MEMS sensor air pressure drift pre-calibration method based on outdoor cold state environment

Abstract

The invention relates to the field of MEMS sensors, and discloses an MEMS sensor air pressure drift pre-calibration method based on an outdoor cold environment, and the method comprises the following steps: S1, low air pressure environment construction and natural recovery, and S2, dynamic data synchronous acquisition. Pollution of vibration of the low-pressure box to calibration data of the sensor is fundamentally eliminated, the purity of the calibration data is ensured, the air pressure drift correction coefficient obtained through fitting truly reflects the physical relation between air pressure and sensor output, and the authenticity and accuracy of the calibration coefficient are remarkably improved. Aiming at the problem that the calibration process in the prior art is not natural, the air pressure natural recovery process is closer to the air pressure change scene experienced by the sensor during actual outdoor application, so that the generalization ability and the compensation effect of the calibrated air pressure coefficient in the actual environment are better, and the applicability problems that the precision required by a laboratory is high and the field error is large are effectively solved.

Description

Technical Field

[0001] This invention relates to the field of MEMS sensors, and more particularly to a method for pre-calibrating the barometric pressure drift of MEMS sensors based on outdoor cold environments. Background Technology

[0002] Microelectromechanical systems (MEMS) sensors have been widely used in consumer electronics, automotive electronics, industrial monitoring, and meteorological observation due to their advantages such as small size, low power consumption, low cost, and ease of integration. Among them, MEMS barometric pressure sensors, as an important environmental sensing sensor, are often used in scenarios such as altitude measurement, weather forecasting, assisted navigation, and indoor and outdoor positioning.

[0003] In existing technologies, the calibration of MEMS sensor pressure drift is usually performed in a high-altitude low-pressure test chamber or a programmable barometric chamber. However, this method has significant drawbacks: the vacuum pump, compressor, and other motor equipment integrated in the low-pressure test chamber generate continuous vibrations during operation. These vibrations couple into the sensor output signal, forming noise, and also change the stress distribution of the sensor chip and PCB, interfering with the accurate extraction of pure pressure drift characteristics. The test chamber maintains a constant low pressure by actively pumping / filling air, and the rate and path of its pressure change differ from the natural pressure changes experienced by the sensor in actual outdoor applications, resulting in a decrease in the applicability of the calibrated pressure coefficient in real-world environments. Steady-state calibration at multiple pressure points is time-consuming, and high-precision programmable barometric equipment is expensive, which is not conducive to rapid factory calibration in large-scale production. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a method for pre-calibrating the air pressure drift of MEMS sensors based on outdoor cold environments, thus solving the above problems.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a pre-calibration method for barometric pressure drift of MEMS sensors based on outdoor cold environments, comprising the following steps: S1. Low-pressure environment construction and natural recovery: Place the MEMS sensor to be calibrated in a sealed calibration chamber, first reduce the air pressure in the chamber to the target low value, then stop all active air pressure control and vibration generation equipment, and allow the air pressure in the chamber to slowly recover to the ambient air pressure under natural conditions. S2. Dynamic Data Synchronous Acquisition: During the natural recovery process of air pressure, the raw output data of the MEMS sensor is acquired synchronously and continuously. and the ambient air pressure data inside the cavity ; S3. Data Preprocessing and Feature Construction: Preprocessing the collected raw data... and Denoising and alignment processes were performed, and the rate of change of air pressure was calculated. Construct a dataset for modeling, including air pressure. Rate of change of air pressure and the corresponding sensor output change ,in , This refers to the sensor's output under known reference conditions. S4. Decoupling Physical-Mathematical Model Construction and Fitting: Based on the physical laws of air pressure and drift, a correction model integrating linear and nonlinear relationships is established.

[0006] in, The change in sensor output For ambient air pressure, The rate of change of air pressure, , , , Here are the pressure drift correction coefficients to be fitted. For model residuals; A nonlinear regression algorithm is used to evaluate the model parameters. , , , Perform fitting; S5. Parameter Verification and Storage: The fitted pressure drift correction coefficients are then stored. , , , ; S6. Long-term correction interface: Provides an updateable interface for the calibration parameters by dynamically correcting the factory calibration parameters based on online monitoring data.

[0007] Preferably, in step S1, after the active air pressure control device is stopped, the calibration chamber maintains limited gas exchange with the outside world so that the air pressure rises slowly, simulating the natural air pressure change process outdoors.

[0008] Preferably, the limited gas exchange is achieved through micropores or controllable valves.

[0009] Preferably, in step S3, the data preprocessing includes using a moving average filter to remove high-frequency noise and using the 3σ criterion to remove data anomalies caused by accidental interference.

[0010] Preferably, in step S4, when establishing the model, temperature T is also introduced as a compensation variable to construct a temperature-dependent pressure coefficient model:

[0011] in, , , , This is a coefficient that varies with temperature; By repeating steps S1-S3 at multiple temperature points, a functional relationship between the coefficients and temperature is fitted.

[0012] Preferably, in step S5, the variance of the residual distribution of the fitted model is also stored as prior knowledge for use in setting the process noise of the subsequent online correction algorithm.

[0013] Preferably, in step S5, the non-volatile memory is a Flash memory.

[0014] Preferably, in step S6, the updatable interface is implemented through a communication interface, allowing the host computer to inject new air pressure correction coefficients.

[0015] Preferably, the communication interface is an RS485 interface.

[0016] Preferably, the MEMS sensor is a MEMS accelerometer or tilt sensor.

[0017] This invention provides a method for pre-calibrating the barometric pressure drift of MEMS sensors based on outdoor cold environments. Compared with existing technologies, it has the following advantages: In this invention, by stopping all active air pressure control and vibration-generating equipment, the air pressure is allowed to recover naturally, fundamentally eliminating the contamination of sensor calibration data by low-pressure box vibration, ensuring the purity of the calibration data, and making the fitted air pressure drift correction coefficient truly reflect the physical relationship between air pressure and sensor output, significantly improving the authenticity and accuracy of the calibration coefficient; addressing the problem of unnatural calibration processes in existing technologies, the natural air pressure recovery process is closer to the air pressure change scenarios experienced by sensors in actual outdoor applications, making the generalization ability and compensation effect of the calibrated air pressure coefficient better in actual environments, effectively solving the applicability problem of high precision requirements in laboratories and large field errors; 2. In this invention, there is no need to maintain multiple steady-state pressure points for a long time. Continuous data from low pressure to ambient pressure can be obtained in a single natural recovery process. The calibration cycle is short, and the equipment requirements are low. It does not require a high-precision programmable pressure source, which is conducive to rapid batch calibration on the production line, significantly reducing equipment investment and production costs. The decoupled physical and mathematical model that integrates linear and nonlinear terms can more accurately characterize the complex influence of air pressure and its rate of change on the sensor output. It also supports temperature compensation and long-term parameter updates, enabling the sensor to have lifelong learning capabilities and adapt to slow performance changes during long-term use, providing a reliable guarantee for high-precision outdoor static measurement. Attached Figure Description

[0018] Figure 1 This is a flowchart of the method for pre-calibrating the air pressure drift of MEMS sensors based on outdoor cold environment proposed in this invention; Figure 2 A comparison graph showing the change of sensor output signal over time; Figure 3 This is a comparison chart of sensor output signals changing with air pressure. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] Please see Figures 1-3 The present invention provides two technical solutions, specifically including the following embodiments: Example 1: A method for pre-calibrating the barometric pressure drift of MEMS sensors in outdoor cold environments includes the following steps: S1. Low-pressure environment construction and natural recovery: The MEMS sensor to be calibrated is placed in a sealed calibration chamber. The MEMS sensor is a MEMS accelerometer or tilt sensor. First, the air pressure in the chamber is reduced to the target low value. Then, all active air pressure control and vibration generation equipment are stopped, allowing the air pressure in the chamber to slowly recover to the ambient air pressure under natural conditions. After the active air pressure control equipment is stopped, the calibration chamber maintains limited gas exchange with the outside world to make the air pressure rise slowly, simulating the natural air pressure change process outdoors. Limited gas exchange is achieved through micropores or controllable valves. S2. Dynamic Data Synchronous Acquisition: During the natural recovery process of air pressure, the raw output data of the MEMS sensor is acquired synchronously and continuously. and the ambient air pressure data inside the cavity ; S3. Data Preprocessing and Feature Construction: Preprocessing the collected raw data... and Denoising and alignment processes were performed, and the rate of change of air pressure was calculated. Construct a dataset for modeling, including air pressure. Rate of change of air pressure and the corresponding sensor output change ,in , For the sensor output under known reference conditions, data preprocessing includes using a moving average filter to remove high-frequency noise and using the 3σ criterion to remove data outliers caused by accidental interference. S4. Decoupling Physical-Mathematical Model Construction and Fitting: Based on the physical laws of air pressure and drift, a correction model integrating linear and nonlinear relationships is established.

[0021] in, The change in sensor output For ambient air pressure, The rate of change of air pressure, , , , Here are the pressure drift correction coefficients to be fitted. For model residuals; A nonlinear regression algorithm is used to evaluate the model parameters. , , , Perform fitting; When building the model, temperature T was also introduced as a compensation variable to construct a temperature-dependent pressure coefficient model:

[0022] in, , , , This is a coefficient that varies with temperature; By repeating steps S1-S3 at multiple temperature points, a functional relationship between the coefficients and temperature is fitted.

[0023] S5. Parameter Verification and Storage: The fitted pressure drift correction coefficients are then stored. , , , Furthermore, the variance of the residual distribution of the fitted model is also stored as prior knowledge for use in setting the process noise of the subsequent online correction algorithm. The non-volatile memory is Flash memory. S6. Long-term correction interface: Based on online monitoring data, the factory calibration parameters are dynamically corrected, providing an updateable interface for the calibration parameters. The updateable interface is implemented through a communication interface, allowing the host computer to inject new air pressure correction coefficients. The communication interface is an RS485 interface.

[0024] Example 2: Based on Example 1, this example uses a MEMS tilt sensor for slope monitoring in plateau areas to illustrate its factory pre-calibration process for air pressure drift. Step 1: Set up the calibration environment: Place the sensor in a well-insulated, sealed calibration chamber. The chamber is equipped with a high-precision pressure sensor (e.g., ±10 Pa) and a temperature sensor. First, use a controllable vacuum pump to rapidly reduce the chamber pressure to 65 kPa (simulating an altitude of approximately 3500 meters), and stabilize the temperature at 25°C. Step 2: Initiate natural recovery: Once the target pressure is reached, immediately shut down the vacuum pump and all fans inside the chamber, leaving only the data acquisition equipment powered. Allow outside air to slowly enter through a controllable micro-orifice valve (adjustable orifice size) on the chamber, causing the chamber pressure to rise naturally at a rate of approximately 10 Pa / min to the local ambient pressure (approximately 101.3 kPa). Step 3: Synchronous Data Acquisition: During the natural recovery of air pressure (approximately 6 hours), the output tilt angle value Y of the MEMS sensor and the air pressure value P inside the cavity are synchronously acquired at a frequency of once per second. Step 4: Data Processing and Modeling: Align and low-pass filter the collected data. Using the sensor output at the initial time (65 kPa air pressure) as the baseline Y0, calculate the drift ΔY at each time step. Then process the data... Import the calculation software. Build the model: A set of calibration coefficients was obtained by fitting using the nonlinear least squares method: = -0.0015, =0.0008, =2.1e-6, =1.5e-7; Step 5: Parameter Storage and Verification: [The text abruptly ends here, likely due to an incomplete sentence or a formatting error.] , , , The model residual variance is written into the sensor's internal Flash memory. To verify this, steps one through three were repeated under another temperature condition (15°C) (the initial air pressure was changed to 70 kPa), and the stored coefficients were used for real-time correction. The standard deviation of the corrected tilt angle data was reduced by 88% compared to the original data, proving that the calibration was effective. Step Six: Provide an update interface: Reserve functions in the sensor firmware to allow the host computer to inject new air pressure correction coefficients through a communication interface (such as RS485) so that the factory parameters can be optimized in the future by combining online monitoring data.

[0025] Example 3: This example addresses the three major shortcomings of traditional high-altitude low-pressure test chamber calibration methods: vibration interference, unnatural process, and low efficiency and cost. It demonstrates the targeted improvement effect of the method of the present invention in industrial mass production. S1. Low-Pressure Environment Construction and Natural Recovery: Improvements to Address Vibration Interference: A 12-station parallel calibration system was designed, with each station equipped with an independent sealed calibration chamber (300mL volume). The MEMS tilt sensor to be calibrated is fixedly installed in each chamber. Unlike traditional low-pressure test chambers, this embodiment physically separates the vacuum pump group from the calibration chambers, controlling them via a solenoid valve group. First, the vacuum pump group synchronously and rapidly reduces the air pressure in all 12 chambers to 60kPa (simulating an altitude of approximately 4000 meters). Once the target air pressure is reached, the vacuum pump and all solenoid valves are immediately and completely shut off, cutting off all active air pressure control and vibration-generating equipment. This fundamentally eliminates the continuous vibration interference generated by the vacuum pump, compressor, and other motor equipment during the calibration data acquisition phase. In contrast, in traditional low-pressure test chambers, the vacuum pump runs continuously while maintaining air pressure, and vibration continuously couples into the sensor output signal, generating noise and altering the stress distribution of the chip and PCB, interfering with the accurate extraction of pure air pressure drift characteristics. Improvements to address the unnatural process: After all active equipment is stopped, each calibration chamber maintains limited gas exchange with the outside environment through an independent air-resistance microporous tube (0.3mm inner diameter, 40mm length), allowing the chamber pressure to slowly recover to ambient pressure under natural conditions, with a pressure change rate of approximately 15Pa / min. This natural recovery process completely simulates the natural pressure changes experienced by the sensor in actual outdoor applications (such as slow pressure changes caused by passing weather systems and altitude changes), rather than the constant low pressure maintained by active evacuation / inflation in traditional test chambers or the artificially controlled pressure change path. The rate and path of naturally recovered pressure change are highly consistent with the real outdoor environment, ensuring that the calibrated pressure coefficient has excellent applicability in actual environments. S2. Dynamic Data Synchronous Acquisition: Throughout the natural pressure recovery process (approximately 4.6 hours), each station synchronously and continuously acquires the raw output data Y (in arcseconds) of the MEMS tilt sensor and the ambient air pressure data P within the cavity at a frequency of 5 times per second. Since the vibration source has been completely eliminated in step S1, the data acquired in this stage is pure air pressure drift characteristic data, unaffected by vibration noise from equipment such as vacuum pumps. S3. Data Preprocessing and Feature Construction: The collected raw data is processed as follows to ensure the quality of the modeling data; S31. Noise Reduction and Alignment: Time alignment is performed on the Y and P of each station, and a moving average filter (window width 10 points) is used to remove residual high-frequency noise. S32. Anomaly Removal: Based on the 3σ criterion, data anomalies caused by accidental electromagnetic interference are automatically identified and removed to avoid accidental interference affecting model accuracy. S33, Feature Structure: Calculating the Rate of Pressure Change (Unit: kPa / hour), using the output of each sensor at the initial time (air pressure 60 kPa) as the baseline Y0, calculate the sensor output change ΔY = Y - Construct 12 high-quality datasets ; S4. Decoupling Physical-Mathematical Model Construction and Fitting: Based on the physical laws of air pressure and drift, a correction model integrating linear and nonlinear relationships is established.

[0026] in, The change in sensor output For ambient air pressure, The rate of change of air pressure, , , , Here are the pressure drift correction coefficients to be fitted. For model residuals; A nonlinear regression algorithm is used to evaluate the model parameters. , , , Perform fitting; A nonlinear least squares method was used to quickly fit the sensors at each workstation, with a single sensor fitting time of less than 0.3 seconds. Steps S1-S3 were then repeated at two temperature points, 15℃ and 35℃, to construct a temperature-dependent pressure coefficient model.

[0027] in, , , , The coefficient that varies with temperature is fitted. A linear relationship is established to achieve temperature compensation; S5. Parameter Validation and Storage: The pressure drift correction coefficients (a1~a4, b1~b4) obtained from the fitting are written into the built-in Flash non-volatile memory of each sensor. Simultaneously, the variance σ of the residual distribution of the fitted model is stored... 2 Stored together as prior knowledge for use in subsequent online correction algorithms (such as Kalman filtering) to set process noise; Quality is graded based on residual standard deviation: σ < 4 arcseconds is Grade A, 4 ≤ σ < 10 arcseconds is Grade B, and σ ≥ 10 arcseconds is unqualified. 8% of each batch is randomly sampled for independent pressure change verification to ensure calibration consistency. S6. Long-term correction interface: Each sensor has a reserved long-term correction interface via RS485 communication, allowing the host computer to inject new air pressure correction coefficients. After the sensors are deployed in the field, the factory calibration parameters can be dynamically corrected based on online monitoring data, adapting to sensor aging and environmental changes, achieving lifelong learning capability.

[0028] The following are the test results of Example 3 above:

[0029] In summary, the proposed method for pre-calibrating MEMS sensor pressure drift in outdoor cold environments overcomes the fundamental shortcomings of existing technologies through a three-pronged technological innovation: eliminating vibration interference, simulating natural processes, and improving efficiency while reducing costs. This method combines theoretical rigor with engineering practicality, providing reliable technical support for the widespread application of MEMS accelerometers, tilt sensors, and other sensors in outdoor static health monitoring of bridges, slopes, and buildings. It possesses significant application value and industrialization prospects.

[0030] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A method for pre-calibrating the barometric pressure drift of MEMS sensors based on outdoor cold environments, characterized in that: Includes the following steps: S1. Low-pressure environment construction and natural recovery: Place the MEMS sensor to be calibrated in a sealed calibration chamber, first reduce the air pressure in the chamber to the target low value, then stop all active air pressure control and vibration generation equipment, and allow the air pressure in the chamber to slowly recover to the ambient air pressure under natural conditions. S2. Dynamic Data Synchronous Acquisition: During the natural recovery process of air pressure, the raw output data of the MEMS sensor is acquired synchronously and continuously. and the ambient air pressure data inside the cavity ; S3. Data Preprocessing and Feature Construction: Preprocessing the collected raw data... and Denoising and alignment processes were performed, and the rate of change of air pressure was calculated. Construct a dataset for modeling, including air pressure. Rate of change of air pressure and the corresponding sensor output change ,in , This refers to the sensor's output under known reference conditions. S4. Decoupling Physical-Mathematical Model Construction and Fitting: Based on the physical laws of air pressure and drift, a correction model integrating linear and nonlinear relationships is established. in, The change in sensor output For ambient air pressure, The rate of change of air pressure, , , , Here are the pressure drift correction coefficients to be fitted. For model residuals; A nonlinear regression algorithm is used to evaluate the model parameters. , , , Perform fitting; S5. Parameter Verification and Storage: The fitted pressure drift correction coefficients are then stored. , , , ; S6. Long-term correction interface: Provides an updateable interface for the calibration parameters by dynamically correcting the factory calibration parameters based on online monitoring data.

2. The method for pre-calibrating the pressure drift of MEMS sensors based on outdoor cold environment according to claim 1, characterized in that: In step S1, after the active air pressure control device is stopped, the calibration chamber maintains limited gas exchange with the outside world so that the air pressure rises slowly, simulating the natural air pressure change process outdoors.

3. The method for pre-calibrating the pressure drift of MEMS sensors based on outdoor cold environment according to claim 2, characterized in that: The limited gas exchange is achieved through micropores or controllable valves.

4. The method for pre-calibrating the pressure drift of MEMS sensors based on outdoor cold environment according to claim 1, characterized in that: In step S3, the data preprocessing includes using a moving average filter to remove high-frequency noise and using the 3σ criterion to remove data anomalies caused by accidental interference.

5. The method for pre-calibrating the pressure drift of MEMS sensors based on outdoor cold environment according to claim 1, characterized in that: In step S4, when building the model, temperature T is also introduced as a compensation variable to construct a temperature-dependent pressure coefficient model: in, , , , This is a coefficient that varies with temperature; By repeating steps S1-S3 at multiple temperature points, a functional relationship between the coefficients and temperature is fitted.

6. The method for pre-calibrating the pressure drift of MEMS sensors based on outdoor cold environment according to claim 1, characterized in that: In step S5, the variance of the residual distribution of the fitted model is also stored as prior knowledge for use in setting the process noise of the subsequent online correction algorithm.

7. The method for pre-calibrating the pressure drift of MEMS sensors based on outdoor cold environment according to claim 1, characterized in that: In step S5, the non-volatile memory is Flash memory.

8. The method for pre-calibrating the pressure drift of MEMS sensors based on outdoor cold environment according to claim 1, characterized in that: In step S6, the updatable interface is implemented through a communication interface, allowing the host computer to inject new air pressure correction coefficients.

9. The method for pre-calibrating the pressure drift of MEMS sensors based on outdoor cold environment according to claim 8, characterized in that: The communication interface is an RS485 interface.

10. The method for pre-calibrating the pressure drift of MEMS sensors based on outdoor cold environment according to claim 1, characterized in that: The MEMS sensor is a MEMS accelerometer or tilt sensor.

Images