CN-121140872-B - Actuator stroke calibration system with multi-mode feature fusion

CN121140872BCN 121140872 BCN121140872 BCN 121140872BCN-121140872-B

Abstract

The application relates to the technical field of actuating sensors, and provides a multi-mode feature fusion actuator stroke calibration system which comprises a sensor module, a signal conditioning module, a data extraction module, a multi-mode feature fusion module and a stroke output and calibration module, wherein the sensor module is used for collecting multi-dimensional information of a plurality of sensors reflecting strokes and states of actuators, the signal conditioning module is used for amplifying and compressing Ji Duowei-degree information, the data extraction module is used for carrying out dimensional compression on aligned multi-dimensional information and outputting feature vectors, the multi-mode feature fusion module is used for fusing the feature vectors and outputting stroke estimation values, and the stroke calibration module is used for correcting the stroke estimation values and outputting final stroke values. The application solves the problems of limited precision, poor robustness, poor dynamic performance and weak environmental adaptability of a single actuation sensor in the prior art.

Inventors

FAN SHENGHONG
FAN WENJIE
JIANG JIAQING
ZHANG TIANFEI
LIU HUI

Assignees

北京普达迪泰科技有限公司

Dates

Publication Date: 20260512
Application Date: 20250903

Claims (6)

1. An actuator travel calibration system with multi-modal feature fusion, comprising: The system comprises a sensor module, a signal conditioning module, a data extraction module, a multi-mode feature fusion module and a stroke output and calibration module; the sensor module is used for collecting multidimensional information of strokes and states of the actuators reflected by the sensors; The signal conditioning module is used for amplifying and processing Ji Duowei-degree information; The data extraction module is used for carrying out dimension compression on the aligned multidimensional information and outputting a feature vector; the multi-mode feature fusion module is used for fusing feature vectors and outputting a travel estimation value; the stroke output and calibration module is used for correcting the stroke estimated value and outputting a final stroke value; The correcting the estimated stroke value and outputting a final stroke value includes: identifying an abnormal sensor, preprocessing abnormal sensor data, calculating a correlation value of the abnormal sensor and the actuator stroke, distributing dynamic weight to the abnormal sensor, generating a correction amount based on the correlation value and the dynamic weight, and outputting a final stroke value based on the correction amount and the stroke estimation value; The identifying anomaly sensor includes: The multidimensional information deviation of each sensor is calculated in real time, and the formula is as follows: In the formula, As the deviation value of the sensor m, For the data value of sensor m at time t, Is a steady state reference value for sensor m; When (when) The time is marked as an abnormal sensor source, Is the historical standard deviation; the abnormal sensor data is preprocessed according to the following formula: In the formula, As the normalized data of the sensor m, As a normalized sequence of the run-length references, For the mean value of sensor m in the observation window, For the standard deviation of sensor m in the observation window, As a reference for the travel distance, Is that Is used for the average value of (a), Is that Standard deviation of (2).
2. The actuator travel calibration system with multi-modal feature fusion as claimed in claim 1, characterized in that the sensor comprises: displacement sensing unit, vibration sensing unit and environment sensing unit.
3. The actuator travel calibration system of claim 1, wherein the fusing feature vectors and outputting travel estimates comprises: and outputting the travel estimation value by the feature vector through the physical model constraint layer and the dynamic weighting fusion layer respectively.
4. The actuator travel calibration system with multi-modal feature fusion of claim 1, wherein the calculating the correlation value of the anomaly sensor and the actuator travel is formulated as follows: In the formula, For the value of the association of the sensor m with the stroke, As a total number of samples, For the purpose of normalizing the data mean value, Is the normalized reference mean.
5. The actuator travel calibration system of claim 1, wherein the dynamic weights are assigned to the anomaly sensors according to the following formula: In the formula, As a fusion coefficient of the sensor m, For the cross-correlation energy of sensors m and n, As the time-lag parameter, Is the maximum time lag.
6. The actuator travel calibration system of claim 4 or 5, wherein the correction amount is generated based on the correlation value and the dynamic weight, and the formula is as follows: In the formula, For the local compensation quantity of the sensor m, For the correction amount of the correction amount, Is an adaptive gain factor.

Description

Actuator stroke calibration system with multi-mode feature fusion Technical Field The invention belongs to the technical field of actuation sensors, and particularly relates to a multi-mode characteristic fusion actuator stroke calibration system. Background Conventional calibration systems typically rely on a single position sensor such as an LVDT, a magnetic grid scale, an encoder. Is susceptible to installation errors, temperature drift, electromagnetic interference, mechanical wear and nonlinear characteristics, and accuracy is reduced or fails under severe working conditions. In addition, most calibration is performed in static or quasi-static state, and it is difficult to reflect the real stroke characteristics of the actuator in the dynamic motion process, so that a high-precision, high-robustness actuator stroke calibration system capable of adapting to the dynamic process and the complex environment is needed. Disclosure of Invention In view of the above-mentioned drawbacks of the prior art, the present invention proposes a multi-mode feature fusion actuator stroke calibration system, which is used for solving the problems of limited accuracy, poor robustness, poor dynamic performance and weak environmental adaptability of a single actuation sensor in the prior art, and the technical scheme steps of the present invention include: The system comprises a sensor module, a signal conditioning module, a data extraction module, a multi-mode feature fusion module and a stroke output and calibration module; the sensor module is used for collecting multidimensional information of strokes and states of the actuators reflected by the sensors; The signal conditioning module is used for amplifying and processing Ji Duowei-degree information; The data extraction module is used for carrying out dimension compression on the aligned multidimensional information and outputting a feature vector; the multi-mode feature fusion module is used for fusing feature vectors and outputting a travel estimation value; The stroke calibration module is used for correcting the stroke estimated value and outputting a final stroke value. Preferably, the sensor comprises: displacement sensing unit, vibration sensing unit and environment sensing unit. Preferably, the fusing the feature vectors and outputting the stroke estimation value includes: and outputting the travel estimation value by the feature vector through the physical model constraint layer and the dynamic weighting fusion layer respectively. Preferably, the correcting the stroke estimation value and outputting the final stroke value includes: The method comprises the steps of identifying an abnormal sensor, preprocessing abnormal sensor data, calculating a correlation value of the abnormal sensor and a stroke of an actuator, distributing dynamic weights to the abnormal sensor, generating a correction amount based on the correlation value and the dynamic weights, and outputting a final stroke value based on the correction amount and a stroke estimated value. Preferably, the identifying an abnormal sensor source includes: The multidimensional information deviation of each sensor is calculated in real time, and the formula is as follows: In the formula, As the deviation value of the sensor m,For the data value of sensor m at time t,Is a steady state reference value for sensor m; When (when) The time is marked as an abnormal sensor source,Is the historical standard deviation. Preferably, the preprocessing is performed on the abnormal sensor data, and the formula is as follows: In the formula, As the normalized data of the sensor m,As a normalized sequence of the run-length references,For the mean value of sensor m in the observation window,For the standard deviation of sensor m in the observation window,As a reference for the travel distance,Is thatIs used for the average value of (a),Is thatStandard deviation of (2). Preferably, the calculation of the correlation value between the abnormality sensor and the actuator stroke is as follows: In the formula, For the value of the association of the sensor m with the stroke,As a total number of samples,For the purpose of normalizing the data mean value,Is the normalized reference mean. Preferably, the dynamic weight is assigned to the abnormal sensor, and the formula is as follows: In the formula, As a fusion coefficient of the sensor m,For the cross-correlation energy of sensors m and n,As the time-lag parameter,Is the maximum time lag. Preferably, the correction amount is generated based on the correlation value and the dynamic weight, and the formula is as follows: In the formula, For the local compensation quantity of the sensor m,For the correction amount of the correction amount,Is an adaptive gain factor. The beneficial effects are that: The application provides a multi-mode feature fusion actuator stroke calibration system, which is characterized in that displacement/vibration/temperature sensors work cooperatively, cross-c