Industrial robot calibration safety protection device and control method
By setting up a power control management unit and an electronic fence on an industrial robot, and combining the decision tree regression prediction model to monitor and control the current and voltage changes of the drive motor, the prevention and control problems of malfunctions during the calibration process of industrial robots are solved, and the safety and efficiency of the calibration process are improved.
Patent Information
- Application Number
- CN202510193290.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-02-21
- Publication Date
- 2025-04-25
AI Technical Summary
During the calibration process of industrial robots, misoperation may lead to the sudden start of the robot, and it is difficult for the prior art to effectively prevent and control such misoperation.
By setting up a power control management unit on an industrial robot, monitoring the current and voltage change gradient of the drive motor, and using the decision tree regression prediction model to judge and control malfunctions, and setting up an electronic fence to ensure safe operation.
It realizes accurate judgment and control of malfunctions in the calibration process of industrial robots, improves the safety and efficiency of the calibration process, and avoids the sudden start-up of the robot due to malfunctions.
Smart Images
Figure CN119871426A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of industrial robot calibration control, and more specifically, relates to an industrial robot calibration safety protection device and a control method. Background Art
[0002] Improving the data intelligence of the manufacturing industry, the development pace of industry-education integration is closely synchronized with industrial upgrading. In the field of industrial teaching, the focus is on cultivating students and engineers' proficiency in calibration technology. The calibration of industrial robots is not only limited to the learning process of learners, but also in industrial production, as production tasks continue to change, robots also need to be calibrated and adjusted in a timely manner.
[0003] In the application scenario of visual inspection of industrial robots, calibration is a crucial link. Regardless of the calibration tool instrument used, engineers need to operate and observe at close range, and perform demonstration operation and debugging at the same time. During the operation, due to the irregular debugging operation or interference of the control signal, the robot may be misoperated, causing it to start suddenly. At present, the cause of this phenomenon is still unclear (the possible reason may be communication delay and abnormal communication signal processing after multiple operations). It not only appears in teaching, but also occurs from time to time during the debugging process of the production line. In order to cope with this situation, the current protective measures include wearing safety helmets and quickly performing emergency stop operations when two-person operation malfunctions (but manual response requires high concentration at all times, which is difficult to ensure in actual operation).
[0004] In the current technology, safety protection devices are installed on industrial robots or machine vision is used to detect image recognition targets. One is the logical judgment and detection of danger signals, and the other is the visual detection method. The two methods carry out protection control from different angles; however, when the robot malfunctions, it generally does not report an error or issue a danger signal. Summary of the invention
[0005] The main purpose of this application is to propose an industrial robot calibration safety protection device and control method, which can use the changing gradient of the driving current to set different thresholds to achieve calibration smooth control.
[0006] In order to achieve the above-mentioned purpose, in a first aspect, the present application provides an industrial robot calibration safety protection device, comprising a power control management unit, wherein the power control management unit is arranged at a power supply end of the industrial robot so that the power enters the industrial robot through the power control management unit; When a person is detected entering the calibration area, the power control management unit switches to the driving motor state tracking mode to monitor the change gradient of the current and voltage of the driving motor of the industrial robot; If the change gradient is less than the preset gradient threshold, it is determined to be normal; if the change gradient is greater than the preset gradient threshold, it is determined to be a malfunction, and the power supply control management unit reduces the supply of driving current to the driving motor.
[0007] Furthermore, it also includes: Collect the working data of the driving motor of the industrial robot, Based on the working data, a decision tree regression prediction model is used for training, and the voltage and current of the power management unit are used as targets of the decision tree regression prediction model until a stop condition is met; wherein the stop condition includes that the variance of the data in the node is reduced to a sufficiently small value or reaches a preset maximum depth; According to the monitored current and voltage, the trained model is used to predict the working state of the drive motor and adjust the control parameters.
[0008] Furthermore, when a person is detected entering the calibration area, the method further includes: setting an electronic fence in the calibration area by a visual detection device, an infrared ranging device, a laser ranging device, or an electronic shielding device.
[0009] Furthermore, when the gradient threshold is set, it is satisfied that the current peak of the driving motor of the industrial robot is significant and the change gradient presents a step response form.
[0010] Furthermore, the power control management unit includes a power drive module and a power control system. The power drive module provides voltage and current according to the drive motor drive control, and the power control system performs real-time status detection on the voltage and current flowing through the drive motor when it is working.
[0011] In a second aspect, the present application also provides an industrial robot calibration safety protection control method, comprising the following steps: Set up an electronic fence in the designated area and detect intrusion of people in the designated area; When a person intrusion is detected, the mode is switched to the driving motor state tracking mode to monitor the change gradient of the current and voltage of the driving motor of the industrial robot; If the change gradient is less than the preset gradient threshold, it is determined to be normal; if the change gradient is greater than the preset gradient threshold, it is determined to be a malfunction, and the supply of driving current to the driving motor is reduced.
[0012] Furthermore, it also includes: Collect the working data of the driving motor of the industrial robot, Based on the working data, a decision tree regression prediction model is used for training, and the voltage and current of the power management unit are used as targets of the decision tree regression prediction model until a stop condition is met; wherein the stop condition includes that the variance of the data in the node is reduced to a sufficiently small value or reaches a preset maximum depth; According to the monitored current and voltage, the trained model is used to predict the working state of the drive motor and adjust the control parameters.
[0013] Furthermore, the method of setting up the electronic fence includes: setting up the electronic fence in the marked area by a visual detection device, an infrared ranging device, a laser ranging device, or an electronic shielding device.
[0014] Furthermore, when the gradient threshold is set, it is satisfied that the current peak of the driving motor of the industrial robot is significant and the change gradient presents a step response form.
[0015] Compared with the prior art, the present invention has the following beneficial effects: by intelligently switching the calibration scene control mode, monitoring the changing gradients of the current and voltage of the industrial robot's drive motor, and making false operation judgments, when the drive motor is not required to work at high load, the system will reduce the supply of drive current to achieve softer control, thereby improving the safety and efficiency of the calibration process. BRIEF DESCRIPTION OF THE DRAWINGS
[0016] Figure 1 Schematic diagram of the industrial robot calibration application provided for this implementation; Figure 2 Schematic diagram of the implementation process of the industrial robot calibration safety protection device provided for this implementation; Figure 3 Schematic diagram of the training process using the decision tree regression prediction model provided for this implementation Figure 4 A flow chart of the industrial robot calibration safety protection control method provided for this implementation.
[0017] In the picture: 1. Industrial robot; 2. Electronic fence. DETAILED DESCRIPTION
[0018] The embodiments of the present disclosure are described in detail below with reference to the accompanying drawings.
[0019] The following describes the embodiments of the present disclosure through specific examples, and those skilled in the art can easily understand other advantages and effects of the present disclosure from the contents disclosed in this specification. Obviously, the described embodiments are only a part of the embodiments of the present disclosure, rather than all of the embodiments. The present disclosure can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present disclosure. It should be noted that the following embodiments and features in the embodiments can be combined with each other without conflict. Based on the embodiments in the present disclosure, all other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of the present disclosure.
[0020] Improving the data intelligence of the manufacturing industry, the development pace of industry-education integration is closely synchronized with industrial upgrading. In the field of industrial teaching, the focus is on cultivating students and engineers' proficiency in calibration technology. The calibration of industrial robots is not only limited to the learning process of learners, but also in industrial production, as production tasks continue to change, robots also need to be calibrated and adjusted in a timely manner.
[0021] In the application scenario of visual inspection of industrial robots, calibration is a crucial link. Regardless of the calibration tool instrument used, engineers need to operate and observe at close range, and perform demonstration operation and debugging at the same time. During the operation, due to the irregular debugging operation or interference of the control signal, the robot may be misoperated, causing it to start suddenly. At present, the cause of this phenomenon is still unclear (the possible reason may be communication delay and abnormal communication signal processing after multiple operations). It not only appears in teaching, but also occurs from time to time during the debugging process of the production line. In order to cope with this situation, the current protective measures include wearing safety helmets and quickly performing emergency stop operations when two-person operation malfunctions (but manual response requires high concentration at all times, which is difficult to ensure in actual operation).
[0022] In the current technology, safety protection devices are installed on industrial robots or machine vision is used to detect image recognition targets. One is the logical judgment and detection of danger signals, and the other is the visual detection method. The two methods carry out protection control from different angles; however, when the robot malfunctions, it generally does not report an error or a danger signal. In addition, visual detection is difficult to balance in terms of cost, response speed, and ease of implementation.
[0023] In the current research field of industrial robot calibration technology, the description of calibration devices and methods in technical literature usually emphasizes its simplicity, accuracy and ease of implementation. Existing industrial robot protection devices and control methods, especially those systems that rely on danger signals or human-machine position relationships based on visual detection to build control logic, have put forward higher standards for detection accuracy and response speed.
[0024] During the calibration operation, due to the proximity between the operator and the robot and the rapid response of the robot's actions, the traditional logic of executing control measures after a dangerous state is detected not only prolongs the operation process but also increases the uncertainty during the operation.
[0025] To overcome this challenge, an embodiment of the present invention proposes an industrial robot calibration safety protection device and control method, which is particularly suitable for teaching and industrial demonstration applications.
[0026] Based on this, firstly, Figure 1 , Figure 2 As shown, this embodiment provides an industrial robot calibration safety protection device, the core components of which include a power control management unit, which is arranged at the power supply end of the industrial robot 1; the unit operates independently of the existing robot system, ensuring that the three-phase power first passes through the management unit, and then safely enters the power supply end of the robot system, realizing non-interference integration with the existing system, and allowing installation as an independent module. Among them, the power management unit includes a power drive and control system. After the three-phase electricity passes through the power drive module, it provides voltage and current according to the drive motor drive control. The power control system performs real-time status detection on the voltage and current flowing through the drive motor when it is working. There are many modes in the power drive control process, including input voltage, output voltage, load current, power line temperature, etc.
[0027] An electronic fence 2 is also set in the calibration area by various means such as an infrared distance measuring device, a laser distance measuring device or an electronic shielding device. The electronic fence 2 is designed to define the boundary of the movement safety area of the robot system, and can be installed in conjunction with a fixture to achieve accurate detection of the closed area.
[0028] After the robot is started, if the electronic fence 2 detects that a person has entered its set area, the system will automatically activate the drive motor tracking state control. This control mainly monitors the current and voltage in real time through the power management unit, that is, monitors the change gradient of the current and voltage of the drive motor of the industrial robot 1. If the change gradient is less than the preset gradient threshold, it is judged as normal; if the change gradient is greater than the preset gradient threshold, it is judged as a malfunction, and the power control management unit reduces the supply of the drive current of the drive motor.
[0029] This embodiment provides three current threshold intervals for different drive control characteristics of industrial robots 1: step current threshold, normal working threshold and boundary threshold interval. The step current threshold is the above-mentioned gradient threshold. When the gradient threshold is preset, the set gradient threshold should be able to satisfy that when it exceeds the threshold, it indicates that the industrial robot 1 is malfunctioning, that is, the current peak of the drive motor of the industrial robot 1 is significant, and the change gradient presents a step response form. Based on the physical principle of motor drive, when the drive motor needs to run at high speed and high torque, its drive control speed and torque increase accordingly, the drive current increases, and the average voltage rises. Therefore, the industrial robot 1 calibration safety protection device provided in this embodiment realizes precise control of the robot's action by intelligently switching the calibration scene control mode. When the electronic fence 2 detects that a person enters the calibration area, the system automatically switches to the driving motor state tracking mode. The system can track the working state of the driving motor in real time by monitoring the main control current and voltage in real time. In the case of malfunction of the robot caused by misoperation or signal delay, the system will be based on the changes in the drive motor current and voltage, and the bus current is gradient managed, and the control loop is added to achieve precise follow-up control of the drive motor. When the drive motor is not required to work at high load, the system will reduce the supply of drive current to achieve softer control, thereby improving the safety and efficiency of the calibration process.
[0030] In a possible embodiment, Figure 2 , Figure 3 As shown, a decision tree regression prediction model is introduced to collect, train and predict working data to achieve dynamic control and adjustment of threshold parameters. The following steps are included: S110, collecting working data of the driving motor of the industrial robot.
[0031] The collected working data include, but are not limited to: input voltage, output voltage, output current, load type, ambient temperature, and power efficiency.
[0032] S120. Based on the working data, a decision tree regression prediction model is used for training, and the voltage and current of the power management unit are used as targets of the decision tree regression prediction model until a stopping condition is met; wherein the stopping condition includes that the variance of the data within the node is reduced to a sufficiently small value or reaches a preset maximum depth.
[0033] The model training process includes: Data preprocessing Clean data: remove or fill missing values and handle outliers.
[0034] Feature engineering: may include feature selection, feature construction, etc.
[0035] Data standardization: Make the data have a uniform scale to facilitate model processing.
[0036] Divide the dataset The data is divided into training set, validation set and test set. Usually 70% of the data is used as training set, 15% as validation set and 15% as test set.
[0037] Building a decision tree model Build a decision tree model using the training dataset. You can use the DecisionTreeRegressor class in Python's scikit-learn library.
[0038] Code example: from sklearn.tree import DecisionTreeRegressor from sklearn.model_selection import train_test_split # Assume X is the feature matrix and y is the target variable X_train,X_test,y_train,y_test=train_test_split(X,y, test_size=0.3,random_state=42) # Initialize the decision tree regression model tree_reg=DecisionTreeRegressor(random_state=42) # Train the model tree_reg.fit(X_train, y_train) Model Evaluation The performance of the model can be evaluated using the validation set. Indicators such as mean square error (MSE) and root mean square error (RMSE) can be used.
[0039] from sklearn.metrics import mean_squared_error # Predict validation set y_pred = tree_reg.predict(X_val) # Calculate MSE mse = mean_squared_error(y_val, y_pred) print(f"Mean Squared Error: {mse}") Model pruning Based on the model evaluation results, pruning is performed to prevent overfitting.
[0040] # Set the maximum depth and other parameters for pruning tree_reg_pruned=DecisionTreeRegressor(max_depth=5, random_state=42) tree_reg_pruned.fit(X_train, y_train) 8. Model Optimization Use methods such as Grid Search to optimize model parameters.
[0041] from sklearn.model_selection import GridSearchCV # Parameter grid param_grid = {'max_depth': [None, 5, 10], 'min_samples_split': [2, 5,10]} # Grid Search grid_search=GridSearchCV(estimator=tree_reg, param_grid=param_grid,cv=5, scoring='neg_mean_squared_error') grid_search.fit(X_train, y_train) # Optimal parameters print(f"Best parameters: {grid_search.best_params_}") Model deployment Deploy the trained model into the power drive control system to achieve real-time prediction and control.
[0042] The model's splitting criteria are usually based on information gain, Gini impurity, etc. to maximize the accuracy of the prediction. The prediction process involves new input data being predicted through the tree structure until it reaches a leaf node that provides a predicted value. The high interpretability of the decision tree model makes its decision process easy to understand and explain, and it is applicable to a variety of data sets including classification and regression problems.
[0043] To prevent overfitting, it may be necessary to prune the decision tree or use an ensemble method such as random forest or gradient boosting tree to improve the predictive performance and robustness of the model. The evaluation of model performance is usually based on indicators such as mean square error (MSE), root mean square error (RMSE) or coefficient of determination (R²).
[0044] Constructing a fully grown tree: First, let the decision tree grow fully without considering pruning until every leaf node is pure or cannot be split further.
[0045] Evaluate model complexity: Use some complexity indicators, such as tree depth, number of nodes, etc., to evaluate the complexity of the model.
[0046] Calculate model error: Use a validation set or cross-validation method to calculate the error of the model.
[0047] Pruning strategy selection: Pre-pruning: During the tree growth process, if the splitting of a node cannot significantly reduce the error or does not meet a certain threshold, stop splitting.
[0048] Post-pruning: First build a complete tree, then prune from the bottom, gradually removing nodes that have little impact on model performance.
[0049] Implement pruning: For pre-pruning, set a threshold, and stop splitting when the information gain or other splitting criteria brought by the split is lower than this threshold.
[0050] For post-pruning, start from the bottom of the tree and remove nodes one by one. After each removal, evaluate the model performance and keep the pruned result if the performance does not drop significantly.
[0051] Evaluate pruning effect: Re-evaluate the error on the pruned model to ensure that the pruned model has reduced complexity while maintaining sufficient generalization ability.
[0052] Adjust pruning parameters: Adjust pruning parameters according to the pruning effect, such as information gain threshold, maximum depth, minimum number of samples, etc.
[0053] Repeat the pruning process: Multiple pruning and evaluation may be required to find the best pruning point.
[0054] Select the best model: Select a model from the pruning process that performs best on the validation set.
[0055] In addition, this embodiment is not limited to the decision tree regression prediction model used. Depending on the actual application conditions, other machine learning prediction models or fuzzy control, synovial control, neural network prediction and other methods in modern control theory can also be used to achieve data-driven prediction purposes.
[0056] S130: According to the monitored current and voltage, use the trained model to predict the working state of the drive motor and adjust the control parameters.
[0057] In the second aspect, based on the above embodiments, Figure 4 As shown, this embodiment also provides an industrial robot calibration safety protection control method, comprising the following steps: S210: Setting up an electronic fence in the designated area and detecting the intrusion of people in the designated area.
[0058] S220: When a person intrusion is detected, switch to a driving motor state tracking mode to monitor the change gradient of the current and voltage of the driving motor of the industrial robot.
[0059] S230: If the change gradient is less than a preset gradient threshold, it is determined to be normal; if the change gradient is greater than the preset gradient threshold, it is determined to be a malfunction, and the supply of driving current to the driving motor is reduced.
[0060] Furthermore, it also includes: Collect the working data of the driving motor of the industrial robot, Based on the working data, a decision tree regression prediction model is used for training, and the voltage and current of the power management unit are used as targets of the decision tree regression prediction model until a stop condition is met; wherein the stop condition includes that the variance of the data in the node is reduced to a sufficiently small value or reaches a preset maximum depth; According to the monitored current and voltage, the trained model is used to predict the working state of the drive motor and adjust the control parameters.
[0061] Furthermore, the method of setting up the electronic fence includes: setting up the electronic fence in the marked area by a visual detection device, an infrared ranging device, a laser ranging device, or an electronic shielding device.
[0062] Furthermore, when the gradient threshold is set, it is satisfied that the current peak of the driving motor of the industrial robot is significant and the change gradient presents a step response form.
[0063] The above is only for explaining the implementation mode of the present invention and is not intended to limit the present invention. For those skilled in the art, any modification, equivalent substitution, improvement, etc. made within the spirit and principle of the present invention without creative work should be included in the protection scope of the present invention.
Claims
1. An industrial robot calibration safety protection device, characterized in that: It includes a power control management unit, which is arranged at the power supply end of the industrial robot so that the power enters the industrial robot through the power control management unit; When a person is detected entering the calibration area, the power control management unit switches to the driving motor state tracking mode to monitor the change gradient of the current and voltage of the driving motor of the industrial robot; If the change gradient is less than the preset gradient threshold, it is judged to be normal; If the change gradient is greater than a preset gradient threshold, it is determined to be a malfunction, and the power control management unit reduces the supply of the driving current to the driving motor.
2. The industrial robot calibration safety protection device according to claim 1, characterized in that: Also includes: Collect the working data of the driving motor of the industrial robot, Based on the working data, a decision tree regression prediction model is used for training, and the voltage and current of the power management unit are used as targets of the decision tree regression prediction model until a stop condition is met; wherein the stop condition includes that the variance of the data in the node is reduced to a sufficiently small value or reaches a preset maximum depth; According to the monitored current and voltage, the trained model is used to predict the working state of the drive motor and adjust the control parameters.
3. The industrial robot calibration safety protection device according to claim 1, characterized in that: When a person is detected to enter the calibration area, the method further includes: setting an electronic fence in the calibration area by using a visual detection device, an infrared distance measuring device, a laser distance measuring device, or an electronic shielding device.
4. The industrial robot calibration safety protection device according to claim 1, characterized in that: When the gradient threshold is set, it is satisfied that the current peak value of the driving motor of the industrial robot is significant and the change gradient presents a step response form.
5. The industrial robot calibration safety protection device according to claim 1, characterized in that: The power control management unit includes a power drive module and a power control system. The power drive module provides voltage and current according to the drive motor drive control, and the power control system performs real-time status detection on the voltage and current flowing through the drive motor when it is working.
6. An industrial robot calibration safety protection control method, characterized in that: The following steps are involved: Set up an electronic fence in the designated area and detect intrusion of people in the designated area; When a person intrusion is detected, the mode is switched to the driving motor state tracking mode to monitor the change gradient of the current and voltage of the driving motor of the industrial robot; If the change gradient is less than the preset gradient threshold, it is judged to be normal; If the change gradient is greater than a preset gradient threshold, it is determined to be a malfunction, and the supply of drive current to the drive motor is reduced.
7. The industrial robot calibration safety protection control method according to claim 6, characterized in that: Also includes: Collect the working data of the driving motor of the industrial robot, Based on the working data, a decision tree regression prediction model is used for training, and the voltage and current of the power management unit are used as targets of the decision tree regression prediction model until a stop condition is met; wherein the stop condition includes that the variance of the data in the node is reduced to a sufficiently small value or reaches a preset maximum depth; According to the monitored current and voltage, the trained model is used to predict the working state of the drive motor and adjust the control parameters.
8. The industrial robot calibration safety protection control method according to claim 6, characterized in that: The method of setting up the electronic fence includes: setting up the electronic fence in the marked area by a visual detection device, an infrared distance measuring device, a laser distance measuring device, or an electronic shielding device.
9. The industrial robot calibration safety protection device according to claim 6, characterized in that: When the gradient threshold is set, it is satisfied that the current peak value of the driving motor of the industrial robot is significant and the change gradient presents a step response form.