Motor housing automatic coding control method and device

Abstract

The application discloses a kind of motor shell automatic coding control method and device, it is related to motor shell processing technical field, comprising: the datum information on the surface of motor shell is identified by vision sensor, and the pose state of marking equipment working end is adjusted;Sensor data set of marking equipment working end is collected in real time by multi-sensor fusion system;The collected sensor data set is analyzed and processed based on intelligent decision algorithm, judges positioning state and generates control instruction, predicts the wear state of positioning system based on machine learning model, and generates intelligent decision instruction in combination with classification evaluation result, position compensation and wear state;According to intelligent decision instruction, marking equipment is controlled to execute coding operation on motor shell. Physical limiting positioning is carried out by setting asymmetric distribution positioning groove, and real-time data of coding operation is obtained by cooperating with multi-sensor system, to realize the automatic coding adjustment control of motor shell, improve the accuracy and the degree of beauty of coding operation.

Description

Technical Field

[0001] This invention relates to the field of motor housing processing technology, specifically to an automated marking control method and device for motor housings. Background Technology

[0002] In the motor manufacturing process, the marking operation on the motor housing mainly relies on physical limiting blocks and mechanical clamping devices to fix the position of the housing. The marking area is aligned through a preset rigid positioning structure, thus enabling the marking operation. However, due to manufacturing tolerances on the surface of the motor housing and positioning errors in the clamping mechanism, marking position deviations or incorrect marking directions can easily occur, affecting the reliability of the motor housing marking operation.

[0003] Furthermore, existing motor housing marking systems lack intelligent adjustments for the marking operation. During long-term marking operations, uneven marking contact force can easily lead to inconsistent marking depths on multiple motor housings, reducing the reliability and aesthetics of the motor housing product markings. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of the prior art. This invention provides an automated marking control method and device for motor housings. By setting asymmetrically distributed positioning grooves for physical limiting and positioning, and cooperating with a multi-sensor system to acquire real-time data of the marking operation, the automated marking adjustment control of the motor housing is realized, thereby improving the accuracy and aesthetics of the motor housing.

[0005] This invention provides an automated marking control method for motor housings, the automated marking control method comprising: The motor housing to be marked is transported to the marking station. The asymmetrical distribution of positioning grooves in the marking area on the surface of the motor housing is identified by a vision sensor to obtain reference information. The position and posture of the working end of the marking equipment are adjusted according to the reference information to enter the marking working state. The multi-sensor fusion system collects in real time the positioning contact force data, position data and image data of the working end of the marking device in the marking area of ​​the motor housing, and obtains a sensor data set. The positioning contact force is classified and evaluated by fuzzy logic rules, the position compensation amount is calculated by adaptive PID control algorithm, the wear state of the positioning system is predicted by machine learning model, and intelligent decision instructions are generated by combining the classification evaluation results, position compensation amount and wear state. The intelligent decision-making command controls the marking equipment to perform a marking operation on the motor housing.

[0006] Furthermore, the step of conveying the motor housing to be marked to the marking station and obtaining reference information by identifying asymmetrically distributed positioning grooves using a vision sensor includes: A servo conveyor belt is used to transport the motor housing to be marked to the positioning fixture at the marking station. The positioning fixture fixes the position of the motor housing by a pneumatic clamping mechanism. Using an industrial area array camera as a vision sensor, and a ring light source to illuminate the marking area on the motor housing, the motor housing to be marked is photographed to obtain image data of the motor housing to be marked. The image data of the motor housing to be marked is processed by image grayscale conversion, Gaussian filtering and edge detection algorithms to obtain the asymmetrical distribution of positioning groove contours of the marking area on the surface of the motor housing to be marked; The center coordinates, normal direction, and distance between two asymmetrically distributed positioning grooves are determined by template matching algorithm. The center coordinates, normal direction, and distance between the two grooves are stored and marked as reference information. Furthermore, adjusting the pose state of the marking device's working end according to the reference information to enter the marking working state includes: The X-axis and Y-axis displacement compensation values ​​of the working end of the marking equipment are calculated based on the center coordinates in the reference information, and the rotation angle compensation value of the working end is calculated based on the normal direction. The position and orientation of the working end of the marking device are adjusted according to the displacement compensation value and the rotation angle compensation value, so that the center of the marking needle at the working end is aligned with the origin of the reference coordinate system formed by the positioning groove, and a preparation signal for the marking working state is generated.

[0007] Furthermore, the multi-sensor fusion system collects contact force data, position data, and image data in real time from the working end of the marking device in the marking area of ​​the motor housing, resulting in a sensor data set including: The force sensor collects the contact pressure data between the marking needle and the motor housing surface in real time, the displacement sensor collects the real-time position coordinate data of the working end, and the vision sensor collects the image data during the marking process. The Kalman filter algorithm is used to perform time synchronization and noise filtering on the collected contact pressure data, real-time position coordinate data and image data to obtain a noise-reduced sensor data set. Furthermore, the classification and evaluation of positioning contact force using fuzzy logic rules includes: A fuzzy logic evaluation model is established, which divides the contact force data into several fuzzy subsets, uses the rate of change of the positioning contact force as an auxiliary input variable, and classifies the positioning contact force state in conjunction with preset fuzzy rules, and outputs the evaluation results.

[0008] Furthermore, the application of the adaptive PID control algorithm to calculate the position compensation includes: The deviation between the actual position and the reference position of the marking device's working end is used as the input data for the PID controller. The parameters of the PID controller are adjusted according to the input data. At the same time, the KI value is finely adjusted according to the contact force fluctuation fed back by the force sensor, and the output value of the PID controller is obtained. The position compensation amount of the marking device's working end is calculated based on the output value of the PID controller. Furthermore, the prediction of the wear status of the positioning system based on the machine learning model includes: A random forest regression model is used as a machine learning model to collect historical operating data of the positioning system, extract wear-related features from the historical data, and construct a training dataset based on the wear-related features. The machine learning model is trained using a training dataset, and the trained machine learning model is then fed with current running data to predict the wear and tear of the positioning system. A wear compensation value is set according to the wear level, and the wear compensation amount is incorporated into the intelligent decision-making command.

[0009] Furthermore, the step of controlling the coding device to perform coding operations according to intelligent decision instructions includes: The marking device's working end is set with descent depth, moving speed, and marking force according to the intelligent decision-making instructions. The working end of the driving marking device moves along a preset trajectory in the marking area of ​​the motor housing, while real-time feedback data of the marking operation is obtained through multiple sensor data. The system detects out-of-limit data on contact force or position deviation based on the real-time feedback data, and adjusts the coding operation accordingly.

[0010] Furthermore, the automated coding control method also includes: The quality of the motor housing after the coding operation is completed is inspected using a visual sensor to obtain inspection data. Based on the detection data, the motor housings that have completed the coding operation are classified and cut into smaller parts.

[0011] The present invention also provides an automated marking control device for motor housings, the automated marking control device comprising: Positioning module: Used to transport the motor housing to be marked to the marking station, identify the asymmetrical distribution of positioning grooves on the marking area of ​​the motor housing surface through a vision sensor, obtain reference information, and adjust the position and posture of the working end of the marking equipment according to the reference information to enter the marking working state; Data acquisition module: used to acquire in real time the contact force data, position data and image data of the marking device working end in the marking area of ​​the motor housing through a multi-sensor fusion system; Intelligent decision module: It is used to classify and evaluate the positioning contact force through fuzzy logic rules, calculate the position compensation amount by applying an adaptive PID control algorithm, predict the wear state of the positioning system based on a machine learning model, and generate intelligent decision instructions by combining the classification and evaluation results, the position compensation amount and the wear state. The coding module is used to control the coding equipment to perform coding operations on the motor housing according to the intelligent decision instructions.

[0012] This invention provides an automated marking control method and device for motor housings. By setting asymmetrically distributed positioning grooves for physical limiting and positioning, and cooperating with a multi-sensor system to acquire real-time data of the marking operation, the automated marking adjustment control of the motor housing is realized, thereby improving the accuracy and aesthetics of the motor housing. Attached Figure Description

[0013] Figure 1 This is a flowchart of the automated marking control method for motor housing in an embodiment of the present invention; Figure 2 This is a schematic diagram of the marking area structure of the motor housing according to an embodiment of the present invention; Figure 3 This is a schematic diagram of the marking state of the motor housing in an embodiment of the present invention; Figure 4 This is a schematic diagram of the automated marking control device for the motor housing in an embodiment of the present invention. Detailed Implementation

[0014] 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.

[0015] Example 1: Figure 1 A flowchart of the automated marking control method for motor housing in an embodiment of the present invention is shown. Figure 2 This diagram illustrates the marking area structure of the motor housing according to an embodiment of the present invention. Figure 3A schematic diagram of the marking state of the motor housing in an embodiment of the present invention is shown. The motor housing 1 itself has at least two asymmetrically distributed positioning grooves, including a first positioning groove 111 located at the upper left corner of the marking area 11 and a second positioning groove 112 located at the lower right corner of the marking area 11, forming an error-proof positioning reference. The marking device 2 is provided with a positioning protrusion that matches the positioning groove. Physical positioning is achieved by the positioning protrusion of the marking device 2 and the positioning groove of the marking area 11 of the motor housing 1, thereby improving the accuracy of the marking operation of the motor housing 1.

[0016] Furthermore, the asymmetrically distributed first positioning groove 111 and second positioning groove 112 can be achieved by combining grooves of different shapes or depths, such as an asymmetrical arrangement of circular and rectangular grooves, or a specific spacing distribution of triangular and trapezoidal grooves, thereby forming an unambiguous spatial reference base, avoiding orientation identification errors caused by symmetrical structures, and improving the accuracy of alignment and positioning between the marking device 2 and the motor housing 1.

[0017] Specifically, the automated coding control method includes: S11: The motor housing 1 to be marked is transported to the marking station. The asymmetrical distribution of positioning grooves in the marking area 11 on the surface of the motor housing 1 is identified by a vision sensor to obtain reference information. The position and posture of the working end of the marking device 2 are adjusted according to the reference information to enter the marking working state.

[0018] Specifically, a servo conveyor belt is used to transport the motor housing 1 to be marked to the positioning fixture at the marking station. The positioning fixture fixes the position of the motor housing 1 by a pneumatic clamping mechanism. The belt conveyor mechanism driven by a stepper motor or servo motor is used to ensure that the motor housing 1 arrives at the marking station stably and reduces position fluctuations during the transport process.

[0019] An industrial area scan camera is used as a vision sensor, along with a ring light source to illuminate the marking area 11 of the motor housing 1, capturing images of the motor housing 1 to be marked and obtaining image data. The industrial area scan camera is an industrial-grade imaging device with two-dimensional image capture capabilities. It forms an area scan imaging module composed of CMOS or CCD image sensors, enabling it to acquire complete area images of the motor housing 1 surface, providing high-resolution raw data for subsequent processing. The ring light source provides uniform illumination to the imaging area of ​​the industrial area scan camera, eliminating local shadows and reflections, improving image contrast, and thus enhancing the accuracy of visual recognition.

[0020] Furthermore, the industrial area array camera can adopt a ring lighting structure composed of multi-segment LED arrays, which can meet the lighting requirements of the marking area 11 of the motor housing 1.

[0021] The image data of the motor housing 1 to be marked is processed by image grayscale conversion, Gaussian filtering, and edge detection algorithms to obtain the asymmetrical distribution of the positioning groove contours of the marking area on the surface of the motor housing 1. Gaussian filtering refers to an image smoothing technique based on the Gaussian distribution function, which is achieved by convolving the image with a convolution kernel. It can suppress high-frequency noise while preserving contour edge features. Edge detection algorithms are processing methods for identifying regions of abrupt changes in image brightness. They can be implemented using algorithms such as the Sobel operator or the Canny operator, thereby accurately separating the boundary contours of the positioning grooves.

[0022] The center coordinates, normal direction, and distance between two asymmetrically distributed positioning grooves are determined using a template matching algorithm. These coordinates, normal direction, and distance are then stored and marked as reference information. The template matching algorithm is a method for target localization by comparing the similarity between an image region and a preset template. It is implemented using normalized cross-correlation or squared difference matching techniques to obtain the precise geometric parameters of the positioning grooves.

[0023] Specifically, adjusting the pose of the working end of the marking device 2 according to the reference information to enter the marking working state includes: The X-axis and Y-axis displacement compensation values ​​of the working end of the marking device 2 are calculated based on the center coordinates in the reference information, and the rotation angle compensation value of the working end is calculated based on the normal direction. The displacement compensation value refers to the correction parameter that quantifies the positional offset of the working end in the planar coordinate system. It can be implemented by coordinate transformation algorithm or vector operation method, which can accurately eliminate the positional deviation in the X-axis and Y-axis directions. The rotation angle compensation value can be understood as the angle parameter that corrects the spatial orientation deviation of the working end. It can be implemented by vector projection method or rotation matrix calculation, thereby accurately correcting the trajectory skew caused by the inconsistency of the normal direction.

[0024] The position and orientation of the working end of the marking device 2 are adjusted according to the displacement compensation value and the rotation angle compensation value, so that the center of the marking needle at the working end is aligned with the origin of the reference coordinate system formed by the positioning groove, and a preparation signal for the marking working state is generated.

[0025] Pose state adjustment refers to the dynamic optimization process of comprehensive position and orientation correction, which can be achieved by using a multi-degree-of-freedom motion control system to avoid the cumulative error caused by single-dimensional adjustment; the origin alignment of the reference coordinate system refers to establishing the spatial coincidence relationship between the center of the marking needle and the geometric center of the positioning groove, which can be achieved through closed-loop feedback control to eliminate the initial positioning error; the preparation signal can be understood as the state feedback mechanism that characterizes the completion of pose correction, which can be achieved by digital signal triggering or state register setting, and ensure that the marking operation is started only under precise alignment conditions.

[0026] Furthermore, by accurately calculating the X-axis and Y-axis displacement compensation values ​​using the center coordinates, the planar position offset is directly reflected and corrected, allowing the working end to move to the theoretical origin position of the reference coordinate system. Simultaneously, the rotation angle compensation value is calculated based on the normal direction to quantify the rotational deviation of the working end, ensuring that the needle direction is consistent with the spatial orientation of the reference coordinate system. Through the synergistic application of displacement compensation values ​​and rotation angle compensation values, the spatial position and orientation of the working end are comprehensively corrected, forming a complete pose optimization mechanism. On this basis, the center of the marking needle and the origin of the reference coordinate system formed by the positioning groove are spatially aligned, completely eliminating the initial positioning error. Finally, a preparation signal for the marking working state is generated, providing real-time feedback on the system's correction completion status, and providing reliable start conditions for subsequent marking operations.

[0027] S12: The positioning contact force data, position data and image data of the working end of the marking device 2 in the marking area 11 of the motor housing 1 are collected in real time through the multi-sensor fusion system to obtain the sensor data set.

[0028] Specifically, the multi-sensor fusion system collects in real time the contact force data, position data, and image data of the marking device 2's working end in the marking area 11 of the motor housing 1, resulting in a sensor data set including: The contact pressure data between the marking needle and the surface of the motor housing 1 is collected in real time by a force sensor, the real-time position coordinate data of the working end is collected by a displacement sensor, and the image data of the marking process is collected by a vision sensor. Furthermore, the working end of the marking device 2 is equipped with a force sensor. The force sensor is a sensing device used to measure contact pressure. The force sensor can be a strain gauge sensor, a piezoelectric sensor, or a capacitive sensor. It can accurately capture the interaction force changes between the marking needle and the shell surface, providing basic data for force state assessment.

[0029] Furthermore, the marking device 2 is equipped with a displacement sensor, which is a measuring device for detecting spatial position coordinates. It can be implemented using a laser displacement sensor, a grating ruler, or a rotary encoder to ensure that the position information of the working end in three-dimensional space is accurate and to support precise positioning operations.

[0030] A Kalman filter algorithm is used to perform time synchronization and noise filtering on the acquired contact pressure data, real-time position coordinate data, and image data, resulting in a denoised sensor data set. Standard Kalman filtering, extended Kalman filtering, or adaptive Kalman filtering are then employed to eliminate time skew in the multi-source data and suppress random noise, ensuring the reliability of the data set.

[0031] Specifically, force sensors, displacement sensors, and vision sensors acquire their respective data streams in parallel, forming a multi-source heterogeneous data input. The Kalman filter algorithm establishes a state prediction model based on the dynamic characteristics of the sensor data, and uses the weighted fusion of measured and predicted values ​​to perform optimal estimation, eliminating time deviations caused by differences in sampling frequencies between different sensors, while suppressing random noise introduced by environmental interference. This processing mechanism ensures that contact force, position, and image data are strictly aligned in the time dimension and significantly improves the signal-to-noise ratio of the data, thereby outputting a noise-reduced synchronous sensor data set to provide high-quality input for the intelligent decision-making module.

[0032] S13: Classify and evaluate the positioning contact force through fuzzy logic rules, calculate the position compensation amount by applying an adaptive PID control algorithm, predict the wear state of the positioning system based on a machine learning model, and generate intelligent decision instructions by combining the classification and evaluation results, the position compensation amount and the wear state. Specifically, the classification and evaluation of positioning contact force using fuzzy logic rules includes: A fuzzy logic evaluation model is established, dividing the contact force data into several fuzzy subsets. The rate of change of the positioning contact force is used as an auxiliary input variable. Preset fuzzy rules are used to classify the positioning contact force state and output the evaluation result. The positioning contact force is classified and evaluated using fuzzy logic rules, dividing the contact force data into four fuzzy subsets: a low contact force set (contact force data in the range of 0-1N); a normal contact force set (contact force data in the range of 1-3N); a high contact force set (contact force data in the range of 3-5N); and an over-limit contact force set (contact force data in the range of 5N or more). The rate of change of the contact force data is used as an auxiliary input variable. Preset fuzzy rules, such as setting "contact force > 5N and rate of change > 2N / s" as the judgment rule for abnormal jamming, are applied. The judgment rule for determining normal contact is set as "contact force within the range of 1~3N and change rate <0.5N / s". By classifying the positioning contact force state, four evaluation results are output: "normal contact, slight over-force, severe over-force, and abnormal jamming". The positioning contact force is divided by fuzzy rule classification, thereby improving the operational reliability of the automatic marking control system.

[0033] Furthermore, the fuzzy logic evaluation model refers to a decision-making framework based on fuzzy set theory, which can be implemented using embedded software or application-specific integrated circuits. It is used to handle the fuzziness and uncertainty of contact force data in order to achieve quantitative analysis and processing of the positioning operation between the marking device 2 and the motor housing 1.

[0034] Furthermore, the fuzzy subset refers to discretizing continuous contact force data into operable state categories, which can define different threshold ranges based on actual marking process requirements to meet the positioning requirements of positioning marking operations.

[0035] Furthermore, the contact force change rate, as an auxiliary input variable, can be understood as the rate of change of contact force over time. It can be calculated in real time using a digital differential algorithm or a hardware filtering circuit to capture the dynamic trend of force change.

[0036] Preset fuzzy rules refer to the pre-defined mapping logic from input variables to output categories. They can be stored in the rule base and executed through the fuzzy inference engine. The rules can be dynamically optimized based on historical marking data. The output evaluation result refers to the contact force status label after classification, which can be used to trigger corresponding control commands to ensure the stable operation of the marking process.

[0037] Specifically, the automated marking control method for motor housing 1 provided in this embodiment of the invention collects contact force data from the working end of the marking device 2 in real time and calculates its rate of change. The force value and rate of change are sent as dual input variables to a fuzzy logic evaluation model. The model performs inference calculations on the input variables according to preset fuzzy rules, accurately classifying the contact force state into normal contact, slight over-force, severe over-force, or abnormal jamming. After the classification result is output, it serves as a key basis for dynamically adjusting the marking parameters. For example, the operation is immediately paused when abnormal jamming is detected, and the descent depth is fine-tuned in the case of slight over-force. This forms a closed-loop control mechanism during the marking process, effectively avoiding the response lag problem caused by static force value evaluation.

[0038] Specifically, the application of the adaptive PID control algorithm to calculate the position compensation includes: The deviation between the actual position and the reference position of the working end of the marking device 2 is used as the input data of the PID controller. The parameters of the PID controller are adjusted according to the input data. At the same time, the KI value is finely adjusted according to the contact force fluctuation fed back by the force sensor, and the output value of the PID controller is obtained. The position compensation amount of the working end of the marking device 2 is calculated according to the output value of the PID controller. Furthermore, the deviation between the actual position and the reference position of the working end of the marking device 2 is used as the input of the PID controller, and the preset standard PID parameters are set, where: KP=2.0, KI=0.5, KD=0.1; the deviation between the actual position and the reference position of the working end of the marking device 2 is detected, and when the position deviation is <0.02mm, KP is adjusted to 1.6, that is, 0.8 times the preset standard value of KP, and KD is adjusted to 1.2, that is, 1.2 times the preset standard value of KD, to avoid overshoot; When the deviation value is between 0.02mm and 0.1mm, the standard parameter value is maintained. When the deviation value is greater than 0.1 mm, KP is adjusted to 3.0, which is 1.5 times the preset standard value of KP, and KD is adjusted to 0.7, which is 0.7 times the preset standard value of KD. At the same time, the KI value is finely adjusted according to the contact force fluctuation fed back by the force sensor. The adjustment range of the KI value is set between 0.3 and 0.7, which can effectively avoid integral saturation. Finally, the X / Y / Z axis position compensation amount of the working end is calculated according to the PID output value.

[0039] Specifically, when the marking device 2 detects a deviation of 0.01mm between the actual position and the reference position, the system automatically adjusts KP to 1.6 and KD to 1.2, and finely adjusts the KI value to 0.4 based on the contact force fluctuations fed back by the force sensor. The calculated X / Y / Z axis position compensation is sent to the servo drive unit to control the working end to perform fine displacement adjustments until the deviation is eliminated. During this process, the PID controller continuously receives real-time deviation signals and force fluctuation data, dynamically updates the parameter configuration, and ensures that the marking needle is stably aligned with the marking area 11.

[0040] Furthermore, the deviation value refers to the displacement difference between the actual position and the reference position of the working end of the marking device 2. It is obtained by real-time measurement using a high-precision displacement sensor, which can quantify the positioning error and serve as the basic input signal for control adjustment. Adjusting the parameters of the PID controller based on the input data can be understood as dynamically modifying the proportional coefficient KP and the derivative coefficient KD based on the deviation range. This can be achieved using threshold judgment logic or an adaptive mapping function to optimize response characteristics for different error scenarios and avoid insufficient adaptability caused by fixed parameters. Fine-tuning the KI value specifically refers to adjusting the integral coefficient based on the amplitude of contact force fluctuation. This can be achieved using proportional scaling or lookup table interpolation, which can dynamically balance the intensity of integral action and prevent integral accumulation problems caused by force fluctuations. The position compensation amount refers to the calculated displacement correction amount in the X / Y / Z axis directions, which can drive the servo motor or hydraulic actuator to complete the pose adjustment and convert the control output into precise physical displacement compensation.

[0041] Specifically, the prediction of the wear status of the positioning system based on the machine learning model includes: A random forest regression model was used as the machine learning model. Historical operational data of the positioning system was collected, wear-related features were extracted from the historical data, and a training dataset was constructed based on these features. The random forest regression model is a non-parametric prediction tool based on ensemble learning. It can be implemented using a decision tree ensemble architecture, specifically by training multiple decision trees in parallel and aggregating the output results. This approach handles non-linear relationships and high-dimensional features in historical data, avoiding the insufficient generalization ability of traditional linear models in complex wear scenarios.

[0042] Furthermore, the historical operating data includes: the number of cycles, the average contact force for each positioning, the maximum contact force, the positioning error, and the positioning time, among other relevant working data of the marking device 2. Wear-related features such as "number of cycles - contact force fluctuation value" and "positioning time - error growth rate" are extracted from the historical data. A training dataset is constructed using these wear-related features, enabling effective training of the machine learning model to accurately analyze the wear degree of the working end of the marking device 2.

[0043] The machine learning model is trained using a training dataset. The trained model is then fed current operating data to predict the wear level of the positioning system. Training the machine learning model with the training dataset improves its convergence in analyzing the wear level of the marking device 2's working end, thereby achieving accurate analysis of the wear level of the positioning system.

[0044] A wear compensation value is set according to the wear degree, and the wear compensation amount is incorporated into the intelligent decision command, so that the positioning operation of the marking device 2 can correct the wear degree of the positioning protrusion at the working end of the marking device 2, thereby improving the consistency of the marking operation of the marking device 2 on multiple motor housings 1.

[0045] Furthermore, the relationship between the number of cycles and the contact force fluctuation value of the marking device 2 is used to reflect the progressive wear of mechanical components. By calculating the correlation between the fluctuation amplitude of the contact force sequence and the number of cycles in historical data, the physical degradation process caused by wear can be quantified. Based on the analysis of the wear degree of the working end of the marking device 2, the working end can be adapted to the working requirements of the marking device 2.

[0046] The prediction results are output according to the degree of wear. By setting a threshold interval, the prediction results can be divided into four levels: no wear, slight wear, moderate wear, and severe wear. For example, the classification boundary can be set according to the numerical range of wear compensation, so that the wear assessment results strictly match the actual engineering needs, which facilitates the fine-grained formulation of subsequent compensation strategies.

[0047] Specifically, in this embodiment, a random forest regression model is used as the machine learning model to collect historical operating data of the positioning system (including the number of cycles, average contact force, maximum contact force, positioning error, and positioning time for each positioning). Wear-related features such as "number of cycles - contact force fluctuation value" and "positioning time - error growth rate" are extracted to construct a training dataset. The trained model is then used to input the current operating data in real time to predict the wear level of the positioning system (divided into "no wear, slight wear, moderate wear, and severe wear"). When the prediction is "slight wear", a wear compensation amount of 0.01-0.03 mm is generated; when the prediction is "moderate wear", a compensation amount of 0.03-0.05 mm is generated. The wear compensation amount is then incorporated into the intelligent decision-making instructions.

[0048] Furthermore, in this embodiment, the open-source machine learning library scikit-learn is used to implement a random forest regression model. The number of decision trees in the random forest regression model is set to 100 to balance computational efficiency and prediction accuracy. In the feature extraction stage, the standard deviation of the contact force within every 100 cycles is calculated as the "contact force fluctuation value" based on the number of cycles and the contact force sequence in the historical running data, and this value is normalized with the number of cycles to form a composite feature. At the same time, based on the trend of the positioning error over time, linear regression is used to fit the error growth rate and correlate it with the positioning time to generate the "positioning time - error growth rate" feature. After the training dataset is constructed, the model optimizes the hyperparameters through cross-validation to ensure the robustness of the prediction of wear status. In the real-time prediction stage, the system inputs the latest running data every 50 positioning operations. After the model outputs the wear level, the control system automatically adds the wear compensation amount of the corresponding interval to the position compensation calculation module. For example, when "slight wear" is detected, the compensation amount is dynamically generated in the range of 0.01mm to 0.03mm, and the adaptive adjustment of the compensation strategy can be completed without manual intervention.

[0049] By analyzing the wear condition of the working end of the marking device 2 using a machine learning model, the wear status of the positioning system can be accurately quantified and dynamically compensated. This effectively solves the problem of improper compensation setting caused by the lack of a wear prediction mechanism, avoids marking position deviation and unstable marking quality, and ensures the continuous consistency of marking accuracy during long-term operation.

[0050] S14: Control the marking device 2 to perform a marking operation on the motor housing 1 according to the intelligent decision command.

[0051] Specifically, controlling the coding device to perform coding operations according to intelligent decision instructions includes: The marking device 2 is configured with its working end descent depth, moving speed, and marking force according to the intelligent decision-making instructions. The descent depth of the working end of the marking device 2 can be set based on the contact force assessment results, and it descends by 0.2mm to 0.3mm during normal contact to ensure that a label is formed within the marking area 11 of the motor housing 1.

[0052] The moving speed of the working end of the marking device 2 can be set according to the position compensation amount. When the position compensation amount is small, the moving speed of the working end of the marking device 2 can be set to 50 mm / s. When the position compensation amount is large, the moving speed of the working end of the marking device 2 can be reduced to 30 mm / s.

[0053] Specifically, the marking force is finely adjusted based on the amount of wear compensation; the more severe the wear, the greater the marking force is increased by 5% to 10%.

[0054] The working end of the driving marking device 2 moves along a preset trajectory in the marking area 11 of the motor housing 1, while real-time feedback data of the marking operation is obtained through multi-sensor data. The system detects out-of-limit data on contact force or position deviation based on the real-time feedback data, and adjusts the coding operation accordingly.

[0055] After receiving the instruction, the control system of the marking equipment 2 drives the marking needle to move in the marking area 11 of the motor housing 1 along a preset trajectory (such as QR code or character trajectory). At the same time, it receives data feedback from multiple sensors in real time. If the contact force exceeds the limit or the position deviation exceeds 0.15mm, the marking is immediately paused and deviation correction is performed. After the correction is completed, the marking continues until the entire marking process is finished.

[0056] Intelligent decision-making commands refer to control commands generated by integrating multi-dimensional information. They can be implemented by combining parameters such as contact force assessment results, position compensation amount, and wear status, making the marking parameters adaptive.

[0057] Furthermore, the descent depth setting can be dynamically adjusted according to the contact force state. By adjusting the needle probing distance of the marking device 2 and using fuzzy logic evaluation results as input variables for adjustment, unclear marking or material damage can be avoided, thereby ensuring the accuracy of the marking operation of the marking device 2.

[0058] Specifically, the movement speed setting refers to the working end movement rate adjusted in real time based on the position compensation amount. It can be implemented using a compensation amount threshold judgment mechanism, which can balance marking efficiency and accuracy.

[0059] In addition, the marking force setting can be understood as a force parameter that is finely adjusted in combination with the wear condition. It can be implemented using a wear degree graded compensation algorithm to compensate for the impact of mechanical wear on the marking effect.

[0060] In practical applications, real-time feedback data acquisition refers to the continuous monitoring of dynamic information in the marking process through multi-sensor fusion. This can be achieved by synchronously acquiring force sensors, displacement sensors, and vision sensors. By monitoring the marking operation of marking device 2 in real time through a multi-sensor system, the accuracy and reliability of the marking operation of marking device 2 can be improved.

[0061] Among them, the over-limit data detection and adjustment refers to the abnormal handling mechanism triggered by threshold judgment. It can be implemented by using dual judgment logic of position deviation and contact force. The threshold data of the marking device 2 can be set according to the actual processing situation to improve the accuracy of the marking operation.

[0062] Furthermore, by constructing a closed-loop control mechanism, real-time feedback and dynamic adjustment are deeply integrated into the coding execution process. First, the descent depth, movement speed, and coding intensity are set according to intelligent decision-making instructions. These instructions integrate multi-dimensional information such as contact force assessment, position compensation, and wear status, making parameter settings adaptive. The descent depth is dynamically set based on the contact force assessment results, ensuring a reasonable depth range is maintained under normal contact conditions. The movement speed is adjusted in real-time according to the position compensation amount; high speed is used to improve efficiency when the compensation amount is small, and deceleration is used to ensure accuracy when the compensation amount is large. The coding intensity is fine-tuned in conjunction with the wear compensation amount, compensating for the impact of mechanical wear on the marking effect. Second, real-time feedback data is acquired simultaneously as the drive end moves along the preset trajectory. Multi-sensor fusion continuously monitors contact force, position, and image information, providing a reliable basis for dynamic adjustment. Third, out-of-limit data is detected based on real-time feedback data, and the coding operation is adjusted accordingly. Threshold judgments of contact force or position deviation are used as trigger conditions; once a limit is exceeded, the process is immediately paused and corrected. Finally, the entire process, through a pause-correction-resume cycle, ensures that the coding process can quickly recover in the event of an anomaly, maintaining trajectory accuracy and marking quality.

[0063] Specifically, the automated coding control method further includes: The motor housing 1, after the coding operation has been completed, undergoes quality inspection using a visual sensor to acquire inspection data. Based on this data, the motor housing 1 is then sorted and cut into smaller units. The visual sensor is a non-contact detection device used to acquire image information of an object's surface. It can be an industrial area scan camera, line scan camera, or infrared imaging equipment, avoiding secondary damage caused by physical contact and providing objective, quantitative image data. Quality inspection refers to the automatic evaluation of the clarity, positional accuracy, and completeness of the coding results. This can be achieved using image grayscale processing, edge detection algorithms, or character recognition technology, replacing subjective errors from manual inspection with objective data. Sorting and cutting refers to the mechanism that automatically guides products to different processing paths based on the quality inspection results. This can be achieved using pneumatic sorting mechanisms, robotic arm guiding devices, or conveyor belt diversion systems, ensuring that sorting decisions rely on real-time inspection data rather than experience-based judgment, thus improving resource utilization efficiency.

[0064] Specifically, the automated marking control method for the motor housing 1 abandons the traditional physical limiting method and adopts a visual guidance and intelligent decision-making mechanism to achieve high-precision marking operation. After the motor housing 1 to be marked is transported to the marking station, a vision sensor identifies the asymmetrically distributed positioning grooves on the surface of the housing to obtain reference information. The vision sensor can specifically be a high-resolution CMOS industrial camera, which has fast image capture capabilities and can clearly present the geometric features of the grooves. Based on this reference information, the pose of the working end of the marking device 2 is dynamically adjusted to ensure that the working end is precisely aligned with the marking area 11. The multi-sensor fusion system collects the positioning contact force data, position data, and image data of the working end of the marking device 2 in the marking area 11 of the motor housing 1 in real time, forming a sensor data set. In specific implementation, the system can integrate a piezoelectric force sensor and a laser displacement sensor. The former is used to monitor micro-Newton level contact force changes, and the latter provides sub-micron level position feedback. The two work together to achieve multi-dimensional state perception. Based on intelligent decision-making algorithms, sensor data sets are analyzed and processed. The positioning contact force is classified and evaluated using fuzzy logic rules, such as classifying the contact force state into normal or abnormal categories. Simultaneously, an adaptive PID control algorithm is applied to calculate the position compensation amount, dynamically optimizing control parameters based on real-time position deviation. Furthermore, a machine learning model predicts the wear state of the positioning system, inferring the system's aging trend using historical operating data. Thus, the classification evaluation results, position compensation amount, and wear state are comprehensively generated into intelligent decision-making instructions. These instructions control the marking device 2 to adjust parameters such as descent depth and movement speed, driving the working end to perform the marking operation along a preset trajectory. As a preferred implementation, the machine learning model can be specifically implemented as a random forest regression algorithm, with input features including the number of iterations and contact force fluctuation values ​​to quantify the degree of wear. During the marking process, a real-time feedback mechanism continuously monitors the operating status; if out-of-limit data is detected, an immediate correction process is triggered. Therefore, this method utilizes the unique geometric features of the asymmetric positioning groove to eliminate directional ambiguity. Combined with multi-source data fusion and a closed-loop decision-making mechanism, it effectively avoids initial positioning deviations caused by physical limitations, significantly improving marking positioning accuracy and operational efficiency, while ensuring the stability of marking quality under complex working conditions.

[0065] Example 2: Figure 4 A schematic diagram of an automated marking control device for motor housings according to an embodiment of the present invention is shown. The automated marking control device includes: Positioning module 10: It is used to transport the motor housing to be marked to the marking station, identify the asymmetrical distribution of positioning grooves in the marking area on the surface of the motor housing through a vision sensor, obtain reference information, and adjust the position and posture of the working end of the marking equipment according to the reference information to enter the marking working state.

[0066] Specifically, a servo conveyor belt is used to transport the motor housing to be marked to the positioning fixture at the marking station. The positioning fixture fixes the position of the motor housing through a pneumatic clamping mechanism. The belt conveyor mechanism driven by a stepper motor or servo motor is used to ensure that the motor housing arrives at the marking station stably and reduces position fluctuations during the transportation process.

[0067] Data acquisition module 20: used to collect in real time the contact force data, position data and image data of the working end of the marking device in the marking area of ​​the motor housing through a multi-sensor fusion system.

[0068] The force sensor collects the contact pressure data between the marking needle and the motor housing surface in real time, the displacement sensor collects the real-time position coordinate data of the working end, and the vision sensor collects the image data during the marking process. Furthermore, the working end of the marking equipment is equipped with a force sensor, which is a sensing device for measuring contact pressure. The force sensor can be a strain gauge sensor, a piezoelectric sensor, or a capacitive sensor, which can accurately capture the interaction force changes between the marking needle and the shell surface, providing basic data for force state assessment.

[0069] Furthermore, the marking equipment is equipped with a displacement sensor, which is a measuring device for detecting spatial position coordinates. It can be implemented using a laser displacement sensor, a grating ruler, or a rotary encoder to ensure that the position information of the working end in three-dimensional space is accurate and to support precise positioning operations.

[0070] A Kalman filter algorithm is used to perform time synchronization and noise filtering on the acquired contact pressure data, real-time position coordinate data, and image data, resulting in a denoised sensor data set. Standard Kalman filtering, extended Kalman filtering, or adaptive Kalman filtering are then employed to eliminate time skew in the multi-source data and suppress random noise, ensuring the reliability of the data set.

[0071] Intelligent Decision Module 30: It is used to classify and evaluate the positioning contact force through fuzzy logic rules, calculate the position compensation amount by applying an adaptive PID control algorithm, predict the wear state of the positioning system based on a machine learning model, and generate intelligent decision instructions by combining the classification and evaluation results, the position compensation amount and the wear state.

[0072] A fuzzy logic evaluation model is established, dividing the contact force data into several fuzzy subsets. The rate of change of the positioning contact force is used as an auxiliary input variable. Preset fuzzy rules are used to classify the positioning contact force state and output the evaluation result. The positioning contact force is classified and evaluated using fuzzy logic rules, dividing the contact force data into four fuzzy subsets: a low contact force set (contact force data in the range of 0-1N); a normal contact force set (contact force data in the range of 1-3N); a high contact force set (contact force data in the range of 3-5N); and an over-limit contact force set (contact force data in the range of 5N or more). The rate of change of the contact force data is used as an auxiliary input variable. Preset fuzzy rules, such as setting "contact force > 5N and rate of change > 2N / s" as the judgment rule for abnormal jamming, are applied. The judgment rule for determining normal contact is set as "contact force within the range of 1~3N and change rate <0.5N / s". By classifying the positioning contact force state, four evaluation results are output: "normal contact, slight over-force, severe over-force, and abnormal jamming". The positioning contact force is divided by fuzzy rule classification, thereby improving the operational reliability of the automatic marking control system.

[0073] Preset fuzzy rules refer to the pre-defined mapping logic from input variables to output categories. They can be stored in the rule base and executed through the fuzzy inference engine. The rules can be dynamically optimized based on historical marking data. The output evaluation result refers to the contact force status label after classification, which can be used to trigger corresponding control commands to ensure the stable operation of the marking process.

[0074] The application of adaptive PID control algorithm to calculate position compensation amount, predicting the wear state of positioning system based on machine learning model, and generating intelligent decision instructions by combining classification evaluation results, position compensation amount and wear state include: The deviation between the actual position and the reference position of the marking device's working end is used as the input data for the PID controller. The parameters of the PID controller are adjusted according to the input data. At the same time, the KI value is finely adjusted according to the contact force fluctuation fed back by the force sensor, and the output value of the PID controller is obtained. The position compensation amount of the marking device's working end is calculated based on the output value of the PID controller. Furthermore, the deviation between the actual position of the marking device's working end and the reference position is used as the input to the PID controller. Preset standard PID parameters are provided, where: KP=2.0, KI=0.5, KD=0.1. The deviation between the actual position of the marking device's working end and the reference position is detected. When the position deviation is <0.02mm, KP is adjusted to 1.6, which is 0.8 times the preset standard value of KP, and KD is adjusted to 1.2, which is 1.2 times the preset standard value of KD, to avoid overshoot. When the deviation value is between 0.02mm and 0.1mm, the standard parameter value is maintained. When the deviation value is greater than 0.1 mm, KP is adjusted to 3.0, which is 1.5 times the preset standard value of KP, and KD is adjusted to 0.7, which is 0.7 times the preset standard value of KD. At the same time, the KI value is finely adjusted according to the contact force fluctuation fed back by the force sensor. The adjustment range of the KI value is set between 0.3 and 0.7, which can effectively avoid integral saturation. Finally, the X / Y / Z axis position compensation amount of the working end is calculated according to the PID output value.

[0075] Coding module 40: Used to control the marking device to perform a coding operation on the motor housing according to the intelligent decision command.

[0076] By combining the visual sensor's identification of asymmetrically distributed positioning grooves with a multi-sensor fusion system using an intelligent decision-making mechanism, the system avoids directional misjudgment and initial positioning deviation caused by physical limitations. Furthermore, by dynamically evaluating the positioning contact force using fuzzy logic rules, calculating the position compensation amount using an adaptive PID control algorithm, and predicting the wear state of the positioning system using a machine learning model, the system achieves real-time optimization and correction of the marking process, thereby improving marking accuracy and efficiency.

[0077] Specifically, the unique spatial feature of the asymmetrically distributed positioning grooves eliminates the risk of misjudgment due to 180-degree rotation caused by symmetrical structures, ensuring the reliability of the reference information; the multi-sensor fusion system achieves spatiotemporal data synchronization and noise suppression through Kalman filtering, providing high-confidence input for decision-making; the intelligent decision-making module dynamically identifies the state based on the rate of change of contact force, adaptively adjusts PID parameters to match changes in working conditions, and predicts wear trends based on historical data, so that the execution parameters are closely matched with the real-time state, significantly reducing invalid operations and maintaining the stability of coding quality.

[0078] The automated marking control device for motor housing provided in this embodiment of the invention achieves physical alignment between the marking device and the motor housing by using the positioning groove of the marking area of ​​the motor housing in conjunction with the positioning protrusion of the marking device. It also improves the accuracy and reliability of the marking operation between the marking device and the motor housing by intelligently adjusting the marking operation between them, and improves the consistency of the marking operation.

[0079] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, which may include: read-only memory (ROM), random access memory (RAM), magnetic disk or optical disk, etc.

[0080] Furthermore, the above provides a detailed description of the automated marking control method and device for motor housing provided by the embodiments of the present invention. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. An automated marking control method for motor housings, characterized in that, The automated coding control method includes: The motor housing to be marked is transported to the marking station. The positioning grooves of the marking area on the surface of the motor housing are identified by a vision sensor to obtain reference information. The position and posture of the working end of the marking device are adjusted according to the reference information to enter the marking working state. The working end of the marking device is provided with positioning protrusions that match the positioning grooves. The multi-sensor fusion system collects in real time the positioning contact force data, position data and image data of the coding device working end in the coding area on the surface of the motor housing, and obtains a sensor data set. The positioning contact force data is classified and evaluated by fuzzy logic rules, the position compensation amount is calculated by adaptive PID control algorithm, the wear state of the positioning system is predicted based on machine learning model, and intelligent decision instructions are generated by combining the classification evaluation results, position compensation amount and wear state. The intelligent decision-making command controls the coding device to perform a coding operation on the motor housing; The application of the adaptive PID control algorithm to calculate the position compensation includes: The deviation between the actual position of the coding device's working end and the reference position is used as the input data for the PID controller. The parameters of the PID controller are adjusted according to the input data. At the same time, the KI value is finely adjusted according to the contact force fluctuation fed back by the force sensor, and the output value of the PID controller is obtained. The position compensation amount of the coding device's working end is calculated based on the output value of the PID controller. Fine-tuning the KI value refers to adjusting the integral coefficient according to the fluctuation range of the contact force, using proportional scaling or table lookup interpolation. The adjustment range of the KI value is 0.3~0.

7. The step of adjusting the pose state of the coding device's working end according to the reference information to enter the coding working state includes: The X-axis and Y-axis displacement compensation values ​​of the coding device working end are calculated based on the center coordinates in the reference information, and the rotation angle compensation value of the working end is calculated based on the normal direction. The position and posture of the coding device working end are adjusted according to the displacement compensation value and the rotation angle compensation value, so that the center of the coding needle at the working end is aligned with the origin of the reference coordinate system formed by the positioning groove, and a preparation signal for coding working state is generated. The method involves real-time acquisition of positioning contact force data, position data, and image data from the marking device's working end in the marking area on the motor housing surface using a multi-sensor fusion system, resulting in a sensor data set including: The contact pressure data between the coding needle and the motor housing surface is collected in real time by a force sensor, the real-time position coordinate data of the working end is collected by a displacement sensor, and the image data of the coding process is collected by a vision sensor. The Kalman filter algorithm is used to perform time synchronization and noise filtering on the collected contact pressure data, real-time position coordinate data and image data to obtain a noise-reduced sensor data set; The step of controlling the coding device to perform a coding operation on the motor housing according to intelligent decision-making instructions includes: The descent depth, moving speed, and coding force of the coding device's working end are set according to the intelligent decision-making instructions. The working end of the driving coding device moves along a preset trajectory on the coding area on the surface of the motor housing, while real-time feedback data of the coding operation is obtained through multiple sensor data. The system detects out-of-limit data on contact force or position deviation based on the real-time feedback data, and adjusts the coding operation accordingly.

2. The automated marking control method for motor housing as described in claim 1, characterized in that, The step of conveying the motor housing to be marked to the marking station and obtaining reference information by identifying asymmetrically distributed positioning grooves through a vision sensor includes: A servo conveyor belt is used to transport the motor housing to be marked to the positioning fixture at the marking station. The positioning fixture fixes the position of the motor housing by a pneumatic clamping mechanism. Using an industrial area array camera as a vision sensor, and a ring light source to illuminate the marking area on the surface of the motor housing, the motor housing to be marked is photographed to obtain image data of the motor housing to be marked. The image data of the motor housing to be marked is processed by image grayscale conversion, Gaussian filtering and edge detection algorithms to obtain the asymmetrical distribution of positioning groove contours of the marking area on the surface of the motor housing to be marked; The center coordinates, normal direction, and distance between two asymmetrically distributed positioning grooves are determined by template matching algorithm. The center coordinates, normal direction, and distance between the two grooves are stored and marked as reference information.

3. The automated marking control method for motor housing as described in claim 1, characterized in that, The classification and evaluation of positioning contact force data using fuzzy logic rules includes: A fuzzy logic evaluation model is established, which divides the positioning contact force data into several fuzzy subsets, uses the rate of change of the positioning contact force as an auxiliary input variable, and classifies the positioning contact force state in conjunction with preset fuzzy rules, and outputs the evaluation results.

4. The automated marking control method for motor housing as described in claim 1, characterized in that, The prediction of wear status of the positioning system based on the machine learning model includes: A random forest regression model is used as a machine learning model to collect historical operating data of the positioning system, extract wear-related features from the historical operating data, and construct a training dataset based on the wear-related features. The machine learning model is trained using a training dataset, and the trained machine learning model is then fed with current running data to predict the wear and tear of the positioning system. A wear compensation value is set according to the wear level, and the wear compensation value is incorporated into the intelligent decision-making command.

5. The automated marking control method for motor housing as described in claim 1, characterized in that, The automated coding control method also includes: The quality of the motor housing after the coding operation is completed is inspected using a visual sensor to obtain inspection data. Based on the detection data, the motor housings that have completed the coding operation are classified and cut into smaller parts.

6. An automated marking control device for motor housings, used to implement the automated marking control method for motor housings as described in any one of claims 1 to 5, characterized in that, The automated coding control device includes: Positioning module: Used to transport the motor housing to be marked to the marking station, identify the asymmetrically distributed positioning grooves on the marking area of ​​the motor housing surface through a vision sensor, obtain reference information, and adjust the position and posture of the working end of the marking device according to the reference information to enter the marking working state; Data acquisition module: used to collect in real time positioning contact force data, position data and image data of the coding device working end in the coding area on the surface of the motor housing through a multi-sensor fusion system; Intelligent decision module: It is used to classify and evaluate positioning contact force data through fuzzy logic rules, calculate position compensation amount by applying adaptive PID control algorithm, predict the wear state of positioning system based on machine learning model, and generate intelligent decision instructions by combining classification evaluation results, position compensation amount and wear state. The application of the adaptive PID control algorithm to calculate the position compensation includes: The deviation between the actual position of the coding device's working end and the reference position is used as the input data for the PID controller. The parameters of the PID controller are adjusted according to the input data. At the same time, the KI value is finely adjusted according to the contact force fluctuation fed back by the force sensor, and the output value of the PID controller is obtained. The position compensation amount of the coding device's working end is calculated based on the output value of the PID controller. Fine-tuning the KI value refers to adjusting the integral coefficient according to the fluctuation range of the contact force, using proportional scaling or table lookup interpolation. The adjustment range of the KI value is 0.3~0.

7. Coding module: Used to control the coding device to perform coding operation on the motor housing according to the intelligent decision command.

Images