection control for MEMS gyroscopes.
Antonelli et al. (2009) used extremum-seeking control to
automatically match the vibration mode in MEMS vibrating
gyroscopes. Feng et al. (2007) presented an adaptive
estimator-based technique to estimate the angular motion by
providing the Coriolis force as the input to the adaptive
estimator and to improve the bandwidth of microgyroscope.
Tasi et al. (2010) proposed integrated model reference
adaptive control and time-varying angular rate estimation
algorithm for micro-machined gyroscopes. Raman et al.
(2009) developed a closed-loop digitally controlled MEMS
gyroscope using unconstrained sigma-delta force balanced
feedback control. Park et al. (2004,2007) presented an
adaptive controller for a MEMS gyroscope which drives both
axes of vibration and controls the entire operation of the
gyroscope. In this paper, the proposed adaptive control is
different from the adaptive controller developed by Park et al.
(2004) in that an addition controller is incorporated into the
state feedback controller to give more freedom in designing
the adaptive controller. Therefore the error dynamics is
determined by the reference model dynamics and addition
control.
This paper investigates the adaptive control approach to
identify the angular velocity of MEMS gyroscope using state
tracking controllers. The contribution of this paper is that
novel adaptive control is proposed to control the MEMS
gyroscope and to estimate the angular velocity and all

978-3-902661-93-7/11/$20.00 © 2011 IFAC 3551 10.3182/20110828-6-IT-1002.00034

18th IFAC World Congress (IFAC'11)
Milano (Italy) August 28 - September 2, 2011

unknown gyroscope parameters. The paper is organized as d( Ωz × rA ) dΩ

follows. In section II, the dynamics of MEMS gyroscope is The property B
= z × rA + Ωz × vA + Ωz × Ωz × rA
described. The adaptive controller is derived in section III. dt dt B B B

Simulation results are presented in section IV. The can be proved as

Conclusion is provided in section V. d ( Ω z × r A ) dΩ
B z
= × rA + Ω z × ( v A + Ω z × r A )
dt dt B B B
(4)
2. DYNAMICS OF MEMS GYROSCOPE dΩ z
= × r A + Ω z × v A + Ω z × Ω z × rA .
The inertial frame and rotating frame are shown in Fig. 1. dt B B B

The typical MEMS vibratory gyroscope includes a proof dv A

B
mass suspended by springs, an electrostatic actuation and The property = a A + Ω z × v A can be proved as
dt B B
sensing mechanisms for forcing an oscillatory motion and
sensing the position and velocity of the proof mass.  dx dy dz 
dv A d i + j+ k
B  dt dt dt 
=
dt dt
y 2 2
Y r x d x d y d 2z dx di dy dj dz dk
rA = 2 i+ 2 j+ 2 k+ ⋅ + ⋅ + ⋅
dt dt dt dt dt dt dt dt dt
v B
2 2 2
rA d x d y d z dx dy dz
= 2 i + 2 j + 2 k + Ωz × i +Ωz × j +Ωz × k
B dt dt dt dt dt dt
v = aA + Ωz ×vA .
rB B B
(5)
Differentiating (3) and using (4) and (5) yields
O X dΩ z
a A = aB + a A + 2 Ωz × v A + × rA + Ω z × ( Ω z × rA ) (6)
B B dt B B

Fig. 1. Inertial frame and rotating frame where a A and a B are the accelerations with respect to the
d2x d2y d2z
In the Fig. 1, rA is position vector of an arbitrary point A as inertial frame, a A = i + j + k is the acceleration
B dt 2 dt2 dt2
measured against the axes of the inertial frame. r A is an
B vector of an arbitrary point A as measured against the axes
arbitrary point A relative to the origin of the rotating axes , of the rotating system.
rB is position vector of the origin of the rotating frame The acceleration of A relative to B is therefore made up of
four terms - the acceleration measured against the rotating
relative to the origin of the inertial frame B . Their relation
axes and three components that result from the rotation of the
can be expressed as
axes, and are thus invisible to the observer in the rotating
r A = rB + rA frame. These are the Euler (tangential), Centripetal and
B
(1)
= rB + x i + y j + z k . Coriolis accelerations respectively.
Multiplying (6) by mass m gives
The velocity vector of an arbitrary point A as measured
 dΩ z 
against the axes of the inertial frame can be derived as ma A = ma B + ma A + 2m Ω z × v A + m × r A 
dx dy dz di dj dk B B  dt B  (7)
vA = vB + i + j+ k +x + y +z . (2)
dt dt dt dt dt dt + mΩ z × ( Ω z × r A )
B
di dj dk
Since = Ωz × i , = Ω z× j , = Ω z × k , substituting where 2m Ω z × v A is the Coriolis force and
dt dt dt B

these properties into (2) yields mΩz × ( Ωz × rA ) is the Centrifugal force.

B
dx dy dz
v A = vB + i + j + k + x( Ωz × i ) + y( Ωz × j ) + k( Ωz × k ) The Coriolis force acting on the proof mass along x direction
dt dt dt (3) is derived as
= vB + vA + Ωz × rA Fcoriolis − x = 2mΩ z k × y& j = 2mΩ z y&i . (8)
B B

where v A and v B are the velocities with respect to the The Coriolis force acting on the proof mass along y direction
is derived as
dx dy dz
inertial frame, v A = i+ j + k is the relative Fcoriolis − y = 2mΩ z k × x&i = 2mΩ z x&j . (9)
B dt dt dt
velocity vector of an arbitrary point A as measured against By using the property of k × (k × i ) = −i , the Centripetal
the axes of the rotating system. force acting on the proof mass along x direction can be
The velocity of A relative to B is therefore made up of two derived as
Fcentripetal − x = mΩ z k × (Ω z k × xi ) = −mΩ z xi .
2
terms - the velocity measured against the rotating axes and a (10)
component that results from the rotation of the axes, and is
thus invisible to the observer in the rotating frame.

3552
18th IFAC World Congress (IFAC'11)
Milano (Italy) August 28 - September 2, 2011

By using the property of k × ( k × j ) = − j , the Centripetal q D ∗ Ω

q∗ = , D∗ = , Ωz = z È (16)
forces acting on the proof mass along y direction can be q0 mw0 w0
derived as
u k xx k yy kxy
Fcentripetal − y = mΩ z k × (Ω z k × yj ) = −mΩ z yj .
2
(11) u∗ = 2
, wx = 2
, wy = 2
, wxy = 2
. (17)
mw0 q0 mw0 mw0 mw0
We assume that the table where the proof mass is mounted
is moving with a constant velocity; the gyroscope is rotating
at a constant angular velocity Ω z over a sufficiently long Ignoring the superscript (*) for notational clarity, the
2 2
nondimensional representation of (12) and (13) is
time interval; the centripetal forces mΩ z x and mΩ z y are q&& + Dq& + K b q = u − 2Ωq& (18)
assumed to be negligible; gyroscope undergoes rotation about
the z axis only, and thereby Coriolis force is generated in a  w 2 wxy 
where K b =  x 2 .
direction perpendicular to the drive and rotational axes. A z-  wxy w y 
axis MEMS gyroscope is depicted in Fig. 2. Rewriting the gyroscope model (18) in state space form as
With the assumptions the dynamics of gyroscope become
m&x& + d xx x& + d xy y& + k xx x + k xy y = u x + 2mΩ z y& (12) X& = AX + Bu (19)

m&y& + d xy x& + d yy y& + k xy x + k yy y = u y − 2mΩ z x& . (13) where

Fabrication imperfections contribute mainly to the symmetric  0 1 0 0 
spring and damping terms, k xy and d xy . The x and y axes − w 2 − d xx − wxy − ( d xy − 2Ω z )
A=
x
spring and damping terms k xx , k yy , d xx and d yy are mostly  0 0 0 1 
known, but have small unknown variations from their  2 
nominal values. The mass of proof mass can be determined  − wxy − ( d xy + 2Ω z ) − wy − d yy 
very accurately, and u x , u y are the control forces in the x  x
T
u x   x& 
and y direction. 0 1 0 0  
B=  , u =   , X = (20)
 0 0 0 1  u
 y  y
 
y Ωz  y& 
The reference model x m = A1 sin( w1t ) , y m = A2 sin( w2 t ) is
k yy d yy defined as
k xx k xx
q&&m + K m qm = 0 (21)

m where K m = diag w1 w2 . { 2 2
}
Proof Mass Similar to (19), the reference model can be written as
d xx d xx x  0 1 0 0
− w 2 0 0 0 
X& m =  1 X m ≡ Am X m (22)
k yy d yy  0 0 0 1
 2 
 0 0 − w2 0

Fig. 2. A simple model of a MEMS Z axis gyroscope where Am is a known constant matrix.
We make the following assumptions.
Dividing (12) and (13) by the reference mass and rewriting
Assumption. There exists a constant matrix K * such that the
the gyroscope dynamics in vector forms result in T
D K u following matching condition A + BK * = Am can always be
q&& + q& + a q = − 2 Ωq& (14)
m m m satisfied.
ux   0 −Ωz  d xx d xy  The control target for MEMS gyroscope is (i) to design an
x
where q =   , u =   , Ω =   , D= , adaptive controller so that the trajectory of X can track the
y uy  Ωz 0  d xy d yy  state of reference model X m ; (ii) to estimate the angular
k xx k xy  velocity of MEMS gyroscope and all unknown gyroscope
Ka =  . parameters.
k xy k yy 
Using non-dimensional time t ∗ = w0t , and dividing both
3. ADAPTIVE CONTROL DESIGN
sides of (14) by w02 and reference length q0 give the final
form of the non-dimensional equation of motion as In this section an adaptive controller is proposed to
identify the angular velocity and all unknown gyroscope
q&& D q& K q u Ω q&
+ + 2
= 2
−2 . (15) parameters. The block diagram is shown in Fig. 3. In the
q0 mw0 q0 mw0 q0 mw0 q0 w0 q0 adaptive control design, we consider (19) as the system
Defining a set of new parameters as follows: model.

3553
18th IFAC World Congress (IFAC'11)
Milano (Italy) August 28 - September 2, 2011

bounded and e& is also bounded, according to Barbalat’s

Disturbance
lemma, e will asymptotically converge to zero, lim e( t ) = 0 .
t →∞

Xm e From the adaptive law(30), , according to the persistence

Reference
-
Adaptive MEMS excitation theory, if X is persistent excitation signal, then
Model Controller Gyroscope ~
+ (30) guarantees that K → 0 , K will converges to its true
X values. It can be shown that there always exist some positive
scalar constants α and T such that for all t > 0 ,
Adaptive Estimation of t +T
Law Angular rate ∫
t
XX T dτ ≥ αI if w1 ≠ w2 . Therefore, if w1 ≠ w2 , i.e. the
frequencies on x and y axes are different, which means
that XX T has full rank, the persistent excitation can be
satisfied., then controller parameter converges to its true
~ ~
values, K → 0 . Since K → 0 , then the unknown angular
Fig.3. Block diagram of adaptive control for MEMS velocity as well as all other unknown parameters can be
gyroscope determined from A + BK T = Am and we obtain
Ω z = 0.25( k 22 − k 41 ) .
The tracking error and its derivative are
Conclusion: if persistently exciting drive signals,
e(t) = X (t) − X m (t) (23) ~
xm = A1 sin(w1t ) and y m = A2 sin( w2t ) are used, then K and
e&(t) = Am e(t) + (A − Am )X (t) + Bu(t) . (24)
The adaptive controller is proposed as e( t ) all converge to zero asymptotically. Consequently the
unknown angular velocity can be determined as
u(t) = K T (t) X (t) + K f e(t) (25)
lim t →∞ Ω z (t ) = Ω z . However it is difficult to establish the
*
where K ( t ) is an estimate of K , the constant matrix K f convergence rate.
satisfies the condition that ( Am + BK f ) is Hurwitz.
~
We define the estimation error as K (t) = K (t) − K * and 4. APPLICATION WITH MEMS GYROSCOPE:
substitute the estimation error and (25) into (19) yields SIMULATION STUDY
~
X& (t) = Am X (t) + BK T (t) X (t) + K f e( t ) (26) We will evaluate the proposed adaptive control on the
Then, we have the tracking error equation lumped MEMS gyroscope model by Batur et al. (2006) using
~ MATLAB/SIMULINK. The control objective is to design
e&(t) = (Am + BK f )e(t) + BK T (t) X (t) (27)
adaptive state tracking controller so that a consistent estimate
Define a Lyapunov function of Ωz can be obtained.
1 1 ~
V = e T Pe + tr KM −1 K T
2 2
~
[ ] (28) In the simulation, we allowed ± 5% parameter variations
for the spring and damping coefficients with respect to their
where P and M = diag{m1 m2 } are positive definite matrix. nominal values. We further assumed ± 5% magnitude
Differentiating V with respect to time yields changes in the coupling terms i.e. d xy and ω xy , again with
~ ~&
V& = eT Pe& + tr  KM −1 K T  respect to their nominal values. The external disturbance is a
  random variable signal with zero mean and unity variance.
(29)
~ ~ ~&
= −e T Qe + e T PBK T X + tr  KM −1 K T 
Parameters of the MEMS gyroscope are as follows:
  m = 0.57e − 8 kg, d xx = 0.429e − 6 N s/m,
where P(Am + BK f ) + (Am + BK f )T P = −Q , Q is positive d xy = 0.0429e − 6 N s/m, d yy = 0.687e − 36 N s/m,
definite matrix. k xx = 80.98 N/m, k xy = 5 N/m, k yy = 71.62 N/m,
To make V& ≤ 0 , we choose the adaptive law as w0 = 1kHz , q 0 = 10 −6 m .
~&
K T (t) = K& T (t) = − MB T P T eX T (30) The unknown angular velocity is assumed Ω z = 5.0 rad/s
with K (0) being arbitrary. This adaptive law yields and the initial condition on K matrix is K ( 0 ) = 0.95 K * . The
V& = −e T Qe ≤ −λ ( Q ) e ≤ 0
min (31) desired motion trajectories are xm = sin(w1t ) and
The inequality V& ≤ − λ min ( Q ) e implies that e is y m = 1.2 sin( w2 t ) , where w1 = 6.17 kHz and w2 = 5.11kHz.
t 1 The adaptive gain of (30) is M = diag {20 20} . The K f in
integrable as ∫
0
e dt ≤
λmin ( Q )
[V ( 0 ) − V ( t )] . Since V ( 0 ) is
−10000 −10000 1000 20000
(29) is chosen as K f =  .
bounded and V ( t ) is nonincreasing and bounded, it can be  −1000 −1000 −1000 −1000
t t
concluded that lim ∫ e dt
t →∞ 0
is bounded. Since lim
t →∞ 0 ∫ e dt is Fig. 4 depicts the tracking errors. It is observed that the
tracking errors converge to zero asymptotically. Figs. 5 and

3554
18th IFAC World Congress (IFAC'11)
Milano (Italy) August 28 - September 2, 2011

Fig. 7 draw the adaptation of the angular velocity and x 10

4

2
controller parameters. It is shown that the estimates of

Control Signal u1
angular velocity and controller parameters converge to their 0

values. Fig.6 plots the control input using adaptive control . -2

The estimate of angular velocity using adaptive control has -4

larger overshoot at the beginning but much smaller rise time. -6

0 10 20 30 40 50 60 70 80 90 100
The adaptive control can deal with the model uncertainties Time(Second)
and external disturbances to some extent although the term of
15000
model uncertainties and external disturbances do not show up

Control Signal u2
in the adaptive control derivation. 10000

5000
Simulations demonstrate that with the control laws (25),
and the parameter adaptation laws (30), if the gyroscope is 0
controlled to follow the mode-unmatched reference model,
-5000
the persistent excitation condition is satisfied, i.e. w1 ≠ w2 , 0 10 20 30 40 50 60 70 80 90 100
Time(Second)
and all unknown gyroscope parameters, including the angular
velocity converge to their true values, and tracking error is
Fig.6. Adaptive control input
going to zero asymptotically as time go on.
-3
x 10
2

0 k11 1540
1535
e1

-2 1530
0 10 20 30 40 50 60 70 80 90 100
Time(Second)
-4 50
0 10 20 30 40 50 60 70 80 90 100
k12

0
Time(Second)
-3 -50
x 10 0 10 20 30 40 50 60 70 80 90 100
6
Time(Second)
4 100
k13

90
e3

2
80
0 10 20 30 40 50 60 70 80 90 100
0 Time(Second)
20
-2
0 10 20 30 40 50 60 70 80 90 100
k14

0
Time(Second)

-20
Fig.4. The tracking error using adaptive control 0 10 20 30 40 50 60 70 80 90 100
Time(Second)

100
k21

90

8
80
0 10 20 30 40 50 60 70 80 90 100
Time(Second)
6 50
k22

0
4
-50
0 10 20 30 40 50 60 70 80 90 100
2 Time(Second)
Angular Rate

1380
k23

0 1360
1340
0 10 20 30 40 50 60 70 80 90 100
-2
Time(Second)
100
-4
k24

0
-6
-100
0 10 20 30 40 50 60 70 80 90 100
-8 Time(Second)
0 10 20 30 40 50 60 70 80 90 100
Time(Second)

Fig.7. Adaptation of control parameters using adaptive

Fig.5. Adaptation of angular velocity using adaptive control control

3555
18th IFAC World Congress (IFAC'11)
Milano (Italy) August 28 - September 2, 2011

5. CONCLUSIONS Raman, J., Cretu, E., Rombouts, P., and Weyten, L. (2009).
A closed-loop digitally controlled MEMS gyroscope
This paper investigates the design of adaptive control for
with unconstrained sigma-delta force-feedback, IEEE
MEMS gyroscope. The dynamics model of the MEMS
Sensors Journal, 9(3), pp. 297-305.
gyroscope is developed and nondimensionized. Novel
Sam, Y. J., Osman, H. and Ghani, M. R. (2004). A class of
adaptive controller is proposed and stability condition is
proportional- integral sliding mode control with
established. Simulation results demonstrate that the
application to active suspension system, Systems &
effectiveness of the proposed adaptive control techniques in
Control Letters, 51(3/4), pp. 217–224.
identifying the gyroscope parameters and angular velocity.
Sung, S., Sung, W., Kim, C., Yun, S., and Lee,Y. (2009).
On the mode-matched control of MEMS vibratory
gyroscope via phase-domain analysis and design,
ACKNOWLEDGMENTS IEEE/ASME Transactions on Mechatronics, 14(4),
pp.446-455.
This work is partially supported by National Science Tsai, N., and Sue, C. (2010). Integrated model reference
Foundation of China under Grant No. 61074056, The Natural adaptive control and time-varying angular rate
Science Foundation of Jiangsu Province under Grant estimation for micro-machined gyroscopes,
No.BK2010201. The authors thank to the anonymous International Journal of Control, 83(2), pp. 246-256.
reviewers for useful comments that improved the quality of
Utkin, V. I. Ä1977Å.Variable structure systems with sliding
the manuscript.
modes, IEEE Trans. on Automatic Control, 22, pp.
212-222 .
REFERENCES Zheng, Q., Dong, L., Lee, D., and Gao, Z.(2009). Active
disturbance rejection control for MEMS gyroscopes,
Antonello ,R., Oboe, R. , Prandi, L., and Biganzoli, F. IEEE Transactions on Control System Technology,17(6),
(2009).Automatic mode matching in MEMS vibrating pp. 1432 -1438.
gyroscopes using extremum-seeking control, IEEE
Transactions on Industrial Electronics, 56(10), pp.3880-
3891.
Batur, C. and Sreeramreddy, T.(2006). Sliding mode control
of a simulated MEMS gyroscope, ISA Transactions,
45(1), pp.99-108 .
Chou, C., and Cheng , C.(2003). A decentralized model
reference adaptive variable structure controller for large-
scale time-varying delay systems, IEEE Transactions on
Automatic Control, 48(7), pp. 1213-1217.
Fei, J. and Batur, C. (2009). Robust adaptive control for a
MEMS vibratory gyroscope, International Journal of
Advanced Manufacturing Technology, 42(3), pp. 293-
300.
Fei, J. and Batur, C. (2009). A novel adaptive sliding mode
control for MEMS gyroscope È ISA Transactions 
48(1), pp. 73-78.
Feng, Z., and Fan, M. (2007). Adaptive input estimation
methods for improving the bandwidth of
microgyroscopes, IEEE Sensors Journal, 7(4), pp.562-
567.
Ioannou, P. and Sun, J. (1996). Robust Adaptive Control.
Prentice-Hall.
Leland, R. (2006).Adaptive control of a MEMS gyroscope
using Lyapunov methods. IEEE Transactions on
Control Systems Technology, 14, pp.278–283.
Park, S., Horowitz, R., Hong, S., and Nam, Y. Nam. (2007).
Trajectory-switching algorithm for a MEMS gyroscope,
IEEE Transactions on Instrumentation and
Measurement, 56(60), pp.2561-2569.
Park, S., and Horowitz, R. (2004). New adaptive mode of
operation for MEMS gyroscopes, ASME Transaction
Dynamic Systems, Measurement and Control, 126,
pp.800-810.

3556