Inertial navigation device based on multiple MEMS sensors

Abstract

The invention relates to an inertial navigation device based on multiple MEMS sensors. Comprising an MEMS sensor array and a control unit, wherein the MEMS sensor array comprises n MEMS sensors with a set layout, and n is greater than or equal to 2; each MEMS sensor is connected to the control unit through a control switch; the output end of each MEMS sensor is connected to the control unit; the control unit is configured to be used for collecting initial measurement data of each MEMS sensor, selecting a plurality of better target MEMS sensors from the MEMS sensor array, and correcting and fusing the initial measurement data of the target MEMS sensors to obtain inertial measurement data; and outputting a switch control signal to the control switch so as to open or close the MEMS sensor to output initial measurement data. According to the technical scheme, the initial measurement data of the better MEMS sensor is used for obtaining the inertial measurement data, so that the measurement precision is improved, and the use effect of inertial navigation is improved.

Description

Technical Field

[0001] This application relates to the field of inertial navigation technology, and in particular to an inertial navigation device based on multiple MEMS sensors. Background Technology

[0002] Inertial navigation is an important navigation and positioning technology. Inertial navigation devices typically have multiple built-in MEMS (Micro-Electromechanical System) sensors. These MEMS sensors can measure parameters such as acceleration and angular acceleration, and then integrate them to obtain displacement information and real-time attitude information, thus providing a strong impetus for the leapfrog development of navigation, positioning, attitude perception and other technologies.

[0003] Typically, in order to reduce equipment costs, inertial navigation devices use consumer-grade MEMS sensors. However, due to the limitations of the manufacturing process of MEMS devices, consumer-grade MEMS devices are prone to large measurement errors due to quantization errors, random walks, zero-bias instability noise, nonlinearity, and thermal instability. Therefore, it is necessary to improve inertial navigation devices based on multiple MEMS sensors to enhance the accuracy of the output inertial measurement data. Summary of the Invention

[0004] To address one of the aforementioned technical shortcomings, this application provides a navigation and positioning device based on multiple MEMS sensors, which has higher measurement accuracy.

[0005] An inertial navigation device based on multiple MEMS sensors includes: a MEMS sensor array and a control unit; wherein the MEMS sensor array includes n MEMS sensors arranged in a predetermined layout, where n≥2;

[0006] Each MEMS sensor is connected to the control unit via a control switch;

[0007] The output terminals of each MEMS sensor are connected to the control unit.

[0008] The control unit is configured to acquire initial measurement data from each MEMS sensor, select several superior target MEMS sensors from the MEMS sensor array, correct and fuse the initial measurement data of the target MEMS sensors to obtain inertial measurement data, and output a switch control signal to the control switch to turn the MEMS sensors on or off to output initial measurement data.

[0009] In one embodiment, the MEMS sensors of the MEMS sensor array are arranged at mutually perpendicular angles; wherein the xyz axes of the MEMS sensors are mutually perpendicular.

[0010] In one embodiment, the control unit is connected to each MEMS sensor and each control switch via an I2C bus;

[0011] The MEMS sensor communicates with the control unit via an I2C communication interface, and the control switch communicates with the control unit via an I2C switch interface; wherein the I2C communication interface and the I2C switch interface are arranged at intervals, and the I2C switch interface precedes the I2C communication interface;

[0012] The I2C communication interface and I2C switch interface operate under the control of clock pulse signals.

[0013] In one embodiment, the control unit is connected to the I2C bus via an encoding module, and the control switch is connected to the I2C bus via a decoding module;

[0014] The control unit sends a switch control signal to the encoding module;

[0015] The encoding module encodes the switch control signal into a serial switch control code and sends it to each decoding module via the I2C bus.

[0016] The decoding module is used to cache the serial switch control code and obtain its own corresponding code value from the serial switch control code and send it to the control switch.

[0017] In one embodiment, the control unit is configured to acquire initial measurement data of each MEMS sensor, obtain inertial measurement data obtained by weighted correction calculation of m target MEMS sensors determined in the previous acquisition and output it; and calculate the deviation value between the initial measurement data output by n MEMS sensors and the inertial measurement data respectively; select the m MEMS sensors with the smallest deviation value as the m target MEMS sensors for the next acquisition; where m≥2, nm≥1, and n represents the number of MEMS sensor arrays.

[0018] In one embodiment, the control unit is further configured to store m target MEMS sensors for calibration calculation using a first sensor set; during each calibration calculation, read the m target MEMS sensors determined in the previous acquisition from the first sensor set; and after each calibration calculation, add the m target MEMS sensors acquired in the next acquisition to the first sensor set.

[0019] In one embodiment, the control unit is configured to read the initial measurement data output by m target MEMS sensors in the first sensor set; correct the initial measurement data using independent pre-stage correction coefficients corresponding to the m target MEMS sensors; input the corrected initial measurement data into a pre-trained neural network model; and perform fusion processing using the neural network model to obtain inertial measurement data.

[0020] In one embodiment, the control unit is further configured to periodically count the number of times each MEMS sensor is added to a preset first sensor set; calculate the frequency corresponding to each MEMS sensor based on the counted number of times; and adjust the value m of the number of target MEMS sensors stored in the first sensor set according to the range of the frequency.

[0021] In one embodiment, the control unit is further configured to determine whether the frequency of each MEMS sensor is lower than a set threshold, and when the frequency of a MEMS sensor is lower than the set threshold, to add the corresponding MEMS sensor to a preset second sensor set; wherein the second sensor set is used to store disabled MEMS sensors.

[0022] In one embodiment, the control unit is further configured to acquire the current operating mode; if in high-precision mode, select and disable a number of MEMS sensors with lower frequencies; if in low-precision mode, select and disable a number of MEMS sensors with higher frequencies; wherein the operating mode includes high-precision mode and low-precision mode.

[0023] The above technical solution involves a control unit acquiring initial measurement data from each MEMS sensor, selecting several superior target MEMS sensors from the MEMS sensor array for calibration and fusion to obtain inertial measurement data, and outputting switch control signals based on the operating status of each MEMS sensor to the control switch to turn the designated MEMS sensor on or off. This technical solution uses initial measurement data from superior MEMS sensors to acquire inertial measurement data, improving measurement accuracy and enhancing the effectiveness of inertial navigation.

[0024] Furthermore, the control unit uses the m target MEMS sensors determined in the previous acquisition to perform weighted correction calculations to obtain inertial measurement data for output. It calculates the deviation between the initial measurement data output by each MEMS sensor and the inertial measurement data, and selects the m MEMS sensors with the smallest deviation as the m target MEMS sensors for the next acquisition. This technical solution can maintain the use of better initial measurement data during the inertial measurement data correction process, so that the output inertial measurement data achieves higher accuracy and improves the accuracy of inertial navigation.

[0025] Additional aspects and advantages of this application will be set forth in part in the description which follows, and will become apparent from the description or may be learned by practice of this application. Attached Figure Description

[0026] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein:

[0027] Figure 1 This is a schematic diagram of an inertial navigation device based on multiple MEMS sensors, according to one embodiment.

[0028] Figure 2 This is a schematic diagram of an example MEMS sensor layout;

[0029] Figure 3 This is a timing diagram of an example I2C interface.

[0030] Figure 4 This is a schematic diagram of an inertial navigation device based on multiple MEMS sensors, according to another embodiment. Detailed Implementation

[0031] The embodiments of this application are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain this application, and should not be construed as limiting this application.

[0032] Those skilled in the art will understand that, unless otherwise stated, the singular forms “a,” “an,” “the,” and “the” used herein may also include the plural forms. It should be further understood that the word “comprising” as used in this application’s specification means the presence of the stated feature, integer, step, or operation, but does not preclude the presence or addition of one or more other features, integers, steps, or operations.

[0033] Since inertial navigation devices based on multiple MEMS sensors often suffer from quantization errors, random walks, zero-bias instability noise, nonlinearity, and thermal instability, this application provides an improved inertial navigation device based on multiple MEMS sensors.

[0034] refer to Figure 1 As shown, Figure 1This is a schematic diagram of an inertial navigation device based on multiple MEMS sensors according to one embodiment. It mainly includes a MEMS sensor array and a control unit. The control unit may include a processor and other supporting devices, and can implement algorithm processing functions. The MEMS sensor array includes n MEMS sensors arranged in a predetermined layout, as shown in the figure. Each MEMS sensor can be numbered 1, 2, ..., n, and the n MEMS sensor array is named MEMS1, MEMS2, ..., MEMSn, where n ≥ 2. Each MEMS sensor is connected to the control unit via a control switch, i.e., each MEMS sensor is connected to a control switch, as shown in the figure, where the control switches are represented as S1, S2, ..., Sn. The output terminals of each MEMS sensor are connected to the control unit.

[0035] For example, as shown in the figure, the control unit is connected to each MEMS sensor and each control switch via an I2C bus; by simulating a parallel I2C bus, the initial measurement values ​​of the MEMS sensors at the same time can be accurately acquired, thereby improving the data processing accuracy.

[0036] The control unit is configured to acquire initial measurement data from each MEMS sensor, such as... Figure 1 In this process, the control unit can acquire initial measurement data from MEMS1, MEMS2, ..., MEMSn via the I2C bus; it then selects several superior target MEMS sensors from the MEMS sensor array, specifically m MEMS sensors from MEMS1, MEMS2, ..., MEMSn as target MEMS sensors. Here, the m MEMS sensors with superior performance can be selected as target MEMS sensors. For example, the m superior MEMS sensors can be the top m MEMS sensors whose initial measurement data deviates least from the inertial measurement data. The initial measurement data from the m target MEMS sensors is then corrected and fused to obtain inertial measurement data, and a switch control signal is output to a control switch to turn on or off the specified MEMS sensor to output initial measurement data. For example, the control unit can output a switch control signal to turn on the m target MEMS sensors and turn off the unselected n-m MEMS sensors. The switch control signal can be a level signal.

[0037] As described in the above embodiment, the control unit collects the initial measurement data of each MEMS sensor, selects several superior target MEMS sensors from the MEMS sensor array for calibration and fusion to obtain inertial measurement data, and outputs switch control signals of the operating status of each MEMS sensor to the control switch to turn the specified MEMS sensor on or off. This technical solution uses the initial measurement data of superior MEMS sensors to obtain inertial measurement data, which improves the measurement accuracy and enhances the effectiveness of inertial navigation.

[0038] In one embodiment, the MEMS sensors of each MEMS sensor array are arranged at mutually perpendicular angles; wherein the x, y, and z axes of the MEMS sensors are mutually perpendicular. For example... Figure 2 As shown, Figure 2 This is a schematic diagram of an example MEMS sensor layout. Taking MEMS1, MEMS2, and MEMS3 as examples, according to the marked x, y, and z axes of the MEMS sensors, the x-axis of MEMS1, the x-axis of MEMS2, and the x-axis of MEMS3 are mutually perpendicular. Similarly, the y-axis of MEMS1, the y-axis of MEMS2, and the y-axis of MEMS3 are mutually perpendicular. The z-axis of MEMS1, the z-axis of MEMS2, and the z-axis of MEMS3 are also mutually perpendicular.

[0039] As in the above embodiments, arranging the MEMS sensors of each MEMS sensor array at mutually perpendicular angles can cancel out directional errors between different devices and improve the measurement accuracy of the inertial navigation device.

[0040] In one embodiment, the control unit is connected to each MEMS sensor and each control switch via an I2C bus; the MEMS sensors communicate with the control unit via an I2C communication interface, and the control switches communicate with the control unit via an I2C switch interface; wherein the I2C communication interface and the I2C switch interface are set at intervals, and the I2C switch interface precedes the I2C communication interface; the I2C communication interface and the I2C switch interface operate under the control of a clock pulse signal.

[0041] refer to Figure 3 As shown, Figure 3This is a timing diagram of an example I2C interface. As shown in the figure, the I2C switch interfaces corresponding to MEMS1, MEMS2, ..., MEMSn are I2C11, I2C12, ..., I2C1n, and the corresponding I2C communication interfaces are I2C21, I2C22, ..., I2C2n. Under the control of the clock pulse signal CLK, a communication interaction between the I2C switch interface and the I2C communication interface is executed on the rising edge of the clock pulse signal CLK. For example, firstly, the I2C21 switch interface receives the switch control signal to turn the output of MEMS1 on or off, and then the I2C21 communication interface transmits the initial measurement data. And so on. All the interaction communication between the I2C communication interface and the I2C switch interface is completed within a certain number of clock pulses, which can achieve the effect of parallel I2C bus.

[0042] The technical solution of the above embodiment adopts a simulated parallel I2C bus scheme, which can accurately acquire MEMS measurement values ​​at the same time. Each MEMS sensor reads the measurement data at the same time, further improving the measurement accuracy of the inertial navigation device.

[0043] In one embodiment, such as Figure 4 As shown, Figure 4 This is a schematic diagram of an inertial navigation device based on multiple MEMS sensors according to another embodiment; the control unit is connected to the I2C bus through the encoding module M, and the control switch is connected to the I2C bus through the decoding module; the control unit sends a switch control signal to the encoding module M; the encoding module encodes the switch control signal into a serial switch control code and sends it to each decoding module through the I2C bus, as shown in the figure, the n decoding modules are represented as D1, D2, ... Dn respectively; the decoding module is used to buffer the serial switch control code and obtain its own corresponding code value from the serial switch control code and send it to the control switch.

[0044] For example, the control unit sends the switch control signal to the encoding module for encoding, which can be encoded into level signals such as "0" and "1". Then, according to the frequency of the clock pulse signal CLK, the switch control signal can be encoded into serial switch control codes corresponding to the I2C switch interfaces I2C11, I2C12, ..., I2C1n. On the rising edge of the clock pulse signal CLK, an I2C switch interface communication interaction is performed. The decoding modules, represented as D1, D2, ..., Dn, can decode their own code values ​​to complete the switch control of the corresponding MEMS sensor.

[0045] As in the above embodiments, the control unit can implement relevant logic control through the encoding module and the decoding module, thereby enabling fast and accurate control of the MEMS sensor and ensuring data processing stability.

[0046] To make the technical solution of this application clearer, more embodiments are described below.

[0047] In one embodiment, the inertial navigation device based on multiple MEMS sensors of this application has a control unit configured to acquire initial measurement data of each MEMS sensor, obtain inertial measurement data obtained by weighted correction calculation of m target MEMS sensors determined in the previous acquisition, and output the data; and calculate the deviation values ​​between the initial measurement data output by n MEMS sensors and the inertial measurement data respectively; and select the m MEMS sensors with the smallest deviation values ​​as the m target MEMS sensors for the next acquisition; where m≥2, nm≥1, and n represents the number of MEMS sensor arrays.

[0048] Specifically, at each sampling time, the m target MEMS sensors determined in the previous sampling are acquired, and the weighted correction calculation method is used to calculate the inertial measurement data of the target MEMS sensors. At the same time, after each sampling time, the deviation between the inertial measurement data and the initial measurement data output by each MEMS sensor is calculated, and the m MEMS sensors with the smallest deviation values ​​are selected as the target MEMS sensors for the next sampling. This forms a loop process.

[0049] For example, in the calculation process of the target MEMS sensor, assuming the initial measurement data are q1, q2, ..., qm, and Q represents the inertial measurement data, the corresponding calculated deviation values ​​are Δδ1 = |q1 - Q|, Δδ2 = |q2 - Q|, ..., Δδ m =|qm-Q|; finally, the set of deviation values ​​{△δ1, △δ2, …△δ} is obtained. m Then, based on the size, pair △δ1, △δ2, ... △δ m Sort the data and select the top m MEMS sensors with the smallest deviation values ​​for use in the next data acquisition.

[0050] For example, during the initial calculation, m target MEMS sensors can be randomly selected, or other selections can be made according to set rules; the value of m can be set according to application requirements, or other values ​​can be made according to set rules.

[0051] As described in the above embodiment, the control unit uses the m target MEMS sensors determined during the previous acquisition to perform weighted correction calculations to obtain inertial measurement data and outputs it. It calculates the deviation between the initial measurement data output by each MEMS sensor and the inertial measurement data, and selects the m MEMS sensors with the smallest deviation as the m target MEMS sensors for the next acquisition. In this way, better initial measurement data can be used during the inertial measurement data correction process, and the output inertial measurement data achieves higher accuracy, thereby improving the accuracy of inertial navigation.

[0052] In one embodiment, the inertial navigation device based on multiple MEMS sensors of this application has a control unit configured to further store m target MEMS sensors for calibration calculation using a first sensor set; during each calibration calculation, read the m target MEMS sensors determined in the previous acquisition from the first sensor set; and after each calibration calculation, add the m target MEMS sensors acquired in the next acquisition to the first sensor set.

[0053] For example, a first sensor set is preset to store m target MEMS sensors used for calibration calculation. Therefore, during each calibration calculation, the m target MEMS sensors determined in the previous acquisition can be read from the first sensor set. At the same time, after each calibration calculation, the m target MEMS sensors acquired in the next acquisition are added to the first sensor set.

[0054] As in the above embodiment, the first sensor set can cache m target MEMS sensors between two sampling times, thereby enabling the calibration calculation and the processing of the target MEMS sensors to be performed synchronously, thus improving data processing efficiency.

[0055] In one embodiment, the inertial navigation device based on multiple MEMS sensors of this application has a control unit configured to read the initial measurement data output by m target MEMS sensors in a first sensor set; correct the initial measurement data using independent pre-stage correction coefficients corresponding to the m target MEMS sensors; input the corrected initial measurement data into a pre-trained neural network model; and perform fusion processing using the neural network model to obtain inertial measurement data.

[0056] Specifically, the initial measurement data output by each of the m target MEMS sensors in the first sensor set is read. For example, if the m MEMS sensor arrays added to the first sensor set are MEMS1, MEMS2, ..., MEMSm, then the initial measurement data q1, q2, ..., qm output by these m MEMS sensors are read respectively. The initial measurement data is corrected using the independent pre-stage correction coefficients corresponding to the m target MEMS sensors. Assuming that the pre-stage correction coefficients corresponding to each MEMS sensor are k1, k2, ..., km, then the initial measurement data q1, q2, ..., qm output by sensors MEMS1, MEMS2, ..., MEMSm are corrected using k1, k2, ..., km respectively. The inertial measurement data calculation formula can be expressed as follows:

[0057] Q = K × ∑(k1q1 + k2q2 + ... + kmqm)

[0058] k1+k2+…km=1

[0059] Where Q is the inertial measurement data and K is the set correction coefficient.

[0060] In one embodiment, when fusing the various corrected initial measurement data, a neural network model can be pre-trained; the various corrected initial measurement data are input into the neural network model, and the neural network model is used to perform fusion processing to obtain inertial measurement data.

[0061] For example, a neural network model, such as an LSTM (Long Short-Term Memory) neural network model, can be constructed. By training and testing samples collected by MEMS sensors, a usable neural network model can be obtained. Compared with a weighted fusion correction scheme that uses a set correction coefficient K, using a neural network model for fusion processing can more finely denoise the inertial measurement data, thereby further improving the accuracy of the inertial measurement data.

[0062] As described in the above embodiment, at each sampling time, the inertial measurement data is calculated by weighted correction using m superior target MEMS sensors from the first sensor set. This can eliminate some random noise from MEMS sensors due to excessive errors under different states, further improving the accuracy of the correction scheme and thus further enhancing the reliability of the output inertial measurement data.

[0063] In one embodiment, the inertial navigation device based on multiple MEMS sensors of this application has a control unit configured to periodically count the number of times each MEMS sensor is added to a preset first sensor set; calculate the frequency corresponding to each MEMS sensor based on the counted number of times; and adjust the value m of the number of target MEMS sensors stored in the first sensor set according to the range of the frequency.

[0064] Specifically, the frequency of each MEMS sensor being added to the first sensor set is counted every time period T. Based on the count, the frequency of each MEMS sensor being added can be determined as h1, h2, ... h1. n The frequency h can basically reflect the measurement accuracy of a MEMS sensor, and then the root frequencies h1, h2, ... h n The value range of m is adjusted by storing the number of target MEMS sensors in the first sensor set.

[0065] For example, suppose there is a MEMS sensor array n=6, and m=5 in the default state. By statistically analyzing the frequencies h1 and h2 of sensors MEMS1 and MEMS2, the range of values ​​is much smaller than that of other MEMS sensors. In this case, m can be set to 4, that is, m=4.

[0066] As in the above embodiment, the value of m is dynamically adjusted based on the number of times each of the n MEMS sensors used in actual use is added to the first sensor set, thereby further improving the accuracy of the inertial measurement data calculated by the m MEMS sensors used for calibration.

[0067] In one embodiment, the inertial navigation device based on multiple MEMS sensors of this application has a control unit configured to further determine whether the frequency of each MEMS sensor is lower than a set threshold, and when the frequency of a MEMS sensor is lower than the set threshold, add the corresponding MEMS sensor to a preset second sensor set; wherein the second sensor set is used to store disabled MEMS sensors.

[0068] Specifically, a pre-defined second sensor set is used to store MEMS sensors that do not participate in the calibration calculation. When calculating the frequency of the statistical count, the measurement data output by some MEMS sensors may have large deviations under different states. Therefore, if the frequency of any MEMS sensor is lower than a set threshold, the measurement value of that MEMS sensor has a large deviation from the normal data, and it is added to the second sensor set.

[0069] As described in the above embodiment, by using a threshold to exclude MEMS sensors with large deviations from the MEMS sensors used in the calibration calculation, the influence of MEMS sensors with large deviations is further reduced, and the accuracy of the inertial measurement data in the calibration calculation is further improved.

[0070] In one embodiment, the inertial navigation device based on multiple MEMS sensors of this application has a control unit configured to further acquire the current operating mode; if in high-precision mode, select and turn off a number of MEMS sensors with lower frequency; if in low-precision mode, select and turn off a number of MEMS sensors with higher frequency; wherein the operating mode includes high-precision mode and low-precision mode.

[0071] Specifically, two modes can be preset: high-precision mode and low-precision mode. In high-precision mode, the MEMS sensor with a lower frequency is selected and turned off, and the low-precision MEMS sensor is not used for measurement. In low-precision mode, the MEMS sensor with a higher frequency is selected and turned off, and the high-precision MEMS sensor is not used for measurement. For example, the control unit can use a control switch to turn off the MEMS sensor by sending a set switch control signal to the corresponding MEMS sensor, or it can set a power switch to control the turn-off, that is, disable the MEMS sensor.

[0072] As in the above embodiments, different MEMS sensors with different accuracies can be selected by using different working modes, thereby optimizing the service life of the MEMS sensors. While ensuring the accuracy of the output inertial measurement data, the various MEMS sensors can be fully utilized, the core MEMS sensors can be protected, and their service life can be extended.

[0073] The above description is only a partial embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. An inertial navigation device based on multiple MEMS sensors, characterized in that, include: MEMS sensor array and control unit; wherein, the MEMS sensor array includes n MEMS sensors in a predetermined layout, n≥2; Each MEMS sensor is connected to the control unit via a control switch; The output terminals of each MEMS sensor are connected to the control unit. The control unit is configured to acquire initial measurement data from each MEMS sensor, select several superior target MEMS sensors from the MEMS sensor array, correct and fuse the initial measurement data of the target MEMS sensors to obtain inertial measurement data, and output a switch control signal to the control switch to turn the MEMS sensors on or off to output initial measurement data.

2. The inertial navigation device based on multiple MEMS sensors according to claim 1, characterized in that, The MEMS sensors in the MEMS sensor array are arranged at mutually perpendicular angles; wherein the xyz axes of the MEMS sensors are mutually perpendicular.

3. The inertial navigation device based on multiple MEMS sensors according to claim 1, characterized in that, The control unit is connected to each MEMS sensor and each control switch via an I2C bus. The MEMS sensor communicates with the control unit via an I2C communication interface, and the control switch communicates with the control unit via an I2C switch interface; wherein the I2C communication interface and the I2C switch interface are arranged at intervals, and the I2C switch interface precedes the I2C communication interface; The I2C communication interface and I2C switch interface operate under the control of clock pulse signals.

4. The inertial navigation device based on multiple MEMS sensors according to claim 1, characterized in that, The control unit is connected to the I2C bus via an encoding module, and the control switch is connected to the I2C bus via a decoding module. The control unit sends a switch control signal to the encoding module; The encoding module encodes the switch control signal into a serial switch control code and sends it to each decoding module via the I2C bus. The decoding module is used to cache the serial switch control code and obtain its own corresponding code value from the serial switch control code and send it to the control switch.

5. The inertial navigation device based on multiple MEMS sensors according to any one of claims 1-4, characterized in that, The control unit is configured to acquire initial measurement data of each MEMS sensor, obtain inertial measurement data from the m target MEMS sensors determined in the previous acquisition by performing weighted correction calculations and outputting the data; and calculate the deviation between the initial measurement data output by the n MEMS sensors and the inertial measurement data respectively; and select the m MEMS sensors with the smallest deviation as the m target MEMS sensors for the next acquisition; where m≥2, nm≥1, and n represents the number of MEMS sensor arrays.

6. The inertial navigation device based on multiple MEMS sensors according to claim 5, characterized in that, The control unit is further configured to store m target MEMS sensors for calibration calculation using a first sensor set; and to read the m target MEMS sensors determined in the previous acquisition from the first sensor set during each calibration calculation. After each calibration calculation, the m target MEMS sensors acquired in the next acquisition are added to the first sensor set.

7. The inertial navigation device based on multiple MEMS sensors according to claim 6, characterized in that, The control unit is configured to read the initial measurement data output by m target MEMS sensors in the first sensor set; correct the initial measurement data using independent pre-stage correction coefficients corresponding to the m target MEMS sensors; input the corrected initial measurement data into a pre-trained neural network model; and perform fusion processing using the neural network model to obtain inertial measurement data.

8. The inertial navigation device based on multiple MEMS sensors according to claim 7, characterized in that, The control unit is also configured to periodically count the number of times each MEMS sensor is added to a preset first sensor set; calculate the frequency of each MEMS sensor based on the count; and adjust the value m of the number of target MEMS sensors stored in the first sensor set according to the range of the frequency.

9. The inertial navigation device based on multiple MEMS sensors according to claim 8, characterized in that, The control unit is further configured to determine whether the frequency of each MEMS sensor is lower than a set threshold. When the frequency of a MEMS sensor is lower than the set threshold, the corresponding MEMS sensor is added to a preset second sensor set. The second sensor set is used to store disabled MEMS sensors.

10. The inertial navigation device based on multiple MEMS sensors according to claim 8, characterized in that, The control unit is also configured to acquire the current operating mode; if it is in high-precision mode, select and turn off a number of MEMS sensors with lower frequencies; if it is in low-precision mode, select and turn off a number of MEMS sensors with higher frequencies; wherein, the operating mode includes high-precision mode and low-precision mode.

Images