2024 IEEE 7th International Conference on Smart Technologies in Power Engineering and Electronics (STEE-2024)

MEMS Gyroscope Based on MPU-6050 Sensor and

ATmega328 Microcontroller
2024 IEEE 7th International Conference on Smart Technologies in Power Engineering and Electronics (STEE) | 979-8-3315-4099-9/24/$31.00 ©2024 IEEE | DOI: 10.1109/STEE63556.2024.10748180

Maksym Pokydko Ostap Oliinyk Vadym Tymchenko

Faculty of Electronics Faculty of Electronics Faculty of Electronics
Igor Sikorsky Kyiv Polytechnic institute Igor Sikorsky Kyiv Polytechnic institute Igor Sikorsky Kyiv Polytechnic institute
Kyiv, Ukraine Kyiv, Ukraine Kyiv, Ukraine
[email protected] https://orcid.org/0000-0003-3557-243 [email protected]

Abstract—This paper provides a reliable practical systems, including optical elements such as a fast-steering
implementation of the MEMS gyroscope on the base of MPU- mirror, whose performance is poor in open loop control mode.
6050 universal triaxial inertial measuring unit with a gyroscope
sensor and accelerometer and the microcontroller ATmega328. Proposed in work [3] improvement of double step
Presented interconnection of functional modules on the base of algorithm is to shorten the settling time and improve its quasi-
I2C protocol has allowed to improve the calibration process and static working bandwidth. Mostly all approaches to improve
make a zero offset for all orthogonal sensors in gyroscope that gyroscope parameters are in sensor design, materials and
in return reduce the additive measurement error. Analysis of integration on optical signal to provide a stable
electromechanical properties of universal gyroscope has shown communication link [4].
different properties of vibrating systems. Sensitive elements
signal processing common with developed error model, that The purpose of this paper is to offer a low-cost gyroscope
includes nonlinearity of sensors characteristics, has allowed to system with a maximum efficiency and maximum accuracy of
reduce the total measurement error of MEMS gyroscope up to angular displacement measurement on the base of MPU-6050
1%. sensor.

Keywords—MEMS gyroscope, ATmega328, MPU-6050 II. GYROSCOPE SYSTEM

There are two basic mode concepts of the vibration MEMS
I. INTRODUCTION gyroscope – a pressure mode that detects vibration movements
MEMS gyroscopes are among the most popular along the pressure axis in case of external rotation, and a
orientation sensors in the space today. Due to the low price, motion mode that produces continuous vibration along the
reliability and compactness, low energy consumption and moving axis [5-8].
general simplicity, they are used in a very wide range of
Most MEMS gyroscopes that are based on the vibration
modern electronic devices. With the development of
effect have a proof mass “m” with vertical and horizontal
technology, it is expected that MEMS gyroscopes will become
orthogonal degrees of freedom. The gyroscope continuously
even smaller, even more accurate and more energy-efficient.
vibrates at its resonant frequency, but when rotation occurs in
Now new materials and technologies are being developed for
the orthogonal direction, fluctuations will occur. When an
production that will increase the performance of these devices.
external rotation occurs, this rotation triggers the energy of the
Despite the fact that modern MEMS gyroscopes have many
primary vibration to shift in an orthogonal direction because
advantages, they still face problems of sensitivity to
of the Coriolis effect. A simplified scheme of two degrees of
environmental factors such as temperature, humidity, and
freedom of the mass vibration system is shown in Figure 1.
vibration. Robots are currently being developed to increase
stability and reliability in such conditions. Scientists are also
working on the creation of (IMU) - universal measuring units,
which are combinations of sensors such as a gyroscope,
accelerometer, magnetometer and some others.
A problem to reach a high resolution of MEMS systems is
usually based on resonant gyroscope using the mode-
localization effect in weekly coupled resonators for sensing
the input angular rate [1]. It increases the complexity of sensor
and its price, meanwhile the sensor cannot operate under
accelerations higher then 1-2g.
A lot of effort is applied to solve a problem of sensor
stability, one of it is adaptive method for MEMSs gyroscope Fig. 1. A basic illustration of a mass vibrating proof system [3].
based on a dynamic surface control with adaptive fuzzy and
sliding mode control for compensation the error of the fuzzy The equation (1) is used to calculate the Coriolis force and
approximation, which reduces the chattering phenomenon [2]. the most basic two degrees of freedom equation of motion for
MEMS vibrating gyroscopes. Where m is the inertial mass, x
Last researches and MEMS gyroscopes development are is the displacement of the driving motion, y is the
using a complex solution in control algorithm for MEMS displacement of the sensing motion, c is the damping

979-8-3315-4099-9/24/$31.00 ©2024 IEEE

TT3.39.1
Authorized licensed use limited to: Institut Teknologi Sepuluh Nopember. Downloaded on November 15,2024 at 02:21:23 UTC from IEEE Xplore. Restrictions apply.
 2024 IEEE 7th International Conference on Smart Technologies in Power Engineering and Electronics (STEE-2024)

coefficient, k is the stiffness constant, and η is the external

rotation rate. The driving, sensing, and Coriolis forces are
represented by the variables 𝐹𝐷, 𝐹𝑆, and 𝐹C, respectively [7].

𝑚 𝑐 𝑘𝑥 𝐹 2𝑚𝛺 (1)
Since the driving mode response is typically much larger
than the sensing mode response, the Coriolis force 2𝑚𝛺
becomes insignificant, and equation (1) can be expressed as
follows:
𝑚 𝑐 𝑘𝑥 𝐹 (2) Fig.3. The mechanical part of an MPU6050 [8]

The MPU-6050 has an I2C connection that it uses to

𝑚 𝑐 𝑘𝑦 𝐹 𝐹 (3) communicate with other no inertial digital sensors, such as
pressure sensors. The MPU-6050 has three 16-bit analog-to-
𝐹 2𝑚𝛺 (4) digital converters (ADCs) that digitize the gyroscope and
accelerometer outputs, respectively. For precise monitoring of
The simplest design of a MEMS vibratory gyroscope is both fast and slow displacements, the parts have a user-
shown in Figure 2, which includes a proof mass above the programmable accelerometer full-scale range of ±2g, ±4g,
substrate, driving and sensing electrodes, and a suspension ±8g, and ±16g, as well as a user-programmable gyroscope
support beam system with four fixed support pillars. An full-scale range of ±250, ±500, ±1000, and ±2000°/sec (dps)
oscillator with a predetermined amplitude is produced by [1]. The problem that appears is a zero-level shift because of
driving electrodes in the driving system, where a Coriolis mechanical properties of suspension system during operation
force detector forms the sensing system. The proof mass under different accelerations.
becomes an oscillator when the driving electrodes provide a
continuous and constant momentum along the driving axis On-chip 1024 Byte FIFO buffer allow to reduce power
direction. When the proof mass rotates externally, it shifts its consumption by reading the sensor data in bursts and then
oscillation motion to align with the driving axis' orthogonal transition to a low-power mode while the MPU gathers more
direction. The Coriolis force, which is generated by the data. Low-power Motion Interface applications in portable
combination of driving momentum and external rotation, is applications with less processing burden on the system central
detected by the detecting electrodes on the orthogonal axis. processor unit (CPU) can be uniquely enabled by the MPU-
6050, since it has all the sensor components and on-chip
computing required to support a wide range of motion.
By offloading the expensive Motion Processing
computing needs from the system CPU and providing an
integrated Motion Fusion output, the DMP in the MPU-6050
reduces the need for frequent polling of the motion sensor
output [1].
Communication with every register on the device is
accomplished using I2C. An integrated temperature sensor
and an on-chip oscillator with ±1% variation throughout the
operating temperature range are two additional features. Using
Fig.2. Design of a single vibrating gyroscope proving mass [7]. its proprietary and volume-proven Nasiri-Fabrication
technology, which integrates MEMS wafers with companion
Including a 3-axis gyroscope, 3-axis accelerometer, and a CMOS circuits by wafer-level bonding, InvenSense has
Digital Motion ProcessorTM (DMP) in a compact 4x4x0.9mm reduced the size of the MPU-6050 package to a breakthrough
case, the MPU-6050 is the second in a series of integrated 6- 4x4x0.9mm (QFN) [1]. Offered solution shows a high
axis Motion Tracking devices. To produce a full 9-axis performance, the lowest vibration and heat noise, low cost.
Motion FusionTM output, it directly accepts inputs from an A programmable low-pass filters for accelerometers,
external 3-axis compass via its dedicated I2C sensor bus [8]. gyroscopes, and the on-chip temperature sensor is used for
The 6-axis integration, on-board Motion Fusion TM, and shock resistance and measurement error control. For
run-time calibration firmware of the MPU-6050 Motion versatility in power supply, the MPU-6050 operates on a VDD
Tracking device allow for manufacturers to do away with the power supply voltage range of 2.375V–3.46V. The MPU-
expensive and time-consuming process of selecting, 6050 contains a VLOGIC reference pin that regulates the logic
qualifying, and system-level integrating discrete devices as it levels of its I2C interface in addition to its analog supply pin,
shown in Figure 3. VDD. The VLOGIC voltage could be either VDD or
1.8V±5%. Figure 4 shows the MPU-60X0 motion tracking
device's block diagram with a factory calibration.

TT3.39.2
Authorized licensed use limited to: Institut Teknologi Sepuluh Nopember. Downloaded on November 15,2024 at 02:21:23 UTC from IEEE Xplore. Restrictions apply.
 2024 IEEE 7th International Conference on Smart Technologies in Power Engineering and Electronics (STEE-2024)

Since I2C is a serial communication protocol, data is sent

over a single wire (the SDA line) bit by bit.
Data is sent over I2C in messages. Data frames are used to
separate messages. Every message consists of one or more
data frames containing the data being sent, as well as an
address frame with the slave's binary address. Between each
data frame, the message also contains ACK/NACK bits,
read/write bits, and start and stop conditions as it is shown in
Figure 5:

Fig.5. I2C protocol message diagram

Fig.4. MPU-60X0 block diagram [5]
Here:
The main primary components and functionalities of the  Start Condition: Prior to the SCL line changing from high
MPU-6050 are as follows: to low voltage, the SDA line changes from a high voltage
 Three-axis MEMS rate gyroscope sensor equipped level to a low voltage level.
with signal conditioning and 16-bit ADCs  Stop Condition: The SCL line changes from low to high
 Three-axis MEMS accelerometer sensor equipped voltage before the SDA line changes from low to high
with signal conditioning and 16-bit ADCs voltage.
 Serial communications interfaces: primary I2C and  Address Frame: When the master wishes to communicate
SPI (MPU-6000 only) with a slave, the slave is uniquely identified via a 7 or 10
 Extra I2C serial interface for external magnetometer bit sequence.
and more sensors  Read/Write Bit: One bit that indicates whether data is
 Gyroscope & Accelerometer Self-test being sent from the master to the slave (low voltage level)
 FIFO or requested from it (high voltage level).
 Digital Motion Processor (DMP) engine  ACK/NACK Bit: A message's acknowledge/no-
acknowledge bit comes after every frame. An ACK bit is
Accelerometers, gyroscopes, and other third-party sent back to the sender by the receiving device if an
sensors, such magnetometers, provide data to the DMP, which address frame or data frame was successfully received.
then processes it. The resultant data can be queued in a FIFO
or read from the DMP's registers. One of the external pins of Since I2C lacks SPI's slave choose lines, it requires an
the MPU is accessible to the DMP, and it can be utilized to additional means of informing the slave that data is being
cause interrupts [1]. transmitted to it rather than another slave. It addresses to
Motion processing algorithms typically yield accurate accomplish this. In a new message, the address frame appears
results with little latency when they are run at a fast rate, first, always following the start bit.
typically 200 Hz. Even in cases when an application updates Every slave that is connected to the master receives the
at a significantly slower rate—a low-power user interface, for address of the slave it wishes to interact with. Next, each slave
instance, might update as slowly as 5Hz—motion processing verifies that the address it received from the master matches
must nevertheless occur at 200Hz. Utilizing the DMP as a tool its own address. It returns a low voltage ACK bit to the master
can help reduce power consumption, streamline software if the addresses match. The slave does nothing and the SDA
architecture, streamline time, and preserve MIPS on the host line stays high if the addresses do not match [4].
processor unit. For faster data analysis and increase the accuracy, reading
and writing of the bit is done as follows: a single bit at the end
of the address frame tells the slave whether the master wishes
III. PROTOCOL I2C FOR GYROSCOPE CONTROL
to write data to it or read data from it. The read/write bit is a
In this paper I2C protocol has allowed to have numerous low voltage level that the master uses to convey data to the
masters commanding one or more slaves, as well as the ability slave. The bit is at a high voltage level if the master is asking
to attach multiple slaves to a single master. When it is the slave for data.
necessary, multiple microcontrollers to log data to a single The initial data frame can be transferred after the master
memory card or show text on a single LCD. recognizes the slave's ACK bit.
Similar to UART communication, data transmission The most important bit is sent first in a data frame that is
between devices using I2C only requires two wires: always eight bits long. An ACK/NACK bit is added right after
 SDA (Serial Data) – The line for the master and slave each data frame to confirm that it was successfully received.
to send and receive data. Before the next data frame can be delivered, the ACK bit
needs to be received by the master or the slave, depending on
 SCL (Serial Clock) – The line that carries the clock
who is transmitting the data [4].
signal. The transmission can be stopped by the master sending a
stop condition to the slave when all data frames have been

TT3.39.3
Authorized licensed use limited to: Institut Teknologi Sepuluh Nopember. Downloaded on November 15,2024 at 02:21:23 UTC from IEEE Xplore. Restrictions apply.
 2024 IEEE 7th International Conference on Smart Technologies in Power Engineering and Electronics (STEE-2024)

transmitted. A low to high voltage shift on the SDA line nowhere connected (or connected to GND) - the address of the
following a low to high voltage transition on the SCL line, sensor on the i2c bus will be 0x68, if it is plugged into the
with the SCL line staying high, is the stop condition. power (VCC) – the address will be 0.x69 [1].
IV. EXPERIMENTAL SETUP AND GYROSCOPE DESIGN
The Arduino Nano board acts as the hardware bandage of
the ATmega328 microcontroller as it is shown in Figures 6.1-
6.3:

Fig.7. MPU-6050 sensor banding scheme principle

For the firmware of the microcontroller it is used the

Arduino IDE development environment which uses C++
Fig.6.1. ATmega328 microcontroller binding scheme principle
code. Also it is used a third-party library i2cdev developed by
Jeff Rowberg. The example code MPU6050_DMP6 is used
from the same library, it uses a digital motion processor
(DMP). It is a built-in MPU6050 processor that combines data
from the accelerometer and the gyroscope to find the exact
angle of inclination.
The IMU needs to be calibrated before it is used. Since the
robot's IMU will not be exactly aligned with the ground,
offsets will need to be produced by taking a number of data
readings from the accelerometer and gyroscope. The offsets
Fig.6.2. Principal power circuit ATmega328 offer a translation from the Body Frame to the Inertial Frame
from a physics standpoint. The Inertial Frame is the frame
from which all angle and velocity calculations are made, and
the Body Frame is the IMU mounted on the robot, as it is
shown in Figure 8.

Fig.8. Inertial Frame and the Body Frame

To match inertial and body frames the calibration code is

Fig.6.3. Additional communication hardware part for ATmega328
used. It allows to get six shift values which are writen in the
The GY-521 module appears as the bandage of the MPU- EEPROM and are used in the main code.
6050 sensor (6DOF). Calibration process of the developed gyroscope consists of
the gyro and accelerometer offsets are set to zero at the start
The Figure 7 shows that in ATmega328 the ports
of the operation. After that, the mean value for each offset is
responsible for sequential data output and the line on which
made by using 1000 IMU readings. After that, the IMU uses
the tactical signal is transmitted are A4 (SDA) and A5 (SLC) these values as the new offsets. Until the calibration is within
respectively. The GY-521 module is connected to them, a a predetermined tolerance, the calibrate procedure will keep
constant voltage of 3.3V is applied in accordance with the taking the mean IMU values.
datasheet and connect the GND contacts of the module and
the microcontroller. The 3,3 V is supplied from the After the code is executed, the following values are
microcontroller 5V since the module has a voltage stabilizer. obtained:
In addition, a AD0 pin is displayed on the module. If it is - Sensor readings with offsets: 2 0 16392 -1 2 0
- Your offsets: -4806 -296 991 57 -35 20

TT3.39.4
Authorized licensed use limited to: Institut Teknologi Sepuluh Nopember. Downloaded on November 15,2024 at 02:21:23 UTC from IEEE Xplore. Restrictions apply.
 2024 IEEE 7th International Conference on Smart Technologies in Power Engineering and Electronics (STEE-2024)

- Data is printed as: acelX acelY acelZ giroX giroY giroZ For the evaluation of calibration procedure, it is necessary
to analyze the experimental results to find sensitivity S and
The MPU6050 can be used to extract meaningful data in offset O.
two different methods. One method is to compute the new Solution of equations (1)-(2) for calculation the amplitude
orientation by reading the raw sensor data values, just as it was of sensitive elements oscillations 𝐴 𝑡 after time 𝑡 of
did during the calibration procedure. Extraction of data from rotation (acceleration) is follow:
the MPU's onboard Digital Motion Processor (DMP) is the ∙
second approach. 𝐴 𝑡 2 𝑒 (7)
There are multiple translations that must be performed in
order to obtain the pitch angle from the raw sensor data. were 𝛾 – an attenuation coefficient, 𝛺 – angular frequency of
 The angular velocity—the rate at which the angular forced oscillations, 𝛺 – minimal measured angular
position changes over time—is provided by the velocity, 𝐴 – displacement along X-axis, 𝑣 – natural angular
gyroscope. frequency of oscillator.
 Consequently, in order to obtain the angle position,
V. EXPERIMENTAL RESULTS
it must be integrated across time.
 You must use the IMU offsets to convert readings During the test the following four parameters were
from the body frame to the inertial frame before analyzed:
performing any computations.  Minimal measurement errors
 The raw IMU sensor readings must be converted to  Accumulation of poxibba
degrees.  Maximum rotation bits
Processing that would typically need to be done on the  Artifacts during measurement
CPU is offloaded to the DMP. It keeps an internal buffer full The measurement of a slow (1,43 rad/s) rotation for each
of information from the accelerometer and gyro that it uses to three dimentions is presented in Figure 10.
calculate orientation. The FiFo buffer structure of the DMP is
shown in Figure 9.

Figure 9: This is the internal FiFo buffer format of the DMP

A Quaternion is used to represent the orientation.

Quaternions are a computationally efficient means of Fig.10. Slow rotation measurement (red line is for X- axis of rotation, blue –
representing an object's orientation in three dimensions. They Y-axis and green – Z-axis respectively). Black line – theoretical sensitivity.
help prevent a problem known as Gimbal Lock, which arises
when you spin through an angle greater than 90°. The DMP Analyzing the experimental results presented in Fig.10.,
records its calculations in quaternions to guarantee that the and using the equations (5)-(6) for maximum angle of 180° it
orientation data is helpful for all applications. With the help of is possible to finalize error mathematical model for designed
Jeff Rowberg's program, can quickly translate quaternions gyroscope:
into other relevant data, such linear acceleration and Euler 𝑆 514,3 𝑔𝑆
angles. 𝑂 0 (8)
In this paper Euler's angles are used to obtain the output 𝜑 514,3 ∙ 𝑡 𝑨/𝑟 15°
data. Quaternion values are transferred to a function from the
same dmpGetEuler() library to be converted to angles. Since were 𝑖𝜖 𝑥, 𝑦, 𝑧 , 𝑔 2,4,8,16 [m/s2] respectively.
the output data is given in radians, the conversion to degrees As the Fig.10 shows, measurement errors are in range 2-
are made. 15%. The best indicators are achieved for X-axis and by using
Presented calibration method also involves approximation
built-in (DMP) as it takes most of the calculations on itself.
the instrument errors additionally to the gyroscope
Due to developed calibration procedure the offset error 𝑂 is
mathematical model. Error model includes 9 parameters:
practically zero and can be neglected.
triaxial acceleration A0=[ax0,ay0,az0]T, measured acceleration
In the next test the sensor was rotated and stopped, it
A=[ax,ay,az], diagonal sensitivity matrix S=[sx,sy,sz], offsets
allows to understand how much noise is present in a system,
vector O=[ox,oy,oz]T. Connection between real and measured
and how large is the accumulation of errors. The sensitivity of
triaxial physical quantities can be expressed as follows:
sensors limits the resolution of gyroscope. To find the main
source of noise at 3dB level it is possible to analyze the
𝑨 𝑺 𝑨 𝑶 (5)
dynamics between thermal and mechanical energy of
Angular position φ=[ φx, φy, φz] is calculated using a time
rezonators. The displacement due to the noise is described by
t, that the system was accelerated and the distance r between
expression [10]:
the axis of rotation and acceleration sensor:
𝝋 𝑡 𝑨/𝑟 (6) 𝑥
∆
(9)

TT3.39.5
Authorized licensed use limited to: Institut Teknologi Sepuluh Nopember. Downloaded on November 15,2024 at 02:21:23 UTC from IEEE Xplore. Restrictions apply.
 2024 IEEE 7th International Conference on Smart Technologies in Power Engineering and Electronics (STEE-2024)

where T is the surrounding temperature, 𝑘 – Boltsman accuracy up to 0,1° using an autocorrelation function as
constant, 𝛺⬚ and 𝑚 are the frequency and mass at resonance, follow:
and 𝑄 is the quality factor. ∙ ∙
𝐴 𝑡 2 𝑒 𝑑𝑡 (10)
The experimental results of sensitive elements oscillations
are presented in Figures 11.1-11.3. were 𝜏 is a shift time, 𝑡 is a time oscillation under 3dB level,
𝜎 is a dispersion of oscillations.
For usual opperation, the angle value can only be used
with an accuarasy up to 1° since the sensor is noisy enough,
and measured values less than 1° are very unstable.
VI. CONCLUSIONS
In this paper a new calibration method of the MPU-6050
gyroscope is presented. Under the control of ATmega328
microcontroller it became possible to avoid the offset errors
before measurement starts. Fast communication and data
analysis is realized using I2C protocol.
The firmware of the microcontroller, the calibration of the
Fig.11.1. X-axis swing sensor shows its semi liner measurement parameters.
The presented gyroscope device is assembled of minimum
components and show minimum error of 2°, maximum angle
of inclination of 180°.
Research of MPU-6050 gyroscope dynamic parameters
have shown different thermomechanical properties and noise
signals of triaxial sensors. To improve the accuracy of
gyroscope measurement it is shown that the error model can
be improved by using the experimental data of sensor
research. Excluding the nonlinearity of Y and Z gyroscope
sensors common with presented calibration it is possible to
minimize the measurement error up to 0,1°.
REFERENCES
Fig.11.2. Y-axis swing [1] S. A. R. Bukhari, M. M. Saleem, A. Hamza and S. A. Bazaz, "A Novel
Design of High Resolution MEMS Gyroscope Using Mode-
Localization in Weakly Coupled Resonators," in IEEE Access, vol. 9,
pp. 157597-157608, 2021, doi: 10.1109/ACCESS.2021.3123152.
[2] D. Lei, T. Wang, D. Cao and J. Fei, "Adaptive Dynamic Surface
Control of MEMS Gyroscope Sensor Using Fuzzy Compensator," in
IEEE Access, vol. 4, pp. 4148-4154, 2016, doi:
10.1109/ACCESS.2016.2596538.
[3] Y. Zihao, W. Lihao, W. Yang, Z. Yonggui, L. Yichen and W. Zhenyu,
"Control of a MEMS Fast Steering Mirror With Improved Quasi-Static
Performance," in IEEE Access, vol. 11, pp. 95307-95314, 2023, doi:
10.1109/ACCESS.2023.3311139.
[4] S. Jeon and H. Toshiyoshi, ‘‘MEMS tracking mirror system for a
bidirectional free-space optical link,’’ Appl. Opt., vol. 56, no. 24, pp.
6720–6727, 2017.
[5] Data Sheet “MPU-6000 and MPU-6050 Product Specification
Revision 3.4” pp. 7-31.
Fig.11.3. Z-axis swing [6] Data Sheet and Schematics “Arduino Nano V3.3”, 2011, 57p.
[7] Weqas A. G, Ian H. Ilyas M. and Kristoffer McKee “A Review of
From Figures 11.1-11.3 and expression (9) it is clear that MEMS Vibrating Gyroscopes and Their Reliability Issues in Harsh
all three sensors in triaxial accelerometer have different Environments” Licensee MDPI 2022, pp. 1-12.
[8] Leens F “An Introduction To I2C And SPI Protocols” IEEE
electromechanical properties, resonant frequencies, quality Instrumentation & Measurement Magazine. vol 12. pp. 8-12.
factors and different noise. It also explains the different [9] S. Kim and M. Kim, "Rotation Representations and Their
measurement errors in Figure 10. Conversions," in IEEE Access, vol. 11, pp. 6682-6699, 2023, doi:
On the base of equations (1-3) and (9) and developed 10.1109/ACCESS.2023.3237864.
[10] R. Perelló-Roig, J. Verd, S. Bota, and J. Segura, “Thermomechanical
calibration method it is required additional signal processing noise characterization in fully monolithic CMOS-MEMS resonators”
from all gyroscope sensors sententiously. It allows to get Sensors, vol. 18, no. 9, p. 3124, Sep. 2018, doi: 10.3390/s18093124.

TT3.39.6
Authorized licensed use limited to: Institut Teknologi Sepuluh Nopember. Downloaded on November 15,2024 at 02:21:23 UTC from IEEE Xplore. Restrictions apply.