CN-122018316-A - Real-time optimization method and system for robot driving module based on multi-mode sensing

CN122018316ACN 122018316 ACN122018316 ACN 122018316ACN-122018316-A

Abstract

The invention is suitable for the technical field of robot control, and provides a real-time optimization method and a real-time optimization system for a robot driving module based on multi-mode sensing, wherein the method comprises the following steps of sampling the driving module to obtain multi-mode original data representing electric, thermal and mechanical vibration; the method comprises the steps of calculating an electric-thermal association degree, an electric-vibration association degree and a heat-vibration association degree as initial characteristics, calculating a coupling stress index based on original data and the initial characteristics, dynamically adjusting current limit values, rotating speed limit values and scaling factors of a controller bandwidth limit value according to real-time numerical values of the coupling stress index in combination with a preset mapping function, determining a real-time feedforward compensation coefficient, reconstructing a safety boundary, constructing an optimization problem with tracking errors, energy loss and control smoothness as optimization targets by taking the safety boundary as hard constraint, and solving the optimization problem to obtain optimal current loop proportional gain and integral gain. The invention can maintain the system performance to the maximum degree on the premise of ensuring the safety.

Inventors

SHEN CHAO
FENG NUAN
MENG ZHAOJUN
LI WENYI
CHEN XIANGMIN
WANG YIYANG
QU BOYANG
GUO HAIFENG
ZHOU ZHENCHAO
CHEN YUE
Tai Yuanzheng

Assignees

辽宁科技学院

Dates

Publication Date: 20260512
Application Date: 20260204

Claims (10)

1. The real-time optimization method of the robot driving module based on multi-mode sensing is characterized by comprising the following steps of: Based on a sliding time window, carrying out time-frequency analysis and association analysis on the multi-mode original data, and dynamically calculating the electric-thermal association degree, the electric-vibration association degree and the heat-vibration association degree as initial characteristics; calculating a coupling stress index in real time based on the original data and the initial characteristics, dynamically adjusting scaling factors of a current limit value, a rotating speed limit value and a controller bandwidth limit value according to a real-time numerical value of the coupling stress index and combining a preset mapping function, determining a real-time feedforward compensation coefficient, and reconstructing a safety boundary; Constructing a real-time optimization problem with tracking error, energy loss and control smoothness as optimization targets by taking a safety boundary as hard constraint, solving the optimization problem based on a real-time state, and obtaining optimal current loop proportional gain and integral gain; And applying the solved optimal control parameters to the current regulator to realize real-time optimal control of the robot driving motor.
2. The method for optimizing a driving module of a robot in real time based on multi-modal sensing according to claim 1, wherein the step of dynamically calculating the electric-thermal association, the electric-vibration association and the thermal-vibration association comprises the following steps: The covariance relation of the effective value of the quadrature axis current and the temperature rise rate of the winding on the time sequence is analyzed, and the deviation of the current rotating speed and the rated rotating speed is combined for weighting correction, so that the electric-thermal association degree is obtained; Respectively carrying out frequency spectrum analysis on the direct-axis current signal and the axial vibration acceleration signal, calculating the coherence intensity of the current frequency spectrum component and the vibration frequency spectrum component at the harmonic frequency of the main current, and carrying out weighted summation by taking the importance of each harmonic as the weight to obtain the electric-vibration association degree; and calculating the pearson correlation coefficient of the winding temperature time sequence signal and the vibration acceleration root mean square value time sequence signal in a window to obtain the heat-vibration correlation degree.
3. The method for optimizing a driving module of a robot in real time based on multi-modal sensing according to claim 1, wherein the step of calculating the coupling stress index in real time based on the raw data and the initial characteristics comprises the steps of: constructing a coupling stress index basic model, wherein the model is based on multiplication of normalized real-time torque current instruction, winding temperature rise rate, product term of electric-vibration association degree and vibration amplitude and power function three terms based on the ratio of rated voltage to real-time voltage; Constructing a multi-physical field cross-coupling compensation model, wherein the model linearly combines the product of the electric-thermal association degree and the relative temperature rise and the product of the thermal-vibration association degree and the relative vibration intensity to serve as compensation items for a basic model; And carrying out weighted fusion on the basic model output and the compensation term output to obtain a final coupling stress index, wherein the formula is CSI (t) =gamma 1 ×CSIbase(t)+γ 2 multiplied by CSIcoupling (t), wherein CSI (t) is the coupling stress index at the moment t, CSIbase (t) is a basic model output value, CSIcoupling (t) is a cross coupling compensation term output value, and gamma 1 and gamma 2 are weight coefficients.
4. The method for optimizing a robot driving module in real time based on multi-modal sensing according to claim 1, wherein the step of dynamically adjusting scaling factors of the current limit, the rotation speed limit and the controller bandwidth limit comprises: the real-time coupling stress index, the current adjustment threshold and the real-time bus voltage drop condition are input into an S-shaped attenuation function together, and a current limit scaling factor between 0 and 1 is calculated; inputting the real-time coupling stress index, the rotating speed adjustment threshold value and the real-time winding temperature into an exponential decay function together, and calculating to obtain a rotating speed limit value scaling factor between 0 and 1; inputting the real-time coupling stress index, the bandwidth adjustment threshold and the real-time electric-vibration association degree into a hyperbolic tangent adjustment function together, and calculating to obtain a bandwidth scaling factor between 0 and 1; And obtaining a real-time current limit value, a real-time rotating speed limit value and a controller parameter value range according to the scaling factor.
5. The method for optimizing a robot driving module in real time based on multi-modal sensing as set forth in claim 4, wherein determining the real-time feedforward compensation coefficient comprises inputting a real-time coupled stress index, a rotational speed change rate and a real-time electro-thermal correlation together into a dynamic response function, calculating to obtain a feedforward gain scaling factor, and multiplying the feedforward gain scaling factor by a nominal feedforward gain to obtain the real-time feedforward compensation coefficient.
6. The method for optimizing a driving module of a robot in real time based on multi-modal sensing according to claim 4, wherein the step of constructing a real-time optimization problem to obtain an optimal current loop proportional gain and integral gain comprises: An optimized objective function comprising four sub-objectives, namely the accuracy of rotation speed tracking, the energy efficiency of motor operation, the stability of control output and the smoothness of control parameter self-variation, is constructed; setting a real-time current limit value, a rotating speed limit value and a controller parameter value range as hard constraint conditions for optimizing a solving process; Taking the real-time feedforward compensation coefficient as a known quantity, and taking the proportional gain and the integral gain of the current loop as variables to be optimized; and adopting a gradient descent algorithm, carrying out online iterative solution on the optimized objective function under the hard constraint condition, and updating the optimal values of the proportional gain and the integral gain in real time.
7. Real-time optimization system of robot drive module based on multimode perception, characterized by that, the system includes: The correlation degree determining module is used for sampling the driving module to obtain multi-mode original data representing electric, thermal and mechanical vibration, carrying out time-frequency analysis and correlation analysis on the multi-mode original data based on a sliding time window, and dynamically calculating the electric-thermal correlation degree, the electric-vibration correlation degree and the thermal-vibration correlation degree to serve as initial characteristics; The safety boundary reconstruction module is used for calculating a coupling stress index in real time based on the original data and the initial characteristics, dynamically adjusting scaling factors of a current limit value, a rotating speed limit value and a controller bandwidth limit value according to a real-time numerical value of the coupling stress index and combining a preset mapping function, determining a real-time feedforward compensation coefficient, and reconstructing a safety boundary; the optimization problem solving module is used for constructing a real-time optimization problem with tracking error, energy loss and control smoothness as optimization targets by taking the safety boundary as hard constraint; and the control parameter application module is used for applying the solved optimal control parameters to the current regulator to realize real-time optimal control of the robot driving motor.
8. The multi-modal awareness based robot drive module real-time optimization system of claim 7 wherein the relevancy determination module comprises: The electric heating association degree unit is used for obtaining the electric heating association degree by analyzing the covariance relation between the effective value of the quadrature current and the temperature rise rate of the winding on the time sequence and carrying out weighted correction by combining the deviation of the current rotating speed and the rated rotating speed; The electric vibration correlation degree unit is used for respectively carrying out frequency spectrum analysis on the direct-axis current signal and the axial vibration acceleration signal, calculating the coherence intensity of the current frequency spectrum component and the vibration frequency spectrum component at the harmonic frequency of the main current, and carrying out weighted summation by taking the importance of each harmonic as the weight to obtain the electric vibration correlation degree; and the thermal vibration correlation degree unit is used for calculating the pearson correlation coefficient of the winding temperature time sequence signal and the vibration acceleration root mean square value time sequence signal in a window to obtain the thermal-vibration correlation degree.
9. The multi-modal awareness based robot drive module real-time optimization system of claim 7 wherein the safety boundary reconstruction module comprises: The basic model unit is used for constructing a coupling stress index basic model, and the model is based on multiplication of a normalized real-time torque current instruction, a winding temperature rise rate, a product term of electric-vibration association degree and vibration amplitude and a power function based on the ratio of rated voltage to real-time voltage; The compensation model unit is used for constructing a multi-physical-field cross-coupling compensation model, and the model linearly combines the product of the electric-thermal association degree and the relative temperature rise and the product of the thermal-vibration association degree and the relative vibration intensity to serve as a compensation item for the basic model; The stress index unit is used for carrying out weighted fusion on the basic model output and the compensation term output to obtain a final coupled stress index, wherein the formula is CSI (t) =gamma 1 ×CSIbase(t)+γ 2 multiplied by CSIcoupling (t), the CSI (t) is the coupled stress index at the time t, CSIbase (t) is a basic model output value, CSIcoupling (t) is a cross-coupling compensation term output value, and gamma 1 and gamma 2 are weight coefficients.
10. The multi-modal awareness based robot drive module real-time optimization system of claim 7 wherein the safety boundary reconstruction module further comprises: The current scaling factor unit is used for inputting the real-time coupling stress index, the current adjustment threshold value and the real-time bus voltage drop condition into an S-shaped attenuation function together, and calculating to obtain a current limit scaling factor between 0 and 1; The rotating speed scaling factor unit is used for inputting the real-time coupling stress index, the rotating speed adjustment threshold value and the real-time winding temperature into an exponential decay function together, and calculating to obtain a rotating speed limit value scaling factor between 0 and 1; the bandwidth scaling factor unit is used for inputting the real-time coupling stress index, the bandwidth adjustment threshold and the real-time electric-vibration association degree into a hyperbolic tangent adjustment function together, and calculating to obtain a bandwidth scaling factor between 0 and 1; And the limit value determining unit is used for obtaining a real-time current limit value, a real-time rotating speed limit value and a controller parameter value range according to the scaling factor.

Description

Real-time optimization method and system for robot driving module based on multi-mode sensing Technical Field The invention relates to the technical field of robot control, in particular to a real-time optimization method and system for a robot driving module based on multi-mode sensing. Background The traditional drive control system is usually controlled based on single or small amount of electric parameters such as current, voltage, rotating speed and the like, and lacks synchronous monitoring and comprehensive analysis of multiple physical fields such as temperature field distribution, mechanical vibration and the like. This results in the system not being able to fully sense its own operating condition, especially under high load, high dynamic conditions, the coupling effect between electro-thermo-mechanical vibrations is neglected, and it is difficult for the controller to predict potential overheating, resonance or stability degradation risks. Part of advanced control strategies introduce online optimization, but are usually optimized only for a single performance index (such as a minimum tracking error), and the cooperation of multiple targets such as energy efficiency, temperature rise, vibration and the like is ignored. In addition, the optimization process is often disjointed from the real-time safety boundary of the system, so that control instructions which are mathematically optimal but have high risks (such as being close to overheat limits) in practice are easy to generate, and a mechanism for integrating real-time risk perception into optimization targets and constraints is lacked. In addition, when dealing with disturbance such as sudden load and rugged road surface, the traditional feedforward compensation coefficient is mostly a fixed value or is calculated based on a simple model, and cannot be dynamically adjusted according to the real-time perceived system coupling state, so that the dynamic response is slow and the disturbance rejection capability is limited. Therefore, there is a need to provide a method and a system for real-time optimization of a robot driving module based on multi-modal sensing, which aims to solve the above problems. Disclosure of Invention Aiming at the defects existing in the prior art, the invention aims to provide a real-time optimization method and a real-time optimization system for a robot driving module based on multi-mode sensing so as to solve the problems existing in the background art. The invention is realized in such a way that a robot driving module real-time optimization method based on multi-mode sensing comprises the following steps: Based on a sliding time window, carrying out time-frequency analysis and association analysis on the multi-mode original data, and dynamically calculating the electric-thermal association degree, the electric-vibration association degree and the heat-vibration association degree as initial characteristics; calculating a coupling stress index in real time based on the original data and the initial characteristics, dynamically adjusting scaling factors of a current limit value, a rotating speed limit value and a controller bandwidth limit value according to a real-time numerical value of the coupling stress index and combining a preset mapping function, determining a real-time feedforward compensation coefficient, and reconstructing a safety boundary; Constructing a real-time optimization problem with tracking error, energy loss and control smoothness as optimization targets by taking a safety boundary as hard constraint, solving the optimization problem based on a real-time state, and obtaining optimal current loop proportional gain and integral gain; And applying the solved optimal control parameters to the current regulator to realize real-time optimal control of the robot driving motor. As a further scheme of the invention, the step of dynamically calculating the electric-thermal association degree, the electric-vibration association degree and the thermal-vibration association degree specifically comprises the following steps: The covariance relation of the effective value of the quadrature axis current and the temperature rise rate of the winding on the time sequence is analyzed, and the deviation of the current rotating speed and the rated rotating speed is combined for weighting correction, so that the electric-thermal association degree is obtained; Respectively carrying out frequency spectrum analysis on the direct-axis current signal and the axial vibration acceleration signal, calculating the coherence intensity of the current frequency spectrum component and the vibration frequency spectrum component at the harmonic frequency of the main current, and carrying out weighted summation by taking the importance of each harmonic as the weight to obtain the electric-vibration association degree; and calculating the pearson correlation coefficient of the winding temperature time sequence signal and the vibration