Sensors and Materials, Vol. 31, No.

8 (2019) 2595–2608 2595

MYU Tokyo

S & M 1959

Bidirectional Carouseling Method Based on Phase Difference

for MEMS Gyro North Finder
Xiaowen Cai, Chunxi Zhang, Shuang Gao,* and Ruoyu Zhang

School of Instrumentation Science and Opto-electronics Engineering,

Beihang University, Beijing 100191, China

(Received March 7, 2019; accepted July 16, 2019)

Keywords: MEMS gyro, north finder, bias drift, bidirectional, phase difference

The north finder has been widely used in precision guided weapons, unmanned aerial
vehicles (UAVs), and other device. In this paper, a north finder based on a low-cost micro-
electro-mechanical system (MEMS) gyro is proposed. In order to reduce the substantial bias
effect on the gyro’s north finding angle, a bidirectional carouseling method is applied. To
improve the north finding accuracy, the phase difference between the turntable rotation rate and
the gyro output is used. After the calculation of the phase difference, a five-order Butterworth
filter is chosen to eliminate the gyro high-frequency noise. Experimental results demonstrate
that the proposed method based on the phase difference can effectively improve the north
finding accuracy by 50% compared with the traditional carouseling method.

1. Introduction

The north finder is widely used in military and civilian fields such as targeting, dead
reckoning, pointing system, and inertial navigation. Traditional gyro north finders commonly
use high-precision gyros, such as fiber optic gyros (FOGs), ring laser gyros (RLGs), and
hemispherical resonator gyros (HRGs). However, the high-precision gyros are bulky
and expensive.(1) Their large volume, high power consumption, and high cost limit their
applications in portable navigation. With the rapid development of the micro-electromechanical
system (MEMS) technique, the precision and stability of MEMS gyros have been significantly
improved in recent years.(2) Owing to its small size, low cost, low power, and environment
adaptability,(3,4) a MEMS gyro north finder has become a research focus in the north finder
field. However, limited by its mechanical precision, a MEMS gyro is one or two orders of
magnitude worse than a high-precision gyro;(5) thus, effective methods must be developed to
improve its measurement accuracy. In order to realize a MEMS gyro north finder, there are
two main traditional methods: a static method (maytagging) and a dynamic method (carouseling).
Maytagging is carried out by discrete ±180° turning of the gyroʼs sensitive axis, and it allows
a gyro to face in a given direction while collecting data at the desired interval. Maytagging
has been widely applied in various methods of north finding angle detection, such as the two-
*
Corresponding author: e-mail: gaoshuang@[Link]
[Link]

ISSN 0914-4935 © MYU K.K.

[Link]
2596 Sensors and Materials, Vol. 31, No. 8 (2019)

point method, four-point method, and multiposition method.(6–8) It is simple to operate and is
widely employed in practical applications. However, this method will increase the error caused
by the rate random walk (RRW), which conflicts with the angular random walk (ARW). In
contrast to maytagging, carouseling has great advantages in terms of the robustness against
bias drift,(9) angular random walk, and temperature drift. The north finding angle is measured
by the turntable continuous rotation. However, the angular position error of the turntable
will bring additional uncertainties to the accuracy of north finding in this method. Draper
Laboratory has been developing a MEMS north finder for over a decade and has become a
world leader in this field.(10) Developers from Honeywell Laboratory have reported a process
for a MEMS tuning-fork-type gyro with the azimuth angle maintained in 30 s,(11) and long-term
bias drift was reduced by carouseling. Trusov and coworkers discussed how carouseling can
reduce the north finding time and decrease the error caused by bias drift and angular random
walk.(12,13) Renkoski showed progress in the development of high-accuracy MEMS gyros with
north-finding capabilities, so that the errors can be eliminated by an improved carouseling
method called the baseball stitch slew.(10) Prikhodko et al. obtained the north finding angle by
detecting the difference between the phase of a sinusoidal fit and the instantaneous position of
the turntable.(14) The uncertainty was reduced by data averaging and the robustness against
bias drift in carouseling was proved evidently. However, few applications of carouseling have
been realized in the industrial field; thus, there is a great potential to develop a novel method to
improve carouseling and put it into practical use.
In order to overcome the shortcomings of traditional methods, in this paper we present a
novel carouseling method. The proposed method aims to achieve the following functions:
(1) An improved carouseling method is proposed to decrease MEMS gyro drift;
(2) The phase difference between the turntable position signal and MEMS gyro output is
utilized to further reduce the gyro drift and improve north finding accuracy;
(3) A five-order Butterworth filter is used to eliminate the gyro high-frequency noise.
The remainder of this work is organized as follows. In Sect. 2, the theoretical basis of the
methodology of MEMS gyro north finder is presented. The implementation of the proposed
system is introduced in detail in Sect. 3. The experimental results in Sect. 4 validate the
feasibility of the proposed method. To sum up, conclusions are drawn in Sect. 5.

2. Methodology of North Finder

At a certain local terrestrial latitude L, the rotation rate is divided into two parts, the
vertical component and the horizontal component,(15) as shown in Fig. 1. The gyro’s reference
coordinate system is coincident with the local geographic coordinate system Xg, Yg, and Zg,
denoting the geographical east, north, and up directions, respectively. The MEMS gyro is fixed
perpendicular to the horizontal plane and the sensitive axis is paralleled to the horizontal plane.(16,17)
α in the gyro’s coordinate system in the O−XgYg plane is the north finding angle.
As shown in Fig. 1, ωe is Earth’s rotation rate, which is 15.041°/h.(18) The horizontal
component of Earth’s rotation rate is described as

ω N = ωe cos L . (1)
Sensors and Materials, Vol. 31, No. 8 (2019) 2597

Fig. 1. (Color online) Geographic coordinate system on the gyro’s location.

Assuming that Yg represents the true north direction, the north component of Earth’s rotation
rate is

ω y = ω N cos α . (2)

According to Eq. (2), the north finding angle α can be obtained by measuring the angular
rate of the gyro.

3. Carouseling Method for North Finder

3.1 Traditional carouseling for north finder

The traditional carouseling method measures Earth’s rotation rate on the basis of the
continuous rotation of the MEMS gyro. As shown in Fig. 2, when the orientation of the gyro’s
sensitive axis is rotated continuously, the gyro output appears to be a sinusoidal signal with the
variation of the angle between Earth’s rotation axis and the gyroʼs sensitive axis.
According to the mathematical model, when the turntable is rotated at an angular rate ωt, the
dynamic output model of the MEMS gyroʼs at the north finding angle α in the horizontal plane
can be expressed as

ω (t ) K ω N cos(ωt t + α ) + bc + ε (t ) ,
= (3)

where ω(t) is the gyro dynamic output, K is the gyro scale factor, ωN is the vertical component
of Earth’s rotation rate, ε(t) is the gyro random drift, and bc is the gyro constant bias. The gyro
output can be discretized as
 2598 Sensors and Materials, Vol. 31, No. 8 (2019)

Fig. 2. Curve of Earth and turntable combined action.

=ω (i / f s ) K ω N cos[2π ( ft / f s )i + α ] +=
bc + ε (i / f s ), (i 1, 2,  , N ) (4)

where ft is the turntable frequency, fs is the sampling frequency, and i is the sampling point
number. According to Eq. (4), Eq. (5) can be obtained by matrix decomposition.

=ω (i / f s ) K ω N cos[2π ( ft / f s )i + α ] + bc + ε (i / f s )
= K ω N {cos[2π ( ft / f s )i ]cos α − sin[2π ( ft / f s )i ]sin α }
+ bc + ε (i / f s )
T
 K ω N cos[2π ( ft / f s )i ]  cos α  (5)
 − K ω sin[2π ( f / f )i ]  sin α 
= = N t s    (i 1, 2,  , N )
 1   bc 
   
 ε (i / f s )   1 

The least squares method (LSM)(19) can be used to solve the problem. Denote

=Y ω=
(i / f s ) [ω (1 / f s ), ω (2 / f s ),  , ω ( N / f s )],
T
X = [ cos α sin α bc 1] , and
 K ω N cos [ 2π ( ft / f s ) ] − K ω N sin [ 2π ( ft / f s ) ] 1 ε (1 / f s ) 
  (6)
K ω N cos [ 2π (2 ft / f s ) ] − K ω N sin [ 2π (2 ft / f s ) ] 1 ε ( 2 / f s ) 
A= .
     
 
 K ω N cos [ 2π ( Nft / f s ) ] − K ω N sin [ 2π ( Nft / f s ) ] 1 ε ( N / f s ) 
Sensors and Materials, Vol. 31, No. 8 (2019) 2599

Equation (5) can be written as

Y = AX . (7)

Equation (7) can be reckoned by LSM solutions. Then, the north finding angle α can be
calculated as


( )
−1
X = AT A ATY , (8)

 A sin α   X11 
=α arctan
=   arctan  . (9)
 A cos α   X12 

3.2 Bidirectional carouseling method based on phase difference

The traditional carouseling method has the advantages of f lexibility in designing

parameters, real-time north finding, and simple operation. Under the same conditions, such
as the north finding time or the number of samples collected, the north finding error of the
carouseling method is much larger than that of the maytagging method. Thus, the north finding
performance of the carouseling method is not ideal for practical use. In order to reduce the
effect of bias drift on north finding accuracy, the mechanism of gyro drift is analyzed and a
bidirectional carouseling method based on the phase difference is proposed.

3.2.1 Effect of MEMS gyro output and rotation rate instability error of the turntable

According to Eq. (3), by ignoring the mechanical vibration of the turntable and considering
the gyro drift and the rotation rate instability error of the turntable, we obtain the actual gyro
output as

=ω (t ) K ω N cos (ωt +∆ωt ) t + α  + bc + ε (t )

(10)
= K ω N cos (ωt +∆ωt ) t  cos α − sin (ωt +∆ωt ) t  sin α + bc + ε (t ),

where Δωt is the rotation rate instability error of the turntable and ε(t) is a slowly time-varying
random signal, which is an important performance parameter of the MEMS gyro and also the
important component affecting north finding accuracy. To reduce the impact of ε(t), there are
two methods to solve the problem: the appropriate north finding solution method and the digital
filtering method. In this paper, we choose a bidirectional carouseling method to solve the
problem.
Equation (10) shows that the north finding angle error is directly caused by the rotation rate
instability error of the turntable. At present, the technology for stable-speed servos is quite
mature. The turntable used in this paper is a high precision turntable, the rotation rate instability
error is small and can be neglected in general.
2600 Sensors and Materials, Vol. 31, No. 8 (2019)

3.2.2 Bidirectional carouseling method based on phase difference

The position signal of the turntable is an angle value, which increases from an initial position
to 360°. The position signal of the turntable can be assumed as a periodic signal; then, the
position signal Pt(t) can be described as

Pt (t ) = cos(ωt t ) , (11)

where Pt(t) has no unit because it represents the value on the unit circle.
Assume that the phase difference between the gyro output and the turntable position signal
is the north finding angle α. The curves of the MEMS gyro output ω(t) and the position signal
Pt(t) are shown in Fig. 3.
When the turntable rotation mode is bidirectional, Eqs. (12) and (13) can be substituted for
Eq. (11) as

Pt _ CW (t ) = cos(ωt _ CW t ) , (12)

Pt _ CCW (t ) = cos(ωt _ CCW t ) , (13)

where ωt_CWt is the gyro clockwise output and ωt_CCWt is the gyro counterclockwise output.
By multiplying the position signal of the turntable and MEMS gyro output, we describe the
differential phase N(t) as

N (t ) = Pt (t )ω (t ) . (14)

By using Eqs. (3), (12), and (13), we obtain the differential phases NCW (t) and NCCW (t)
through bidirectional rotation as

Fig. 3. (Color online) Curves of MEMS gyro output and the position signal.
 Sensors and Materials, Vol. 31, No. 8 (2019) 2601

NCW (t ) = Pt _ CW (t )ωCW (t )
= cos(ωt _ CW t )[ K ω N cos(ωt _ CW t + α ) + bc_CW + ε CW (t )]
= cos(ωt _ CW t ) K ω N cos(ωt _ CW t + α )
+ cos(ωt _ CW t )bc_CW + cos(ωt _ CW t )ε CW (t ) (15)
1 1
= ω N cos(2ωt _ CW t + α ) + ω N cos α
2 2
+ cos(ωt _ CW t )(bc_CW + ε CW (t )),

NCCW (t ) = Pt _ CCW (t )ωCCW (t )

= cos(ωt _ CCW t )[ K ω N cos(ωt _ CCW t + α ) + bc_CCW + ε CCW (t )]
= cos(ωt _ CCW t ) K ω N cos(ωt _ CCW t + α )
+ cos(ωt _ CCW t )bc_CCW + cos(ωt _ CCW t )ε CCW (t ) (16)
1 1
= ω N cos(2ωt _ CCW t + α ) + ω N cos α
2 2
+ cos(ωt _ CCW t )(bc_CCW + ε CCW (t )),

where NCW (t) and NCCW (t) represent the differential phases of clockwise and counterclockwise
rotations, respectively. Pt_CW (t) and Pt_CCW (t) are the bidirectional rotation of the turntable
position signals.
On the basis of the phase difference, the constant and random drifts of MEMS gyro are
modulated into periodic signals using a low-pass filter. Considering the calculation complexity
and filtering performance, a five-order Butterworth filter is chosen to eliminate the periodic
signal.(20) Through the Butterworth filter, all periodic signals in Eqs. (15) and (16) are filtered
out, leaving the second term with the phase difference. After the modulation using the low-pass
filter, Eqs. (15) and (16) can be simplified as

1
= t _ CW (t )ωCW (t )
NCW (t ) P= ω N cos α , (17)
2

1
= t _ CCW (t )ωCCW (t )
NCCW (t ) P= ω N cos α . (18)
2

By adding Eqs. (17) and (18), we can calculate the north finding angle as

ω N cos α ,
NCW (t ) + NCCW (t ) = (19)

NCW (t ) + NCCW (t )
α = arccos . (20)
ωN
2602 Sensors and Materials, Vol. 31, No. 8 (2019)

According to the derivation process, the signal modulation based on the phase difference
between the position signal of the turntable and the MEMS gyro output in bidirectional rotation
eliminates the effects of gyro drift and is not affected by the turntable accuracy.

4. Experiments and discussion

4.1 North finder scheme and MEMS gyro characteristics

On the basis of the aforementioned north finder that was previously described, the proposed
hardware design is shown in Fig. 4. The MEMS gyro used was ADIS16495, which is produced
by Analog Devices, Inc.(21) MEMS gyro characteristics are shown in Table 1. The experiments
were conducted at room temperature in Beijing, China, at latitude 39.97° and longitude 116.34°.
The Allan variance was used to analyze the characteristics of the MEMS gyro. To plot
the Allan deviation, data of the MEMS gyro were collected over 55 h. The Allan deviation is
shown in Fig. 5. As shown in Fig. 5, gyro noise contributes to the Allan deviation. There is

Fig. 4. (Color online) The proposed MEMS north finder.

Table 1
MEMS gyro characteristics.
Parameter Typical value Parameter Typical value
Gyro range ±125°/s In-Run Bias Stability 0.8°/h
Nonlinearity ±0.2% of FS Rate Noise Density 0.002°/s/√Hz (RMS)
Misalignment ±0.005° Angle Random Walk 0.56°/√h
Repeatability ±0.07°/s Error Over Temperature ±0.2%

Fig. 5. (Color online) MEMS gyro Allan deviation.

Sensors and Materials, Vol. 31, No. 8 (2019) 2603

a direct relationship between the standard deviation of the output vs time and the slope of the
Allan deviation. From Fig. 5, we can determine that the gyro ARW is 0.54°/ h and the gyro
instability is 0.14°/h.

4.2 Experiment on phase difference

Because the low-pass filter was used, a higher rotation rate was found to be suitable for
the experiment. The initial position was set to 0°, the designed rotation rate was ±100°/s, the
sampling frequency was 10 Hz, and output data for 100 cycles were collected during a period of
360 s. The MEMS gyro original output data are shown in Fig. 6.
As shown in Fig. 6, the gyro output signal contains a large number of different frequencies
of gyro drift. The gyro drift at different frequencies was analyzed on the basis of power
spectral densities (PSD) shown in Fig. 7. In order to reduce the noise and improve the signal-
to-noise ratio, the gyro output requires a suitable low-pass filter to deal with the filtered gyro
output. A five-order Butterworth filter was chosen to filter the gyro output.(22) The frequency
response of the Butterworth filter was maximally flat in the passband and rolled off to zero in
the stopband. This characteristic is suitable for the gyro in this experiment. The filtered results

Fig. 6. (Color online) MEMS gyro original output.

Fig. 7. (Color online) PSD and low-pass filter results.

2604 Sensors and Materials, Vol. 31, No. 8 (2019)

are also shown in Fig. 7. The result demonstrates that the high-frequency component has almost
been eliminated apparently, but the low-frequency component still remains.

4.3 Results of carouseling method at different rotation rates

Assuming that the initial position of the MEMS gyro is coincident with the true north,
the turntable rotation rate increases from 1°/s to 100°/s at 5°/s-rate intervals, and the sample
frequency is 10 Hz. When the turntable rotation is bidirectional, the gyro outputs are collected
on the basis of three typical rotation rates of the turntable, as shown in Fig. 8.
In order to compare the proposed method and the traditional method, the north finding angle
was also obtained by the traditional carouseling method. The gyro output data for 20 cycles
were collected and the relationship between the result of north finding and the turntable rate is
shown in Fig. 9. The 1-sigma error bars of the proposed and traditional methods are shown in
Fig. 10.

(a) (b)

(c)

Fig. 8. (Color online) Gyro output when the turntable rotation rates are (a) ±1°/s, (b) ±10°/s, and (c) ±100°/s.
Sensors and Materials, Vol. 31, No. 8 (2019) 2605

Fig. 9. (Color online) North finding angles obtained Fig. 10. (Color online) 1-sigma error bars of proposed
by proposed and traditional methods. and traditional methods.

As shown in Figs. 9 and 10, the north finding accuracy degrades with increasing turntable
rotation rate. Thus, it is necessary to select the appropriate rotation rate carefully. The north
finding accuracy is improved by the proposed method, which verifies the validity of this
method. In order to compare the accuracy between the traditional method and the proposed
method, the results at different rotation rates are shown in Table 2. As shown in Table 2, the
north finding angle determined by the improved method is superior to that determined by the
traditional method, and the stability of the MEMS gyro is improved by 32.6%.

4.4 Experimental results at different initial positions

In order to further verify the performance of the proposed method, the initial direction of the
gyro’s sensitive axis was set to eight positions at 45° intervals within a cycle, and the turntable
rotation rate was set to 10°/ s. Ten trials of the north finding experiments were carried out at
each position, and the north finding results obtained by the traditional method and the proposed
method based on the phase difference are shown in Table 3. From Table 3, it is obvious that the
proposed method improved the north finding accuracy by 50% compared with the traditional
method, which clearly proves the effectiveness of the method.

4.5 Discussion

In this study, a series of experiments were performed to validate the feasibility and
effectiveness of the proposed method, and the experimental results are discussed in detail. The
results prove the superior performance of the proposed method, which successfully improves
the north finding accuracy. The main advantages of the proposed method are as follows.
(1) The north finding results are not affected by the turntable rotation rate instability error;
(2) The proposed bidirectional carouseling method is effective;
(3) Experiments under different conditions can be carried out to verify the theoretical deduction.
2606 Sensors and Materials, Vol. 31, No. 8 (2019)

Table 2
North finding angle of two methods based on different rotation rates.
Method Turntable rotation rate (°/s) North finding angle (°)
1 3.53
Traditional method 10 4.91
100 16.23
±1 2.38
Proposed method ±10 3.16
±100 12.41

Table 3
North finding accuracy of two methods at different initial positions.
Initial position (°) Traditional method [° (1σ)] Proposed method [° (1σ)]
0 0.86 0.47
45 0.90 0.51
90 0.82 0.43
135 0.87 0.49
180 0.93 0.53
225 0.89 0.48
270 0.83 0.44
315 0.88 0.50

5. Conclusions

In order to improve the north finding accuracy, a bidirectional carouseling method based
on the phase difference is proposed. With the low-pass filter removing the periodic term in
the differential signal, the effect of gyro bias and random error on the north finding accuracy
is effectively suppressed, and the turntable rotation rate instability error does not affect the
north finding accuracy. The experimental results demonstrate that the proposed method can
effectively achieve fast and high-precision MEMS gyro north finding. Our future work will be
focused on reducing the effects of temperature and g-sensitivity to further improve the north
finding accuracy.

Acknowledgments

This work was supported by the National Key Scientific Instrument and Equipment
Development Project (2013YQ50791).

References
1 A. A. Trusov, I. P. Prikhodko, D. M. Rozelle, A. D. Meyer, and A. M. Shkel: Int. Conf. Solid-State Sens. (2013)
2531. [Link]
2 Y. L. Wei and M. Lee: Proc. 2012 SICE Annu. Conf. (SICE, 2012) 2262. [Link]
23914-4_4
3 L. I. Iozan, M. Kirkko-Jaakkola, J. Collin, J. Takala, and C. Rusu: Meas. Sci. Technol. 23 (2012) 025005.
[Link]
4 Y. J. Zhang, B. Zhou, M. L. Song, B. Hou, H. F. Xing, and R. Zhang: Sensors 17 (2017) 973. [Link]
org/10.3390/s17050973
5 H. F. Xing, B. Hou, and Z. H. Lin: Sensors 17 (2017) 2335. [Link]
Sensors and Materials, Vol. 31, No. 8 (2019) 2607

6 R. Arnaudov and Y. Angelov: Meas. Sci. Technol. 16 (2005) 2300. [Link]

7 D. Lapadatu, B. Blixhavn, R. Holm, and T. Kvisteroy: IEEE/ION Position Location Navig. Symp. 1 (2010) 6.
[Link]
8 S. C. Wu, C. Liu, S. B. Yu, Z. Q. Wang, and J. R. Li: Int. Conf. MEC (IEEE, 2011) 994. [Link]
MEC.2011.6025632
9 I. P. Prikhodko, S. A. Zotov, and A. A. Trusov: Proc. 2011 16th Int. Solid-State Sensors, Actuators and
Microsystems Conf. (IEEE, 2011) 2809. [Link]
10 B. M. Renkoski: M. S. thesis, Dept. Aeronaut. Astronaut., Massachusetts Inst. Technol. (2008).
11 B. R. Johnson, E. Cabuz, H. B. French, and R. Supino: IEEE/ION Position Location Navig. Symp. (2010) 168.
[Link]
12 A. A. Trusov, I. P. Prikhodko, S. A. Zotov, and A. M. Shkel: IEEE/ION IEEE Position Location Navig. Symp. (2012)
247. [Link]
13 I. P. Prikhodko, S. A. Zotov, A. A. Trusov, and A. M. Shkel: J. Microelectromech. Syst. 22 (2013) 1257. https://
[Link]/10.1109/JMEMS.2013.2282936
14 I. P. Prikhodko, A. A. Trusov, and A. M. Shkel: IEEE Int. Conf. Micro Electro. Mech. Syst. (2012) 164. https://
[Link]/10.1109/MEMSYS.2012.6170119
15 Q. Wang, J, Xie, C. C. Yang, C. H. He, X. Y. Wang, and Z. Y. Wang: [Link]. 54 (2015) 8944. [Link]
org/10.1364/ao.54.008944
16 R. Arnaudov and Y. Angelov: Meas. Sci. Technol. 16 (2005) 2300. [Link]
17 N. Kolev, R. Arnaudov, and Y. Angelov: IEEE Technol. Appl. (2007) 464. [Link]
IDAACS.2005.283025
18 C. H. He, C. Yang, X. Wang, and Z. Wang: Appl. Opt. 52 (2013) 5303. [Link]
19 J. Collin, M. Kirkko-Jaakkola, and J. Takala: IEEE Trans. Instrum. Meas. 64 (2013) 230. [Link]
org/10.1109/tim.2014.2335921
20 D. K. Upadhyay and T. Khanna: IEEE Int. Conf. Electron. Comput. Commun. Technol. (2015). [Link]
org/10.1109/CONECCT.2015.7383938
21 Analog Devices, Inc.: [Link]
22 Butterworth Filter Design: [Link]

About the Authors

Xiaowen Cai received her Master’s degree in Control Theory and Control
Engineering from Xiangtan University of China, Xiangtan, China. She is
currently pursuing her Ph.D. degree at Beihang University. Her research
interests are related to inertial navigation and integrated navigation.

Chunxi Zhang received his Ph.D. degree in Optical Engineering from

Zhejiang University. He has been a professor at Beihang University since
1999. His current research interests are in integrated navigation and optical
fiber sensing technology.

Shuang Gao received her Ph.D. degree in Optical Engineering from the
School of Instrumentation Science and Opto-electronics Engineering (SIOE)
at Beijing University of Aeronautics and Astronautics, Beijing, China, in
2008. Her research interests focus on inertial navigation and integrated
navigation.
2608 Sensors and Materials, Vol. 31, No. 8 (2019)

Ruoyu Zhang received her B.S. degree in Automation from China University
of Petroleum. She is currently an M.S. candidate in Optical Engineering at
Beijing University of Aeronautics and Astronautics. Her research interests
focus on inertial navigation and integrated navigation.