A method for calibrating a position sensor for a brake motor

Abstract

The application discloses a kind of brake motor position sensor calibration methods, belong to sensor calibration field, comprising the following steps: S1, calibration system is built and initialization;S2, original data acquisition under multiple conditions;S3, data preprocessing and error preliminary analysis, obtain temperature error distribution data under the temperature of full-condition;S4, compensation standard data set;S5, the optimal calibration parameter of main position sensor is exported to make fitness function value minimum;S6, fitting the calibration parameter correction amount of backup position sensor relative to main position sensor, and the correction amount is superimposed to optimal calibration parameter, obtain backup position sensor adaptation calibration parameter.The above-mentioned brake motor position sensor calibration method is used, the core pain point of traditional calibration low efficiency, wide temperature precision degradation, poor consistency of multiple sensors is solved, and an integrated automatic calibration scheme of adaptation mass production, low cost and high precision is provided.

Description

Technical Field

[0001] This invention relates to the field of sensor calibration technology, and in particular to a method for calibrating a position sensor for a brake motor. Background Technology

[0002] As a core component of automotive brake-by-wire systems, the Onebox braking system integrates components such as a motor, reducer, and master cylinder to achieve precise and rapid control of braking pressure. Its control accuracy directly depends on the real-time feedback of the brake motor rotor position. The motor position sensor (such as a rotary transformer or Hall effect sensor), as the core feedback element, needs to output four signals—COS+, COS-, SIN+, and SIN-—that precisely correspond to the rotor's electrical angle. This ensures that the Onebox system can adjust the motor torque based on the position signals under conditions such as emergency braking, gentle braking, and cold start, thereby controlling the braking pressure deviation to ≤5%. However, current calibration technology for the motor position sensor in Onebox braking systems still suffers from the following key technical deficiencies, severely restricting its performance and mass production applications: 1. Existing calibration methods are inefficient and have poor adaptability: Current motor position sensor calibration is mainly divided into two categories: manual calibration and automated equipment calibration. Manual calibration relies on engineers observing the sensor output signal through an oscilloscope and manually adjusting parameters such as offset and gain. A single calibration takes more than 30 minutes and is affected by human operating experience. The calibration error of the same batch of sensors can fluctuate by 0.01 rad, which cannot meet the requirements of Onebox system for mass production (cycle time ≤ 2 minutes / unit) and accuracy (position error ≤ 0.5°). Automated equipment calibration, although it achieves automatic parameter optimization through industrial robots and large calibration stands, has a large equipment size (length ≥ 1.5m) and weight ≥ 50kg. Onebox system is mostly used in compact passenger cars, and its brake module installation space is only 300mm × 200mm × 150mm, which cannot be adapted to large calibration equipment. Moreover, the equipment cost exceeds 500,000 yuan, which is difficult for small and medium-sized enterprises to afford, thus limiting the popularization of the technology.

[0003] 2. Temperature drift leads to deterioration of calibration accuracy under wide temperature conditions: The Onebox braking system operates in a temperature range of -40℃ (winter in cold regions) to 125℃ (heat radiation from the engine compartment). The core components of the position sensor (such as the windings of the rotary transformer and Hall elements) are prone to parameter drift due to temperature variations. Existing technologies often employ quadratic polynomial temperature compensation to address this issue. However, this method can only fit linear temperature errors within a small range (e.g., -20~80℃). When the temperature exceeds this range, the compensation residual increases significantly. Actual measurements show that the compensation error reaches 0.02 rad at -40℃ and 0.018 rad at 125℃, both exceeding the Onebox system's allowable error limit of 0.5°. This results in a braking pressure response delay of ≥100ms during low-temperature starts and a braking pressure overshoot of ≥10% under high-temperature conditions, posing a safety hazard.

[0004] 3. Lack of Consistency Calibration in Multi-Sensor Redundancy Scenarios: To improve reliability, mid-to-high-end Onebox systems often adopt a "primary + backup" dual-position sensor redundancy design (such as a primary rotary transformer + backup Hall sensor), requiring the position signal deviation between sensors to be ≤0.1°, ensuring that braking pressure fluctuation is ≤3% during redundancy switching. However, existing technologies use an "independent calibration + post-comparison" mode for multi-sensor calibration, that is, after separately calibrating the primary and backup sensors, the signal deviation is calculated and corrected by software. This mode lacks a unified spatial reference and is easily affected by installation errors (such as the coaxiality deviation between the sensor and the motor shaft), resulting in a consistency error between sensors reaching 0.008 rad, exceeding the redundancy switching requirements; moreover, the correction process must be completed offline, making real-time calibration on the mass production line impossible, increasing subsequent maintenance costs. Summary of the Invention

[0005] The purpose of this invention is to provide a calibration method for a position sensor used in a brake motor, thereby solving the aforementioned technical problems.

[0006] To achieve the above objectives, the present invention provides a calibration method for a position sensor for a brake motor, comprising the following steps: S1. Set up and initialize the calibration system: Fix the reference calibration component on the output shaft of the motor to be calibrated to provide a physical position reference for the main position sensor and the backup position sensor on the output shaft of the motor to be calibrated, and arrange the working condition simulation component on the outside of the motor to be calibrated to simulate the full working condition temperature. S2. Raw data acquisition under multiple operating conditions: The operating condition simulation component is used to simulate the temperature under all operating conditions, and the COS+, COS-, SIN+, SIN- voltage signals, ambient temperature, motor speed and timestamps output by the main position sensor and the backup position sensor are collected according to the set temperature gradient and speed gradient to obtain the raw dataset. S3. Data preprocessing and preliminary error analysis: The original dataset is subjected to outlier removal, noise filtering, signal consistency verification and normalization in sequence to obtain a standard dataset. Based on the standard dataset, the actual electrical angle and original position error of the main position sensor are calculated to obtain the temperature error distribution data under all operating conditions. S4. Construct a temperature error prediction model with a bidirectional LSTM layer, and use the standard dataset as input and the temperature error distribution data as labels to train the temperature error prediction model with the bidirectional LSTM layer. Then, use the trained temperature error prediction model with the bidirectional LSTM layer combined with non-uniform B-spline interpolation to compensate the standard dataset. S5. Sensor calibration parameter optimization: Based on the standard dataset after temperature compensation, the offset and gain of the main position sensor are defined as calibration parameters to be optimized. A fitness function with the root mean square error of the calibrated position as the objective is constructed. After iterative optimization for a preset number of iterations, the optimal calibration parameters of the main position sensor that minimize the fitness function value are output. S6. Multi-sensor consistency calibration: Using the optimal calibration parameters of the main position sensor as a benchmark, the target features of the main position sensor and the backup position sensor are extracted through the benchmark calibration component. A linear regression algorithm is used to fit the calibration parameter correction amount of the backup position sensor relative to the main position sensor, and the correction amount is superimposed on the optimal calibration parameters to obtain the adaptation calibration parameters of the backup position sensor, thereby achieving the unification of the calibration parameters of the main position sensor and the backup position sensor.

[0007] Therefore, the above-mentioned method for calibrating a position sensor for a brake motor has the following advantages: 1. Significantly improve calibration efficiency and adapt to mass production cycle: Through full-process automation (data acquisition → compensation → optimization → calibration), the calibration time is reduced to ≤2 minutes / unit, which fully matches the production cycle requirements of the Onebox braking system mass production line and significantly reduces production time; 2. Optimize wide-temperature calibration accuracy and avoid potential safety hazards in operating conditions: Within the full operating temperature range of -40~125℃, the temperature error time sequence is captured by bidirectional LSTM and the residual is corrected by non-uniform B-spline, so that the mean error after temperature compensation is ≤0.01rad. Compared with traditional polynomial compensation (error 0.018~0.02rad), the accuracy is improved by ≥40%, which effectively avoids safety risks such as braking pressure response delay (≥100ms) during low-temperature start-up and braking pressure overshoot (≥10%) under high-temperature operating conditions. 3. Ensure consistency between primary and backup sensors and meet redundancy reliability requirements: Using a 3D composite target as a unified physical benchmark, and combining linear regression to fit the correction amount of the backup sensor, the consistency error between primary and backup sensors is reduced from ≥0.008rad in traditional independent calibration to ≤0.002rad, ensuring that the braking pressure fluctuation is ≤3% during redundancy switching, and improving the operational reliability of the Onebox system. 4. Reduce equipment costs and space adaptation difficulty: Use a small 3D composite target with a diameter of 50mm to replace the traditional large calibration equipment (length ≥1.5m, weight ≥50kg), reducing the equipment cost from ≥500,000 yuan to ≤50,000 yuan, and adapting to the compact installation space of the Onebox system of 300mm×200mm×150mm, reducing the technical application threshold for small and medium-sized enterprises. 5. Improve calibration stability and data traceability: Automated data acquisition is stored in CSV format (including timestamp, temperature, speed, and sensor signals). Combined with error heatmap analysis, it ensures that the calibration results are stable and controllable, and the data can be traced throughout the process, which facilitates later maintenance and quality control.

[0008] This invention employs a calibration method for position sensors used in brake motors, achieving automated, wide-temperature adaptability, and consistent primary and backup calibration. Calibration time is ≤2 minutes per unit (mass production compatible), the final position error of the primary sensor is ≤0.005 rad (≈0.286°), and the consistency error between primary and backup sensors is ≤0.002 rad (≈0.114°). It fully meets the mass production requirements (high efficiency, low cost) and accuracy requirements (wide-temperature stability, primary and backup consistency) of Onebox braking systems. It can be directly applied to the mass production calibration of sensors in passenger vehicle brake-by-wire systems, providing core support for precise brake pressure control (deviation ≤5%). Simultaneously, it lowers the technical threshold for small and medium-sized enterprises (reducing equipment costs by 90%), possessing broad industrialization prospects.

[0009] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0010] Figure 1 This is a flowchart of a method for calibrating a position sensor for a brake motor according to the present invention. Detailed Implementation

[0011] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the embodiments of the present invention will be further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are merely illustrative of the embodiments of the present invention and are not intended to limit the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of this application. Examples of the 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.

[0012] It should be noted that the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion, such as a process, method, system, product, or server that includes a series of steps or units, not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such process, method, product, or device.

[0013] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0014] like Figure 1 As shown, a method for calibrating a position sensor for a brake motor includes the following steps: S1. Set up and initialize the calibration system: Fix the reference calibration component on the output shaft of the motor to be calibrated to provide a physical position reference for the main position sensor and the backup position sensor on the output shaft of the motor to be calibrated, and arrange the working condition simulation component on the outside of the motor to be calibrated to simulate the full working condition temperature. In step S1, the main position sensor and the backup position sensor are mounted on the same circumference of the motor housing at a 180° angle around the axis of the output shaft of the motor to be calibrated. The reference calibration component is a 3D composite target coaxially fixed to the end of the output shaft of the motor to be calibrated. The 3D composite target is a 6061 aluminum alloy cylindrical target with a diameter of 50mm. Target features are provided on the outer circumference of the 3D composite target. The distance between the target features and the main position sensor and the distance between the target features and the backup position sensor are equal, so as to provide a physical position reference using the target features. The target features include two concentric circles and a cross-shaped planar groove laser-engraved on the outer circumference of the 3D composite target. The intersection of the cross-shaped planar groove coincides with the center of the concentric circles. One of the cross lines of the cross-shaped planar groove is the 0° reference line, and the other is the 90° reference line. The 0° reference line is along the line connecting the main position sensor and the backup position sensor, and the 90° reference line is perpendicular to the 0° reference line. The two concentric circles, the 0° reference line, and the 90° reference line all fall within the detection field of view of the main position sensor and the backup position sensor. The operating condition simulation component includes heating plates and temperature sensors fixed to both sides of the motor housing of the motor to be calibrated, used to achieve full operating condition temperature simulation from -40℃ to 125℃.

[0015] S2. Raw data acquisition under multiple operating conditions: The operating condition simulation component is used to simulate the temperature under all operating conditions, and the COS+, COS-, SIN+, SIN- voltage signals, ambient temperature, motor speed and timestamps output by the main position sensor and the backup position sensor are collected according to the set temperature gradient and speed gradient to obtain the raw dataset. In step S2, the temperature gradient is set to -40℃, -25℃, 85℃ and 125℃; the speed gradient is set to 0rpm, 1000rpm, 2000rpm, 3000rpm, 4000rpm, 5000rpm and 6000rpm. Any temperature gradient and any speed gradient are combined to obtain the operating conditions of the motor to be calibrated.

[0016] S3. Data preprocessing and preliminary error analysis: The original dataset is subjected to outlier removal, noise filtering, signal consistency verification and normalization in sequence to obtain a standard dataset. Based on the standard dataset, the actual electrical angle and original position error of the main position sensor are calculated to obtain the temperature error distribution data under all operating conditions. In step S3, the actual electrical angle of the main position sensor The calculation formula is as follows: ; in, ; ; In the formula, and These represent the SIN+ voltage signals output by the main position sensor. With SIN voltage signal amplitude difference and COS+ voltage signal With COS-voltage signal The amplitude difference; Indicates the reference voltage of the main position sensor; raw position error of the main position sensor The calculation formula is as follows: ; in, ; In the formula, This represents the theoretical electrical angle of the main position sensor; Indicates the number of pole pairs of the motor to be customized; This indicates the actual speed of the motor to be customized; Indicates the data collection time.

[0017] In step S3, the temperature gradient values ​​are divided into four major temperature categories: -40℃, -25℃, 85℃, and 125℃. Within each major temperature category, the rotational speed is further divided into seven subcategories based on the rotational speed gradient: 0 rpm, 1000 rpm, 2000 rpm, 3000 rpm, 4000 rpm, 5000 rpm, and 6000 rpm. For each subcategory, the mean value of the original position error of the main position sensor is calculated. and standard deviation : ; ; In the formula, This represents the number of original position error samples of the main sensor under the rotational speed subclass within a certain temperature category; Based on the mean of the original position error for each rotational speed subclass and standard deviation A heatmap was plotted using Python's matplotlib library, where the x-axis represents the temperature gradient and the y-axis represents the rotational speed gradient. The peak error conditions were identified, and the temperature error distribution data was obtained.

[0018] S4. Construct a temperature error prediction model with a bidirectional LSTM layer, and use the standard dataset as input and the temperature error distribution data as labels to train the temperature error prediction model with the bidirectional LSTM layer. Then, use the trained temperature error prediction model with the bidirectional LSTM layer combined with non-uniform B-spline interpolation to compensate the standard dataset. In step S4, the loss function of the temperature error prediction model for the bidirectional LSTM layer is as follows: ; In the formula, This represents the mean squared error loss during model training. Indicates the first position sensor Predicted temperature error value for each sample; Indicates the first position sensor True label of temperature error for each sample; The temperature error prediction model using the trained bidirectional LSTM layers, combined with non-uniform B-spline interpolation to compensate for the standard dataset, is expressed as follows: ; in, ; ; ; In the formula, This represents the actual electrical angle of the master position sensor after temperature compensation. This represents the temperature error predicted by the temperature error prediction model of the bidirectional LSTM layer; Represents the interpolation function; Represents the B-spline interpolation function; Represents B-spline nodes; Represents the B-spline coefficients; Let represent the order of the B-spline, and ; Indicates ambient temperature; This represents the function that generates univariate spline nodes; This represents the temperature error residual of the temperature error prediction model for the bidirectional LSTM layer. This indicates the order of the B-spline; This represents the smoothing factor for spline fitting; This label indicates the actual temperature error.

[0019] S5. Sensor calibration parameter optimization: Based on the standard dataset after temperature compensation, the offset and gain of the main position sensor are defined as calibration parameters to be optimized. A fitness function with the root mean square error of the calibrated position as the objective is constructed. After iterative optimization for a preset number of iterations, the optimal calibration parameters of the main position sensor that minimize the fitness function value are output. The fitness function expression described in step S5 is as follows: ; in, ; In the formula, Indicates the fitness value. This indicates the calibration parameters to be optimized, and , These represent the offset and gain of the main position sensor, respectively. Indicates the first The electrical angle after calibration of each sample; Indicates the first The theoretical electrical angle after calibration of a single sample; This indicates the number of samples in the dataset after temperature compensation.

[0020] S6. Multi-sensor consistency calibration: Using the optimal calibration parameters of the main position sensor as a benchmark, the target features of the main position sensor and the backup position sensor are extracted through the benchmark calibration component. A linear regression algorithm is used to fit the calibration parameter correction amount of the backup position sensor relative to the main position sensor, and the correction amount is superimposed on the optimal calibration parameters to obtain the adaptation calibration parameters of the backup position sensor, thereby achieving the unification of the calibration parameters of the main position sensor and the backup position sensor.

[0021] Step S6 specifically includes the following steps: S61. Extract target features: Extract the center coordinates of the target features corresponding to the main position sensor and the backup position sensor respectively, and verify coaxiality: ; In the formula, and These represent the target feature center coordinates corresponding to the main position sensor and the target feature center coordinates corresponding to the backup position sensor, respectively. This represents the spatial distance between the target feature center corresponding to the primary position sensor and the target feature center corresponding to the backup position sensor. ; S62. Based on the target feature center coordinates, calculate the actual electrical angles corresponding to the target feature center coordinates of the main position sensor and the backup position sensor: ; ; In the formula, and These represent the actual electrical angles of the target feature center coordinates corresponding to the main position sensor and the backup position sensor, respectively. S63. Calculate the initial consistency error : ; in, ; ; In the formula, and These represent the electrical angles of the primary position sensor and the backup position sensor after optimal parameter calibration, respectively. and These represent the electrical angles of the primary position sensor and the backup position sensor after compensation in step S4, respectively. and These represent the optimal gain and offset of the main position sensor obtained through iterative optimization in step S5, respectively. S64. Introduce the calibration parameters of the backup position sensor to obtain the corrected angle after calibration of the backup position sensor. : ; In the formula, and These represent the gain correction and offset correction amounts for the backup position sensor, respectively. and with Minimize as the objective and solve using the least squares method and : ; In the formula, This indicates the number of samples in the multi-sensor data.

[0022] Following step S6, step S7, calibration result solidification and full-condition verification, is also included: the optimal calibration parameters of the main position sensor and the adaptation calibration parameters of the backup position sensor are written into the memory of the corresponding sensor, and the accuracy of parameter writing is verified. The full-condition temperature of the brake motor is reproduced using the operating condition simulation component, the output signal of the calibrated sensor is collected, and the final position error and multi-sensor consistency error under full-condition temperature are verified. If the error meets the requirements of the braking system, a standardized calibration report is generated to complete the calibration. If it does not meet the requirements, the process returns to S5 to re-optimize the parameters.

[0023] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for calibrating a position sensor for a brake motor, characterized in that: Includes the following steps: S1. Set up and initialize the calibration system: Fix the reference calibration component on the output shaft of the motor to be calibrated to provide a physical position reference for the main position sensor and the backup position sensor on the output shaft of the motor to be calibrated, and arrange the working condition simulation component on the outside of the motor to be calibrated to simulate the full working condition temperature. The reference calibration component is a 3D composite target coaxially fixed to the end of the output shaft of the motor to be calibrated. The 3D composite target is a cylindrical target with target features on its outer circumference. The distance between the target features and the main position sensor and the distance between the target features and the backup position sensor are equal. The target features include two concentric circles and a cross-shaped planar groove laser-engraved on the outer circumference of the 3D composite target, with the intersection of the cross-shaped planar groove coinciding with the center of the concentric circles; The operating condition simulation component includes heating plates and temperature sensors fixed to both sides of the motor housing of the motor to be calibrated, used to achieve full operating condition temperature simulation from -40℃ to 125℃; S2. Raw data acquisition under multiple operating conditions: The operating condition simulation component is used to simulate the temperature under all operating conditions, and the COS+, COS-, SIN+, SIN- voltage signals, ambient temperature, motor speed and timestamps output by the main position sensor and the backup position sensor are collected according to the set temperature gradient and speed gradient to obtain the raw dataset. S3. Data preprocessing and preliminary error analysis: The original dataset is subjected to outlier removal, noise filtering, signal consistency verification and normalization in sequence to obtain a standard dataset. Based on the standard dataset, the actual electrical angle and original position error of the main position sensor are calculated to obtain the temperature error distribution data under all operating conditions. S4. Construct a temperature error prediction model with a bidirectional LSTM layer, and use the standard dataset as input and the temperature error distribution data as labels to train the temperature error prediction model with the bidirectional LSTM layer. Then, use the trained temperature error prediction model with the bidirectional LSTM layer combined with non-uniform B-spline interpolation to compensate the standard dataset. S5. Sensor calibration parameter optimization: Based on the standard dataset after temperature compensation, the offset and gain of the main position sensor are defined as calibration parameters to be optimized. A fitness function with the root mean square error of the calibrated position as the objective is constructed. After iterative optimization for a preset number of iterations, the optimal calibration parameters of the main position sensor that minimize the fitness function value are output. S6. Multi-sensor consistency calibration: Using the optimal calibration parameters of the main position sensor as a benchmark, the target features of the main position sensor and the backup position sensor are extracted through the benchmark calibration component. A linear regression algorithm is used to fit the calibration parameter correction amount of the backup position sensor relative to the main position sensor, and the correction amount is superimposed on the optimal calibration parameters to obtain the adaptation calibration parameters of the backup position sensor, thereby achieving the unification of the calibration parameters of the main position sensor and the backup position sensor.

2. The method for calibrating a position sensor for a brake motor according to claim 1, characterized in that: In step S1, the main position sensor and the backup position sensor are mounted on the same circumference of the motor housing at a 180° angle around the axis of the output shaft of the motor to be calibrated. The 3D composite target uses a 50mm diameter 6061 aluminum alloy cylindrical target, which provides a physical position reference by utilizing the target features; One of the cross lines of the cross-shaped planar groove is the 0° reference line, and the other is the 90° reference line. The 0° reference line is along the line connecting the main position sensor and the backup position sensor, and the 90° reference line is perpendicular to the 0° reference line. The two concentric circles, the 0° reference line, and the 90° reference line all fall within the detection field of view of the main position sensor and the backup position sensor.

3. The method for calibrating a position sensor for a brake motor according to claim 2, characterized in that: In step S2, the temperature gradient is set to -40℃, -25℃, 85℃ and 125℃; the speed gradient is set to 0rpm, 1000rpm, 2000rpm, 3000rpm, 4000rpm, 5000rpm and 6000rpm. Any temperature gradient and any speed gradient are combined to obtain the operating conditions of the motor to be calibrated.

4. The method for calibrating a position sensor for a brake motor according to claim 3, characterized in that: In step S3, the actual electrical angle of the main position sensor The calculation formula is as follows: ； in, ； ； In the formula, This represents the SIN+ voltage signal output by the main position sensor. With SIN voltage signal The amplitude difference; This represents the COS+ voltage signal output by the main position sensor. With COS-voltage signal The amplitude difference; This represents the reference voltage of the main position sensor; raw position error of the main position sensor The calculation formula is as follows: ； in, ； In the formula, This represents the theoretical electrical angle of the main position sensor; Indicates the number of pole pairs of the motor to be customized; This indicates the actual speed of the motor to be customized; Indicates the data collection time.

5. The method for calibrating a position sensor for a brake motor according to claim 4, characterized in that: In step S3, the temperature gradient values ​​are divided into four major temperature categories: -40℃, -25℃, 85℃, and 125℃. Within each major temperature category, the rotational speed is further divided into seven subcategories based on the rotational speed gradient: 0 rpm, 1000 rpm, 2000 rpm, 3000 rpm, 4000 rpm, 5000 rpm, and 6000 rpm. For each subcategory, the mean value of the original position error of the main position sensor is calculated. and standard deviation : ； ； In the formula, This represents the number of original position error samples of the main sensor under the rotational speed subclass within a certain temperature category; Based on the mean of the original position error for each rotational speed subclass and standard deviation A heatmap was plotted using Python's matplotlib library, where the x-axis represents the temperature gradient and the y-axis represents the rotational speed gradient. The peak error conditions were identified, and the temperature error distribution data was obtained.

6. The method for calibrating a position sensor for a brake motor according to claim 5, characterized in that: In step S4, the loss function of the temperature error prediction model for the bidirectional LSTM layer is as follows: ； In the formula, This represents the mean squared error loss during model training. Indicates the first position sensor Predicted temperature error value for each sample; Indicates the first position sensor True label of temperature error for each sample; The temperature error prediction model using the trained bidirectional LSTM layers, combined with non-uniform B-spline interpolation to compensate for the standard dataset, is expressed as follows: ； in, ； ； ； In the formula, This represents the actual electrical angle of the master position sensor after temperature compensation. This represents the temperature error predicted by the temperature error prediction model of the bidirectional LSTM layer; Represents the interpolation function; Represents the B-spline interpolation function; Represents B-spline nodes; Represents the B-spline coefficients; Let represent the order of the B-spline, and ; Indicates ambient temperature; This represents the function that generates univariate spline nodes; This represents the temperature error residual of the temperature error prediction model for the bidirectional LSTM layer. This indicates the order of the B-spline; This represents the smoothing factor for spline fitting; This label indicates the actual temperature error.

7. A method for calibrating a position sensor for a brake motor according to claim 6, characterized in that: The fitness function expression described in step S5 is as follows: ； in, ； In the formula, Indicates the fitness value. This indicates the calibration parameters to be optimized, and , These represent the offset and gain of the main position sensor, respectively. Indicates the first The electrical angle after calibration of each sample; Indicates the first The theoretical electrical angle after calibration of each sample; This indicates the number of samples in the dataset after temperature compensation.

8. A method for calibrating a position sensor for a brake motor according to claim 7, characterized in that: Step S6 specifically includes the following steps: S61. Extract target features: Extract the center coordinates of the target features corresponding to the main position sensor and the backup position sensor respectively, and verify coaxiality: ； In the formula, and These represent the target feature center coordinates corresponding to the main position sensor and the target feature center coordinates corresponding to the backup position sensor, respectively. This represents the spatial distance between the target feature center corresponding to the primary position sensor and the target feature center corresponding to the backup position sensor. ; S62. Based on the target feature center coordinates, calculate the actual electrical angles corresponding to the target feature center coordinates of the main position sensor and the backup position sensor: ； ； In the formula, and These represent the actual electrical angles of the target feature center coordinates corresponding to the main position sensor and the backup position sensor, respectively. S63. Calculate the initial consistency error : ； in, ； ； In the formula, and These represent the electrical angles of the primary position sensor and the backup position sensor after optimal parameter calibration, respectively. This indicates the electrical angle of the standby position sensor after temperature compensation. and These represent the optimal gain and offset of the main position sensor obtained through iterative optimization in step S5, respectively. S64. Introduce the calibration parameters of the backup position sensor to obtain the corrected angle after calibration of the backup position sensor. : ； In the formula, and These represent the gain correction and offset correction amounts for the backup position sensor, respectively. and with Minimize as the objective and solve using the least squares method and : ； In the formula, This indicates the number of samples in the multi-sensor data.

9. A method for calibrating a position sensor for a brake motor according to claim 8, characterized in that: Following step S6, step S7, calibration result solidification and full-condition verification, is also included: the optimal calibration parameters of the main position sensor and the adaptation calibration parameters of the backup position sensor are written into the memory of the corresponding sensor, and the accuracy of parameter writing is verified. The full-condition temperature of the brake motor is reproduced using the operating condition simulation component, the output signal of the calibrated sensor is collected, and the final position error and multi-sensor consistency error under full-condition temperature are verified. If the error meets the requirements of the braking system, a standardized calibration report is generated to complete the calibration. If it does not meet the requirements, the process returns to S5 to re-optimize the parameters.

Images