Mimu temperature gradient calibration and compensation method

Abstract

The application provides a micro-inertial measurement unit temperature gradient calibration and compensation method, which comprises the following steps: modeling and stress and strain measurement: modeling the MIMU as a cube, six surfaces of the cube are respectively provided with a temperature sensor, three MEMS gyroscopes are welded on the front surfaces of three PCBs A, B and C, and three MEMS accelerometers are welded on the back surfaces of the three PCBs A, B and C; the three PCBs A, B and C are orthogonal to each other, the accelerometers on the surfaces A, B and C measure linear velocities in X-axis, Y-axis and Z-axis directions, and the gyroscopes on the surfaces A, B and C measure angular velocities in Z-axis, X-axis and Y-axis directions; two orthogonally arranged strain sensors are arranged at the center positions of the surfaces A, B and C respectively; a strain sensor of the same type is hung in a temperature box; S2, modeling and calibration of the zero offset output of the MEMS inertial sensor based on the MIMU; S3, temperature and strain compensation of the zero offset of the MIMU; and the temperature error suppression effect of the MIMU is improved.

Description

Technical Field

[0001] This invention belongs to the field of inertial measurement and navigation technology, and in particular to a method for MIMU temperature gradient calibration and compensation. Background Technology

[0002] With the development of inertial navigation technology, Miniature Inertial Measurement Units (MIMUs) have broad prospects in both military and civilian applications due to their advantages such as small size, low cost, and mass production. An MIMU consists of a Micro-Electro-Mechanical Systems (MEMS) gyroscope, a MEMS accelerometer, and an embedded microcomputer, and can provide position, velocity, and other information to a carrier.

[0003] Among the various error sources affecting MIMU accuracy, the zero-bias error of MEMS gyroscopes and accelerometers plays a decisive role. Due to the materials and manufacturing processes of MEMS inertial devices, their zero-point values ​​are highly sensitive to temperature. Multiple heat sources are contained within the MIMU, and even when the external temperature is stable, a temperature gradient still exists within the MIMU. This temperature gradient becomes more pronounced during external temperature changes. The impact of the temperature gradient on MEMS inertial sensors manifests in two ways: firstly, the temperature directly affects the sensor; secondly, the temperature gradient acts on the fixture / PCB, causing PCB deformation and generating stress / strain that directly affects the sensor's output. Traditional temperature compensation methods address the first problem but struggle with the second. Summary of the Invention

[0004] The purpose of this patent is to propose a method for temperature gradient calibration and compensation of a Miniature Inertial Measurement Unit (MIMU). This method not only considers the influence of temperature on inertial devices, but also compensates for the zero-position drift of devices caused by the deformation of printed circuit boards (PCBs) due to temperature gradients, thereby improving the temperature error suppression effect of the MIMU.

[0005] The technical solution to achieve the purpose of this invention is as follows:

[0006] A method for MIMU temperature gradient calibration includes:

[0007] S1, MIMU modeling and stress-strain measurement:

[0008] The MIMU is modeled as a cube, with a temperature sensor on each of its six faces. Three MEMS gyroscopes are soldered to the front of three PCBs (A, B, and C), and three MEMS accelerometers are soldered to the back of the same three PCBs. The three PCBs (A, B, and C) are orthogonal to each other. The accelerometers on faces A, B, and C measure the linear velocity along the X, Y, and Z axes, while the gyroscopes on faces A, B, and C measure the angular velocity along the Z, X, and Y axes. Two orthogonally arranged strain sensors are placed at the center of each of the three faces (A, B, and C) to measure the deformation of the PCBs in two orthogonal directions. A strain sensor of the same model is suspended inside the temperature chamber.

[0009] S2. Zero-bias output modeling and calibration of MEMS inertial sensors based on MIMU:

[0010] (1) Establish a zero-bias temperature error model for MEMS accelerometers and MEMS gyroscopes under MIMU installation conditions;

[0011] (2) Place the MIMU in a temperature chamber and perform calibration tests according to the temperature change curve of constant temperature rate to obtain the zero-position output data of the accelerometer and gyroscope in the MIMU, the output value of the temperature sensor and the output value of the strain sensor; fit the obtained temperature and strain data to obtain the corresponding fitting coefficients.

[0012] The significant advantages of this invention compared to existing technologies are:

[0013] This solution utilizes a stress sensor to directly measure the PCB deformation of the FR-4 material used in the soldering of the inertial sensor, accurately obtaining the output error of the inertial sensor caused by strain, and designs a new and effective compensation method. This method simultaneously considers the effects of temperature and strain on the MEMS inertial sensor in the MIMU, thus more accurately compensating for errors and solving the problem of insufficient zero-bias temperature compensation accuracy caused by PCB deformation due to temperature influence. Furthermore, this method only requires a stress sensor, thus having low implementation costs and making it highly feasible in practical applications. Attached Figure Description

[0014] Figure 1 This is a schematic diagram of the PCB design of the MIMU provided in an embodiment of the present invention;

[0015] Figure 2 This is a schematic diagram illustrating the deformation of a MEMS inertial sensor soldered onto a PCB, as provided in an embodiment of the present invention.

[0016] Figure 3 This is a schematic diagram showing the bonding position of the strain gauge on the PCB according to an embodiment of the present invention;

[0017] Figure 4The constant-temperature variable-rate temperature change curve of the full-temperature measurement experiment provided in the embodiments of the present invention;

[0018] Figure 5 This is a schematic diagram of the overall mounting of the temperature sensor and strain gauge of the MIMU provided in an embodiment of the present invention. Detailed Implementation

[0019] The present invention will be further described below with reference to the accompanying drawings and specific embodiments.

[0020] This embodiment provides a method for MIMU temperature gradient calibration and compensation, including:

[0021] S1. MIMU Modeling and Stress-Strain Measurement Method: The MIMU structure is based on a typical meter header soldered onto a PCB. After soldering the MEMS inertial sensor to the MIMU, the MIMU is modeled as a cube, such as... Figure 5 As shown in the diagram, five PCBs (A through E) are connected by a flexible board. Three MEMS gyroscopes are soldered to the front of PCBs A, B, and C, and three MEMS accelerometers are soldered to the back of PCBs A, B, and C. PCBs D and E house power supply circuitry, data acquisition circuitry, etc. PCBs A, B, and C are orthogonal to each other. The accelerometer on PCB A measures the linear velocity along the X-axis, and the gyroscope on PCB A measures the angular velocity along the Z-axis. The accelerometer on PCB B measures the linear velocity along the Y-axis, and the gyroscope on PCB B measures the angular velocity along the X-axis. The accelerometer on PCB C measures the linear velocity along the Z-axis, and the gyroscope on PCB C measures the angular velocity along the Y-axis. This PCB design is as follows. Figure 1 As shown. A temperature sensor is installed on each of the surfaces A through E and the MIMU base, totaling six sensors, forming a multi-temperature sensor matrix. Two orthogonally arranged strain sensors are installed on each of surfaces A, B, and C, totaling six strain sensors. The strain sensors are arranged as follows... Figure 3 The sample is orthogonally mounted at the center of the PCB to measure the deformation of the PCB in two orthogonal directions. Simultaneously, a strain sensor of the same model, S7, is suspended inside the temperature chamber. Since this strain sensor is suspended and does not contact the object, it is only sensitive to temperature and not strain. Therefore, it can be used as a reference to eliminate the influence of temperature on the strain sensor, improving the accuracy of strain measurement.

[0022] In the MIMU, the sensor is soldered onto a PCB made of FR-4 material. After the PCB is installed in the MIMU, it will deform due to several factors, which will be transmitted to the MEMS sensor and cause additional errors. These factors include the difference in thermal expansion between the PCB and the MIMU frame, the difference in screw positions during the installation process, etc. These deformations are all related to temperature. Figure 2 A schematic diagram of PCB deformation and bending is provided. By directly measuring the deformation, this error can be effectively quantified and compensated for.

[0023] S2. Modeling and calibration of zero-bias output of MEMS inertial sensor based on MIMU: Establishing a zero-bias temperature error model for MEMS accelerometer and MEMS gyroscope under MIMU installation condition.

[0024]

[0025]

[0026] Where A(T) and G(T) are the fitted outputs of the MEMS accelerometer and MEMS gyroscope, and A0 and G0 are the zero-position output values ​​of the MEMS accelerometer and MEMS gyroscope at room temperature (25℃). i Let S be the i-th temperature fitting coefficient of the MEMS accelerometer, T be the temperature data of the inertial sensor, and S be the temperature fitting coefficient of the MEMS accelerometer. x S y S7 is the output of the strain sensor in both the x and y directions of the inertial sensor, and S8 is the output of the suspended strain sensor. i e i f is the i-th strain fitting coefficient of the MEMS accelerometer. i Let g be the i-th temperature fitting coefficient of the MEMS gyroscope. i h i is the i-th strain fitting coefficient of the MEMS gyroscope, where n of the parameters are the order of fitting, i = 1, 2, 3..., and i is the i-th fitting coefficient.

[0027] Place the MIMU in the incubator and follow the instructions as follows: Figure 4 The designed constant-temperature variable-rate temperature change curve was calibrated to obtain the zero-point output data of the accelerometer and gyroscope in the MIMU, the output value of the temperature sensor, and the output value of the strain sensor. The obtained temperature and strain data were fitted in Matlab to obtain the corresponding fitting coefficients, which were used in the subsequent compensation algorithm. Figure 4 The experiment shown illustrates a constant temperature rate change. The temperature change range and rate are determined based on the sensor's operating temperature conditions. For example, the temperature range is -40℃ to +85℃, and the temperature rate is generally selected from 0.5℃ / min to 5℃ / min.

[0028] S3, MIMU zero bias temperature and strain compensation: The MIMU data obtained in the temperature chamber will be used to compensate the inertial sensor using two parameters: operating temperature T and strain S, and two methods: operating temperature T and strain S fitted by temperature.

[0029] The first compensation method uses the temperature T and strain S obtained from calibration to compensate together, as shown in formulas 3 and 4.

[0030]

[0031]

[0032] Where a j A j (T)(j=x,y,z) represents the output acceleration of the accelerometer along the j-th axis, w j G j (T)(j=x,y,z) represents the output angular rate of the j-th axis of the gyroscope. lk In the formula (l=x,y,z,k=x,y), l represents the mounting position of the strain sensor on the MIMU. The x-axis corresponds to mounting on PCB surface A, the y-axis to mounting on PCB surface B, and the z-axis to mounting on PCB surface C. These three surfaces are orthogonal to each other in space. k represents the x and y measurement directions of the strain sensor within the PCB surface. A 0j (j=x,y,z) represents the zero-point output value of the accelerometer along the j-th axis at room temperature (25℃), G 0j (j=x,y,z) represents the zero-position output value of the j-th axis of the gyroscope at a normal temperature of 25℃. i Let be the i-th temperature fitting coefficient of the MEMS accelerometer, T1, T2, T3 be the temperature data of the inertial sensor on surfaces A, B, and C, and d be the temperature coefficient of the inertial sensor on surfaces A, B, and C. i e i f is the i-th strain fitting coefficient of the MEMS accelerometer. i Let g be the i-th temperature fitting coefficient of the MEMS gyroscope. i h i is the i-th strain fitting coefficient of the MEMS gyroscope, where n of the parameters are the order of fitting, i = 1, 2, 3..., and i is the i-th fitting coefficient.

[0033] Compensation method two uses temperature sensor data from specific locations to fit the strain value. This eliminates the need for individual strain sensors, achieving an approximate effect using only temperature sensors, thus simplifying the MIMU structure. Since strain primarily originates from the temperature gradient, multiple temperature sensors can be used to fit the strain pattern. Data from six temperature sensors, T1 to T6, is used to fit the strain S on the three surfaces A, B, and C where the inertial sensors are mounted. The strain on each surface is then transmitted through temperature sensors T1 and T6 on adjacent PCB locations. L and T R And the temperature T at that location M A total of three temperature data points are used as input, such as Figure 5 As shown, each face S x S y A total of six temperature data points were used to fit the temperature gradient to the strain S in Matlab, obtaining the corresponding coefficients o,p,q,o,,p,,q. The fitting formulas are shown in Equations 5 and 6.

[0034]

[0035]

[0036] Where S lk (T)(l=x,y,z,k=x,y) represents the temperature-strain fitting output, where l indicates the strain sensor's mounting position on the MIMU. The x-axis corresponds to mounting on PCB surface A, the y-axis to mounting on PCB surface B, and the z-axis to mounting on PCB surface C, with the three surfaces orthogonal to each other in space. k represents the x and y measurement directions of the strain sensor within the PCB surface. T1~T6 represent the temperature data for surfaces A~E and the bottom six surfaces, respectively. i ,p i ,q i For S lx The strain-temperature fitting coefficient for (T)(l=x,y,z), o i ,,p i ,,q i , for S ly The strain-temperature fitting coefficients of (T)(l=x,y,z), where n is the order of fitting for several parameters, i=1,2,3…, and i is the i-th fitting coefficient.

[0037] The fitted S lk (T)(l=x,y,z,k=x,y) Substitute the x, y, and z axes of the MIMU into formulas 1 and 2 respectively. x and S y Since temperature has already been used to characterize strain, S7 is not needed to eliminate the strain error caused by temperature factors affecting the six strain sensors. Therefore, the influence of S7 can be ignored in formulas 1 and 2. This yields the output compensation formulas for the MIMU installation under the dual influence of strain and temperature, characterized only by temperature, as shown in formulas 7 and 8.

[0038]

[0039]

Claims

1. A method for calibrating the temperature gradient of a MIMU, characterized in that, include: S1, MIMU modeling and stress-strain measurement: The MIMU is modeled as a cube, with a temperature sensor on each of the six faces. Three MEMS gyroscopes are soldered to the front of three PCBs A, B, and C, and three MEMS accelerometers are soldered to the back of three PCBs A, B, and C. The three PCBs A, B, and C are orthogonal to each other. Accelerometers on surfaces A, B, and C measure the linear velocities along the X, Y, and Z axes, while gyroscopes on surfaces A, B, and C measure the angular velocities along the Z, X, and Y axes. Two orthogonally arranged strain sensors are placed at the center of each of the three surfaces A, B, and C to measure the deformation of the PCBs in two orthogonal directions. A strain sensor of the same model is suspended inside the temperature chamber. S2. Zero-bias output modeling and calibration of MEMS inertial sensors based on MIMU: (1) Establish a zero-bias temperature error model for MEMS accelerometers and MEMS gyroscopes under MIMU installation conditions; (2) Place the MIMU in a temperature chamber and perform calibration tests according to the temperature change curve of constant temperature rate to obtain the zero-point output data of the accelerometer and gyroscope in the MIMU, the output value of the temperature sensor and the output value of the strain sensor. The obtained temperature and strain data are fitted to obtain the corresponding fitting coefficients.

2. The MIMU temperature gradient calibration method according to claim 1, characterized in that, The zero-bias temperature error model for MEMS accelerometers and MEMS gyroscopes under MIMU installation conditions is as follows: Where A(T) and G(T) are the fitted outputs of the MEMS accelerometer and MEMS gyroscope, and A0 and G0 are the zero-position output values ​​of the MEMS accelerometer and MEMS gyroscope at room temperature (25℃). i Let S be the i-th temperature fitting coefficient of the MEMS accelerometer, T be the temperature data of the inertial sensor, and S be the temperature fitting coefficient of the MEMS accelerometer. x S y S7 is the output of the strain sensor in both the x and y directions of the inertial sensor, and S8 is the output of the suspended strain sensor. i e i f is the i-th strain fitting coefficient of the MEMS accelerometer. i Let g be the i-th temperature fitting coefficient of the MEMS gyroscope. i h i is the i-th strain fitting coefficient of the MEMS gyroscope, and n is the order of fitting.

3. The MIMU temperature gradient calibration method according to claim 1, characterized in that, Its applications in MIMU zero-bias temperature and strain compensation include: Compensation is achieved by combining the calibrated temperature and strain: Where a j This represents the output acceleration along the j-th axis of the accelerometer, where j = x, y, z; w j S represents the output angular rate of the gyroscope along the j-th axis; lk This represents the strain measured in the MIMU along the l-axis in the k-th direction on the PCB, where l = x, y, z, k = x, y; A 0j G represents the zero-point output value of the accelerometer along the j-th axis at a normal temperature of 25°C. 0j The zero-point output value of the j-th axis of the gyroscope at room temperature (25°C); c i Let be the i-th temperature fitting coefficient of the MEMS accelerometer, and let T1, T2, and T3 be the temperature data of the inertial sensor on surfaces A, B, and C, respectively. i e i f is the i-th strain fitting coefficient of the MEMS accelerometer. i Let g be the i-th temperature fitting coefficient of the MEMS gyroscope. i h i S1 represents the i-th strain fitting coefficient of the MEMS gyroscope, n represents the order of fitting, and S7 represents the output of the suspended strain sensor.

4. The MIMU temperature gradient calibration method according to claim 1, characterized in that, Its applications in MIMU zero-bias temperature and strain compensation include: The inertial sensor is compensated using temperature and strain fitted with temperature: Where a j This represents the output acceleration along the j-th axis of the accelerometer, where j = x, y, z; w j A represents the output angular rate of the j-th axis of the gyroscope; 0j G represents the zero-point output value of the accelerometer along the j-th axis at a normal temperature of 25°C. 0j The zero-point output value of the j-th axis of the gyroscope at room temperature (25°C); c i f is the i-th temperature fitting coefficient of the accelerometer. i Let o be the i-th temperature fitting coefficient of the gyroscope. i ,p i ,q i ,o i ',p i ',q i ' represents the strain-temperature fitting coefficient, i = 1, 2, 3..., n is the fitting order; T1 to T6 are the temperature data of the inertial sensor at A to E and the six bottom surfaces, respectively.

Images