Steel wire bundling manipulator wire feeding speed compensation and precision calibration system
By installing sensors on the wire binding robot to acquire data and performing hierarchical dynamic compensation, the problem of difficulty in quantifying the coupling relationship of error factors in traditional methods is solved, realizing high-precision adaptive binding of the wire binding robot and improving binding quality and stability.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-03-12
- Publication Date
- 2026-05-01
AI Technical Summary
Traditional wire binding robots cannot effectively quantify the dynamic coupling relationship of multiple error factors, making it difficult to accurately suppress the superposition effect of compound errors, resulting in binding defects such as excessive positioning deviation of binding points and uneven wrapping tightness.
Multiple sensors are installed at intervals along the wire conveying path of the wire binding robot to acquire data on wire tension, displacement, robot joint coordinates, and vibration data of the binding action. The data are then analyzed by a processing terminal to achieve hierarchical dynamic compensation and accurately superimpose compensation for multiple error factors.
It effectively solves the problem of delayed response of traditional methods to the superposition effect of compound errors, significantly suppresses the risk of deformation accumulation and resonance amplification, and improves the operational control stability of the wire binding robot.
Smart Images
Figure CN121947848A_ABST
Abstract
Description
A wire feeding speed compensation and accuracy calibration system for a wire bundling robot Technical Field
[0001] This invention relates to the technical field of robotic arm control, specifically to a wire feeding speed compensation and accuracy calibration system for a wire binding robotic arm. Background Technology
[0002] Wire binding robots have been widely adopted in large-scale production scenarios such as automobile frame welding and fixing, bridge steel component splicing, and container body assembly. Wire binding robots have achieved high efficiency and standardization in the binding process. The wire feeding speed, as a core parameter, determines the continuity of wire feeding and the stability of tension, while the wire feeding accuracy is directly related to the alignment of the binding points and the consistency of wire winding. Together, these two factors constitute the core indicators of the robot's binding performance.
[0003] In the wire feeding operation of wire bundling robots, traditional error compensation methods cannot effectively quantify the dynamic coupling relationship of multiple error factors, making it difficult to accurately suppress the superposition effect of compound errors. This results in bundling defects such as excessive positioning deviation of bundling points and uneven winding tightness. This defect has become a technical bottleneck restricting high-precision adaptive bundling operations. Summary of the Invention
[0004] To address the technical problem of accurately compensating for multiple error factors in a wire binding robot, this invention aims to provide a wire feeding speed compensation and accuracy calibration system for a wire binding robot. The specific technical solution is as follows: This invention provides a wire feeding speed compensation and accuracy calibration system for a wire binding robot. The system includes: multiple sensors installed at intervals along the wire feeding path of the wire binding robot; multiple sensors are used to acquire and output data on wire tension, wire displacement, coordinate data of each joint on the robot, and vibration data of the binding action; a processing terminal connected to each sensor; the processing terminal is used to obtain the cumulative deviation of wire feeding based on wire tension and wire displacement; obtain the error superposition degree of the robot joint movement based on the coordinate data of each joint; obtain the risk coefficient of wire feeding resonance caused by the robot performing the binding action based on the vibration data; and perform wire feeding superposition error analysis based on the cumulative deviation, error superposition degree, and risk coefficient, and implement graded dynamic compensation for the robot.
[0005] In one optional embodiment, the cumulative deviation of wire conveying is obtained based on wire tension and wire displacement, including: obtaining the rate of change of tension and the rate of change of displacement of the wire during conveying for at least two adjacent tracking cycles based on wire tension and wire displacement; obtaining a rate of change deviation coefficient characterizing the deformation of the wire conveying based on the rate of change of tension and the rate of change of displacement; and obtaining the cumulative deviation of wire conveying based on the cumulative results characterized by multiple consecutive rate of change deviation coefficients within a preset time window.
[0006] In one optional embodiment, the rate of change deviation coefficient characterizing the deformation of the steel wire conveying is obtained based on the rate of change of tension and the rate of change of displacement, including: obtaining the rate of change difference based on the difference between the rate of change of tension and the rate of change of displacement, and configuring the maximum absolute value of the rate of change of tension and the rate of change of displacement as the absolute extreme value; and determining the ratio of the rate of change difference to the absolute extreme value as the rate of change deviation coefficient.
[0007] In one optional embodiment, the error superposition degree of the robot joint motion is obtained based on the coordinate data of each joint, including: obtaining the pose switching disorder degree of the robot based on the change characteristics of the coordinate data of each joint in the time dimension and the motion direction dimension; obtaining the joint switching scale of the robot based on the change characteristics of the coordinate data of each joint in the spatial dimension; and obtaining the error superposition degree of the robot joint motion based on the pose switching disorder degree and the joint switching scale.
[0008] In one optional embodiment, the pose switching disorder of the robot is obtained based on the change characteristics of the coordinate data of each joint in the time dimension and the motion direction dimension. This includes: obtaining the coordinate change amount corresponding to each joint based on the coordinate data difference of at least two adjacent tracking cycles; marking each coordinate change amount as a target change amount when it is greater than a corresponding set change amount threshold; statistically analyzing the target change amounts to obtain the time overlap of all joints within a preset time window; obtaining the direction switching rate of all joints based on the number of direction adjustments represented by the coordinate data of each joint within the preset time window; and obtaining the pose switching disorder of the robot based on the time overlap and the direction switching rate.
[0009] In one optional embodiment, the joint switching scale of the robot is obtained based on the spatial dimension variation characteristics of the coordinate data of each joint, including: obtaining the angle between the direction vectors of each joint in two adjacent tracking cycles based on the coordinate data of each joint within a preset time window; and obtaining the joint switching scale of the robot based on the angle between the direction vectors of all joints and the cumulative window time, wherein the cumulative window time is the product of the total number of joints and the number of time windows.
[0010] In one optional embodiment, the risk coefficient of wire conveying resonance caused by the robot performing the binding action is obtained based on vibration data, including: performing resonance characteristic analysis of the wire initiation conveying based on vibration data to obtain the initiation resonance ratio; and obtaining the risk coefficient based on the initiation resonance ratio, the standard deviation of the wire feeding speed, the rising slope of the vibration data in the synchronization time window, and the mean square error of the frequency.
[0011] In one optional embodiment, a risk coefficient is obtained based on the starting resonance ratio, the standard deviation of the wire feed speed, the rising slope of the vibration data within the synchronization time window, and the mean square error of the frequency. This includes: obtaining a first vibration interference coefficient based on the product of the starting resonance ratio and the rising slope; obtaining a second vibration interference coefficient based on the product of the standard deviation of the wire feed speed and the mean square error of the frequency; and obtaining a risk coefficient based on the ratio of the first vibration interference coefficient to the second vibration interference coefficient.
[0012] In one optional embodiment, a superimposed error analysis of wire conveying is performed based on the cumulative deviation, the superposition of errors, and the risk coefficient, and graded dynamic compensation is implemented for the robot arm, including: obtaining a first error transmission coefficient based on the change in the superposition of errors over multiple preset time windows and the change in wire tension over corresponding time windows; obtaining a second error transmission coefficient based on the change in the cumulative deviation over multiple preset time windows and the standard deviation of wire tension over corresponding time windows; obtaining an error compensation coefficient based on the resonance amplification result of wire conveying characterized by the first error transmission coefficient, the second error transmission coefficient, and the risk coefficient over the corresponding time windows; obtaining a compensation speed for wire conveying based on the error compensation coefficient, a preset graded control constant, and the reference speed of wire conveying; and adjusting the wire conveying speed of the robot arm in the next time window based on the compensation speed.
[0013] In one optional embodiment, the compensation speed of the wire conveying is obtained based on the error compensation coefficient, a preset graded control constant, and the reference speed of the wire conveying, including: when the error compensation coefficient is less than a preset coefficient range, determining the graded control constant as a first constant value; when the error compensation coefficient is within the preset coefficient range, determining the graded control constant as a second constant value, the second constant value being greater than the first constant value; when the error compensation coefficient is greater than the preset coefficient range, determining the graded control constant as a third constant value, the third constant value being greater than the second constant value; obtaining the graded compensation coefficient based on the constant value determined by the graded control constant and the error compensation coefficient; and obtaining the compensation speed of the wire conveying based on the product of the reference speed and the graded compensation coefficient.
[0014] The present invention has the following beneficial effects: The technical solution of the present invention includes multiple sensors and a processing terminal connected to each sensor. The multiple sensors are installed at intervals along the wire conveying path of the wire binding robot. The multiple sensors are used to acquire wire tension, wire displacement, coordinate data of each joint on the robot, and vibration data of the binding action, and output these data to the processing terminal. The processing terminal is used to obtain the cumulative deviation of wire conveying based on wire tension and wire displacement; to obtain the error superposition degree of the robot's joint movement based on the coordinate data of each joint; to obtain the risk coefficient of wire conveying resonance caused by the robot performing the binding action based on the vibration data; and to perform wire conveying superposition error analysis based on the cumulative deviation, error superposition degree, and risk coefficient, and to implement graded dynamic compensation for the robot. This technical solution accurately superimposes and compensates for multiple error factors of the wire binding robot from three dimensions: the state of the wire itself, the motion state of the mechanism, and the end-effector's working state. This effectively solves the problem of delayed response to the superposition effect of compound errors in traditional methods. While ensuring wire feeding stability, it significantly suppresses the risk of deformation accumulation and resonance amplification, and improves the stability of the wire binding robot's operation control. Attached Figure Description
[0015] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0016] Figure 1 is a structural schematic diagram of a wire feeding speed compensation and accuracy calibration system for a wire bundling robot provided in an embodiment of the present invention; Figure 2 is a schematic diagram of the spaced installation of multiple sensors along the wire conveying path provided in an embodiment of the present invention; Figure 3 is a flowchart of the control of the wire bundling robot implemented by the processing terminal provided in an embodiment of the present invention.
[0017] Explanation of reference numerals in the attached diagram: 1-Tension sensor, 2-Displacement sensor, 3-Position sensor, 4-Vibration sensor, 5-Processing terminal, 6-Steel wire, 7-Binding mechanism. Detailed Implementation
[0018] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a wire feeding speed compensation and accuracy calibration system for a wire bundling robot according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.
[0019] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.
[0020] Currently, in robotic wire feeding and binding operations, it is impossible to effectively quantify the dynamic coupling relationship between path deviation, tension fluctuation, and end-effector vibration. For example, the triggering efficiency between wire feeding path deviation caused by joint posture adjustment and wire tension fluctuation lacks dynamic evaluation, and the transmission rate of tension fluctuation to deformation accumulation has not been quantitatively modeled. Furthermore, the resonant coupling between end-effector vibration and wire feeding speed switching further amplifies errors, ultimately causing binding process defects such as excessive binding point positioning deviation and uneven wrapping tightness. The following, in conjunction with the accompanying drawings, details the specific technical solution of the wire feeding speed compensation and accuracy calibration system for a robotic wire binding arm provided by this invention.
[0021] Please refer to Figure 1, which is a schematic diagram of a wire-binding robot's wire feeding speed compensation and accuracy calibration system according to an embodiment of the present invention. This calibration system can be applied to wire-binding mechanical tools, equipment, etc., to implement calibration control of the wire-binding robot and improve the quality and stability of wire binding. The calibration system includes multiple sensors and a processing terminal connected to each sensor.
[0022] Multiple sensors are installed at intervals along the wire feeding path of the wire binding robot. These sensors acquire data on wire tension, wire displacement, coordinates of each joint on the robot, and vibration data of the binding action, and output these data. See Figure 2, which illustrates the interval installation of multiple sensors along the wire feeding path. During wire feeding and binding, the wire needs to be fed in a straight line and then bent at least once for binding. The V-direction in Figure 2 represents the bending path of the wire. The multiple sensors include tension and displacement sensors deployed along the wire feeding path of the robot's wire feeding mechanism, at least one pose sensor deployed at the robot joints, and at least one vibration sensor deployed at the end of the binding mechanism. Each sensor synchronously acquires real-time wire tension, feed displacement, real-time robot pose coordinates (X, Y, Z), and end-effector vibration of the binding mechanism. The acquired data is transmitted to the processing terminal at a frequency of 10 ms / time to form a real-time operating parameter dataset. Multiple sensors enable dynamic parameter sensing of the wire binding robot under multi-dimensional working conditions, allowing for the superposition and compensation of multiple error factors.
[0023] It is understandable that factors affecting wire feeding accuracy include the state of the wire itself, the motion state of the wire feeding mechanism, and the end-effector state of the binding mechanism. During wire feeding, the wire undergoes elastic deformation under tension. The greater the tension and the longer the duration, the more pronounced the deformation. This deformation directly leads to a deviation between the actual effective feed length and the mechanical displacement length collected by the sensor. Therefore, the error corresponding to the wire's state can be determined through real-time synchronized data from the tension and displacement sensors. Similarly, when the robotic arm joints adjust their posture, the wire feeding mechanism synchronously changes its spatial position, causing the theoretical straight path of the wire feed to deviate from the preset trajectory, resulting in path offset error. The greater the posture adjustment amplitude and the more joint rotation angles, the more significant the cumulative effect of the path offset. Therefore, it is necessary to deploy posture sensors at the robotic arm joints to collect the coordinate data of the corresponding joints. Furthermore, the vibration at the end of the binding mechanism will cause instantaneous swaying at the wire exit position. The greater the vibration amplitude and the closer the frequency is to the resonant frequency band corresponding to the wire feeding speed, the more obvious the instantaneous deviation of the wire feed will be, thereby destroying the uniformity of winding. Therefore, it is equally important to collect vibration data of the binding action to compensate for the wire feeding speed.
[0024] The following section details how to implement hierarchical dynamic compensation based on data collected from multiple sensors. The data collected by each sensor is processed through a processing terminal, and corresponding controls are implemented. The processing terminal can be a microcontroller or a PLC (Programmable Logic Controller), or other terminal devices or components capable of data processing and terminal control. Please refer to Figure 3, which is a flowchart of the processing control of the processing terminal, specifically including: S11, acquiring the wire tension, wire displacement, coordinate data of each joint on the robotic arm, and vibration data of the binding action collected by multiple sensors.
[0025] Specifically, after each sensor collects data, it sends it to the processing terminal, enabling the terminal to acquire various types of data. After acquiring the data, it can be filtered to ensure accuracy. Wire tension and wire displacement represent the tension and displacement of the wire during transport, respectively. Data can be acquired based on a preset tracking cycle, such as once every 10ms. The joints on the wire binding robot include binding joints that rotate in a preset direction; some robots also have hinged joints. Coordinate data represents the rotation of the joints on the robot. Vibration occurs during the wire binding process; vibration data represents the magnitude of the vibration.
[0026] It should be noted that, for ease of calculation, all indicator data involved in the calculation in this embodiment of the invention have undergone data preprocessing to eliminate the influence of dimensions. The specific methods for eliminating the influence of dimensions are well known to those skilled in the art and are not limited here.
[0027] S12. Obtain the cumulative deviation of wire conveying based on wire tension and wire displacement; obtain the error superposition of robot joint movement based on the coordinate data of each joint; obtain the risk coefficient of wire conveying resonance caused by the robot performing the binding action based on vibration data.
[0028] Specifically, calculation models can be set based on the calculation requirements of deviation accumulation, error superposition, and risk coefficient, and the corresponding data can be input into the calculation models to obtain the corresponding results. Deviation accumulation reflects the degree of deviation accumulation between the actual effective feed length and the sensor-detected displacement caused by elastic deformation of the steel wire due to tension during the conveying process. The larger this index, the more the steel wire is stretched, the lower the feed accuracy, and the more likely it is to cause misalignment of the binding point or uneven winding tightness. Error superposition reflects the error accumulation caused by the compound offset and disordered switching of the spatial path when the robot's multiple joints adjust their posture. The larger this index, the more frequent the joint movements, the larger the directional switching amplitude, the more aggravated the path offset, and thus amplify the wire feeding position error. Risk coefficient reflects the resonance coupling risk between end-effector vibration and wire feeding action, including the synchronization between the moment of wire feeding start and the vibration peak, and the closeness of the vibration frequency to the wire feeding speed frequency. The larger this index, the higher the possibility of resonance between vibration and wire feeding, resulting in instantaneous shaking at the wire feeding exit and damage to winding stability.
[0029] For example, step S12 includes sub-steps S12-1 to S12-8. Specifically, in one embodiment, the cumulative deviation of the wire conveying is obtained based on the wire tension and wire displacement, including: S12-1, obtaining the tension change rate and displacement change rate of the wire during conveying for at least two adjacent tracking cycles based on the wire tension and wire displacement. The tracking cycle can be configured with a minimum tracking unit of 10ms, continuously recording the tension change rate and displacement change rate within two adjacent units. That is, the difference between the tension / displacement of the target unit and the previous adjacent unit, and then the ratio of this difference to the tension / displacement of the previous adjacent unit, are used to obtain the tension change rate and displacement change rate. Real-time synchronized data from the tension sensor and displacement sensor are used to focus on monitoring the relative changes in the tension change rate and displacement change rate. It should be noted that if the displacement or tension in the previous cycle is 0, the change rate is recorded as the current value.
[0030] S12-2. Based on the rate of change of tension and the rate of change of displacement, obtain the deviation coefficient of the rate of change of steel wire conveying deformation. A positive deviation coefficient, and the larger the value, the faster the tension changes compared to the displacement, indicating that the rate of change of tension is accelerating while the rate of change of displacement is not increasing synchronously or is even decreasing. This directly indicates that the effective feed deviation caused by the deformation of the steel wire is expanding. At this time, the rate of increase of tension exceeds the rate of increase of displacement, the risk of the steel wire being overstretched is higher, and the actual effective feed length is less than the value collected by the displacement sensor.
[0031] The calculation of the rate of change deviation coefficient includes: obtaining the rate of change difference based on the difference between the tension rate of change and the displacement rate of change, and setting the maximum absolute value of the tension rate of change and the displacement rate of change as the absolute extreme value; the ratio of the rate of change difference to the absolute extreme value is determined as the rate of change deviation coefficient. It can be understood that a positive rate of change deviation coefficient indicates that the tension rate of change is greater than the displacement rate of change, meaning the force on the steel wire increases faster than the wire feeding displacement, indicating that the steel wire is in a stretched state, and the effective feed length is shorter than the value detected by the displacement sensor. Conversely, a negative rate of change deviation coefficient indicates that the tension rate of change is less than the displacement rate of change, the steel wire deformation is easing, and the effective wire feeding length is close to or exceeds the sensor displacement.
[0032] S12-3. Obtain the cumulative deviation of the wire feeding based on the cumulative results represented by multiple continuous rate-of-change deviation coefficients within a preset time window. The preset time window can be set to 5 tracking cycles or other numbers of tracking cycles. Taking 5 consecutive tracking cycles (i.e., 50ms) as a preset time window for calculation as an example, count the number of units with positive tension-displacement rate-of-change deviation coefficients within the calculation window; add up all the positive tension-displacement rate-of-change deviation coefficients within the window, and then divide by the number of time units included in the window, 5. The average value obtained is the cumulative deviation, denoted as P. The larger the value of this index, the higher the cumulative degree of the wire being continuously in an overstretched state within 50ms, and the greater the effective feed deviation; conversely, the smaller the value, the smaller the effective feed deviation.
[0033] The calculation process for error superposition is explained in detail below, specifically including: S12-4, obtaining the pose switching disorder degree of the robot arm based on the change characteristics of the coordinate data of each joint in the time and motion direction dimensions. The pose switching disorder degree determines the synchronicity of joint movements and the degree of disorder in direction switching during wire feeding. A larger pose switching disorder degree indicates more overlapping joint movements, more frequent direction switching, and more unstable path adjustment, leading to a higher risk of error superposition and wire feeding trajectory deviation.
[0034] The first step is to obtain the coordinate change of each joint based on the coordinate data difference between at least two adjacent tracking cycles. The start and end times of adjustment for each joint are recorded in real time. If two or more joints are simultaneously in adjustment within the same 10ms time unit (i.e., the action is not yet finished), it is determined that multi-joint linkage adjustment is occurring, increasing the risk of cumulative wire feeding path deviation. This is because synchronous multi-joint actions easily lead to compound deviations in the spatial position of the wire feeding mechanism. The corresponding coordinate change is obtained based on the coordinate data difference between two adjacent tracking cycles for each joint.
[0035] The second step involves marking each coordinate change exceeding a corresponding set threshold as a target change. These target changes are then statistically analyzed to obtain the time overlap of all joints within a preset time window. A minimum motion threshold (e.g., 0.01mm) is set for each joint of the robotic arm, denoted as the change threshold. For each 10ms time unit, the spatial coordinate change (using Euclidean distance) between the joint in that time unit and the previous adjacent time unit is calculated. If the coordinate change exceeds the set threshold, the joint is in an adjustment state within that time unit and is marked as 1; otherwise, it is marked as 0. The marked values of all joints within that time unit are then summed and divided by the number of joints to obtain the joint adjustment time overlap. This index ranges from 0 to 1; the closer the value is to 1, the more joints are being adjusted simultaneously, and the greater the risk of path composite offset caused by multi-joint linkage. The time overlap of all joints within five time units is averaged, and the time overlap of all joints within the preset time window is denoted as C1.
[0036] The third step is to obtain the direction switching rate of all joints based on the number of direction adjustments represented by the coordinate data of each joint within a preset time window. The adjustment direction (e.g., positive and negative X-axis) of a single joint is continuously tracked within a calculated preset time window (within 5 consecutive time units). The number of adjustments is divided by 5 to obtain the direction switching rate. A higher value indicates that the joint frequently alternates between positive and negative directions. The deviation gradually accumulates with the number of adjustments. The direction switching rates of all joints are summed and averaged to obtain the direction switching rate of all joints, denoted as U.
[0037] The fourth step is to obtain the pose switching disorder degree of the robot arm based on the time overlap degree and the orientation switching rate. The product of the time overlap degree C1 and the orientation switching rate U is recorded as the pose switching disorder degree within the calculation window, and is denoted as the pose switching disorder degree M.
[0038] S12-5. Based on the spatial variation characteristics of the coordinate data of each joint, the joint switching scale of the robot is obtained. The joint switching scale characterizes the magnitude of the change in the robot's motion direction in the spatial dimension, reflecting the degree of inflection and the magnitude of inertia in pose adjustment. The angle between the direction vectors of each joint in two adjacent tracking cycles can be obtained based on the coordinate data of each joint within a preset time window; then, based on the angle between the direction vectors of all joints and the cumulative window time, the joint switching scale of the robot is obtained, where the cumulative window time is the product of the total number of joints and the number of time windows.
[0039] Specifically, based on coordinate data, the angle between direction vectors within every two consecutive time units is calculated within the same time window. The angle is obtained by dividing the dot product of the two direction vectors by the product of their magnitudes and then taking the inverse cosine. The angles between the direction vectors of all joints in all time units within the calculation window are added together and then divided by the product of the total number of joints and the number of window units. The result is the joint switching scale. The larger the value of this index, the larger the joint adjustment direction switching scale, the greater the pose adjustment inertia, which is more likely to cause sudden changes or fluctuations in the path, and errors are unavoidable. Conversely, the smaller the value, the smaller the pose adjustment inertia. The joint switching scale is denoted as H.
[0040] S12-6. Based on the pose switching disorder and joint switching scale, obtain the error superposition degree of the robot's joint motion. Multiply the pose switching disorder degree M and the joint switching scale H within each preset time window to obtain the error superposition degree of joint adjustment within that time window; the larger this value, the more frequent the orientation adjustment, the larger the adjustment scale, the more drastic the joint pose adjustment, and the more significant the superposition effect of path deviation; conversely, it indicates that the joint pose adjustment is more stable.
[0041] The calculation process of the risk factor will be explained in detail below, specifically including: S12-7, analyzing the resonance characteristics of the wire feeding start-up based on vibration data to obtain the starting resonance ratio at which wire feeding starts. Vibration at the end of the binding mechanism will cause instantaneous swaying at the wire outlet position. The greater the vibration amplitude and the closer the frequency is to the resonance frequency band corresponding to the wire feeding speed, the more obvious the instantaneous deviation of the wire feed, thus disrupting the uniformity of wire winding. An analysis model can be constructed based on the needs of resonance characteristic analysis, such as constructing a convolutional neural network model, training the model, inputting vibration data into the trained model, and obtaining the starting resonance ratio at which wire feeding starts based on the model's output.
[0042] Of course, it is also possible to capture in real time the moment when the vibration dynamic peak appears in the vibration data at the end of the binding mechanism, that is, the moment when the maximum value is located; to obtain the moment when the steel wire displacement changes from a static state to a feeding state, which is recorded as the wire feeding start moment; for the current operation of the wire feeding mechanism, to obtain all cases in the calculation window where the time interval between the wire feeding start moment and the most recent vibration dynamic peak moment is less than the length of a time window, and to calculate the proportion of such cases in all wire feeding start times, that is, the start resonance ratio, denoted as W. The higher this index is, the higher the synchronization between the wire feeding start and the strongest vibration moment, and the greater the risk of instantaneous deviation.
[0043] S12-8. The risk coefficient is obtained based on the starting resonance ratio, the standard deviation of the wire feeding speed, the rising slope of the vibration data within the synchronous time window, and the mean square error of the frequency. The standard deviation of the wire feeding speed, the rising slope, and the mean square error of the frequency are three end-point risk characteristics of the binding mechanism. The wire feeding speed can be calculated based on the wire displacement and acquisition time to obtain the wire feeding speed within the time window. The standard deviation of the wire feeding speed is calculated using the standard deviation formula, which characterizes the stability of the wire feeding speed. Based on the vibration data, the rising trend of the vibration frequency within the time window is calculated, i.e., the rising slope. If there is no frequency rise within the window, it is recorded as 0. Further, based on the vibration data, the frequency spectrum of the end-point vibration signal and the frequency spectrum of the wire feeding speed signal within the time window are obtained. The mean square error of the two frequency spectra, i.e., the frequency mean square error, is calculated. The smaller this value, the closer the two are, and the higher the basic resonance risk. The risk coefficient is then derived by calculating the relationship between these three factors and the risk coefficient.
[0044] Specifically, the risk factor is calculated based on the following steps: First, obtain the first vibration interference coefficient by multiplying the starting resonance ratio by the rising slope. Second, obtain the second vibration interference coefficient by multiplying the standard deviation of the wire feed speed by the mean square error of the frequency. Third, obtain the risk factor by the ratio of the first vibration interference coefficient to the second vibration interference coefficient. That is, using the formula: A = The risk coefficient A is calculated, where W is the starting resonance ratio, a is the rising slope, b is the standard deviation of the wire feeding speed, and c is the mean square error of the frequency. The denominator is increased by 0.1 to prevent the formula from being invalid due to a denominator of 0.
[0045] The greater the upward slope, the smaller the standard deviation of the wire feeding speed, and the smaller the mean square error of the frequency, the greater the risk factor. That is, when the vibration frequency continues to increase, the wire feeding speed remains stable, and the frequencies of the two are close, the risk of resonance increases; conversely, the risk is lower. This means that when the frequency increases, the speed is not adjusted synchronously, which can easily cause the two to enter a state of resonance coupling, and the interference of vibration on the mechanism is amplified. Furthermore, the larger the starting resonance ratio W, the greater the risk of instantaneous offset at the starting end of the wire feeding. The product of the two can characterize the greater the impact of the end vibration on the resonance offset after the wire feeding starts.
[0046] At this point, the calculation of deviation accumulation, error superposition, and risk coefficient has been completed, and we proceed to step S13.
[0047] S13. Analyze the superimposed error of wire conveying based on the cumulative deviation, superposition of errors and risk coefficient, and implement graded dynamic compensation for the robot.
[0048] Specifically, this yields three feature vectors for each preset time window: a deviation accumulation factor representing the cumulative degree of continuous deviations within that time period, an error superposition factor representing the superposition of joint adjustment errors, and a risk coefficient representing the impact of mechanical end-effector vibrations. These feature vectors respectively characterize the error performance characteristics of the wire's own state, the mechanism's motion state, and the end-effector's operational state. The three feature vectors are then normalized to obtain the compensation amount for the robot's wire delivery, and a tiered compensation process is implemented. It can be understood that, due to the dynamic characteristics and superposition relationship of errors, the technical solution of this invention decomposes the error correlation through evolutionary analysis of the three feature vectors within all continuous calculation windows, thus avoiding insufficient compensation caused by isolated analysis.
[0049] For example, step S13 includes sub-steps S13-1 to S13-5, which are described in detail as follows: S13-1, obtain the first error transmission coefficient based on the change in error superposition degree in multiple preset time windows and the change in wire tension in the corresponding time windows. The essence of wire feeding path deviation is a composite deviation in the spatial position of the wire feeding mechanism. This deviation causes the wire to change from a preset straight feed to a non-linear detour, resulting in an increase in the actual force-bearing length of the wire, which in turn triggers tension fluctuations. To quantify this triggering effect: take all continuous calculation windows before the current moment and record the joint adjustment error superposition degree S and tension change rate T for each time window; calculate the change in joint adjustment error superposition degree ΔS and the change in tension change rate ΔT within each time window; count the number of windows in the same direction (both positive or both negative) of ΔS and ΔT in all continuous calculation windows before the current moment, and use the number of windows in the same direction / the total number of windows to obtain the first error transmission coefficient K1; the value of K1 ranges from 0 to 1. The closer K1 is to 1, the stronger the triggering correlation is, indicating that the increase / decrease in path deviation will directly trigger the enhancement / weakening of tension fluctuations. When K1 is close to 0, it indicates that there is no obvious correlation between path deviation and tension fluctuations, and the triggering end of the error chain is not activated.
[0050] S13-2. Based on the change in cumulative deviation over multiple preset time windows and the standard deviation of wire tension within the corresponding time windows, obtain the second error transmission coefficient. Since tension fluctuation is the direct cause of increased wire deformation, it is necessary to quantify the extent to which a unit tension fluctuation can induce cumulative deformation growth.
[0051] Using a single preset time window as a unit, first calculate the standard deviation g of all tension data within that window, reflecting the severity of tension fluctuations; then calculate the change ΔP of the cumulative degree of continuous deviation within that time window. Since only the aggravating effect is considered, if ΔP is negative, it indicates that the deformation has eased, so ΔP = 0 is taken. ΔP / g is used to obtain K2, which is the second error transmission coefficient. A larger second error transmission coefficient K2 indicates that under the same intensity of tension fluctuations, the cumulative increase in wire deformation is more significant, and the transmission efficiency is higher; a smaller K2 indicates that the effect of tension fluctuations on deformation is weak, and the path transmission effect of the error chain fails.
[0052] S13-3. Based on the resonance amplification results of wire conveying characterized by the first error transmission coefficient, the second error transmission coefficient, and the risk coefficient within the corresponding time window, the error compensation coefficient is obtained. The error coupling coefficient K can be calculated by multiplying the first error transmission coefficient K1 and the second error transmission coefficient K2. Each coefficient can be normalized to ensure that the calculated error coupling coefficient K is between 0 and 1. A larger K value indicates a stronger chain reaction where the wire conveying path deviates from the wire tension fluctuation, leading to increased wire deformation. For example, when the error coupling coefficient K is greater than 0.7, it means that the path deviation has a strong triggering correlation with the tension fluctuation, and the transmission efficiency of the tension fluctuation to increased deformation is high, indicating a strongly coupled error chain.
[0053] The end vibration generated by the binding mechanism not only directly causes feed deviation, but also amplifies the coupling effect of the error chain through resonance interference. When the vibration frequency is close to the frequency of the wire tension fluctuation, it will aggravate the instability of the tension and thus amplify the accumulation of deformation.
[0054] The risk coefficient is denoted as R. We can compare the risk coefficient R of the current time window with that of the previous time window. If the current R is greater than the previous window R, it indicates an increased risk of end-resonance, and the end-amplification factor F = 1 + R. The larger the R value, the more significant the amplification of tension fluctuations by resonance. If the current R is less than or equal to the previous window R, it indicates that the risk of end-resonance is stable or decreasing, and the end-amplification factor F = 1, indicating no amplification effect. Multiplying the F value by the K value yields the error compensation coefficient for each time window in real time. .
[0055] S13-4. Based on the error compensation coefficient, the preset graded control constant, and the reference speed of the wire conveyor, the compensation speed of the wire conveyor is obtained. Five consecutive time windows can be set as the compensation period to calculate the error compensation coefficient. The average value is used to smooth out potential oscillations caused by excessively large changes in wire feed speed; error compensation coefficient. The larger the average value, the stronger the coupling relationship of error overlap; conversely, the smaller the value, the weaker the coupling relationship of error overlap. The wire feeding speed is compensated in stages based on the error compensation coefficient, the preset graded control constant, and the reference speed of the wire conveyor.
[0056] For example, sub-step S13-4 includes: First, determining the constant value of the graded control constant. Specifically, this includes determining the graded control constant as a first constant value when the error compensation coefficient is less than a preset coefficient range; determining the graded control constant as a second constant value when the error compensation coefficient is within the preset coefficient range, the second constant value being greater than the first constant value; and determining the graded control constant as a third constant value when the error compensation coefficient is greater than the preset coefficient range, the third constant value being greater than the second constant value. For example, when the error compensation coefficient... When the average value is less than or equal to 0.3, the error generated during the wire feeding process is determined to be weakly coupled. Setting the graded control constant to less than 1 and slightly reducing the wire feeding speed can suppress the error. When the error compensation coefficient... When the average value is greater than 0.3 but less than or equal to 0.7, the error generated during the wire feeding process is determined to be of medium coupling. The graded control constant is set to 1 to linearly cancel the coupling effect. When the error compensation coefficient... When the average value is greater than 0.7, the error generated in the wire feeding process is determined to be strongly coupled. The graded control constant is set to be greater than 1, the wire feeding speed is greatly reduced, and the compensation is strengthened to suppress the resonance effect.
[0057] The second step is to obtain the graded compensation coefficient based on the constant value determined by the graded control constant and the error compensation coefficient. The graded compensation coefficient is the compensation value for correcting the wire conveying speed of the robotic arm. It can be calculated based on the constant value and the error compensation coefficient.
[0058] The third step is to obtain the compensation speed of the wire conveyor based on the product of the base speed and the graded compensation coefficient. The compensation speed can be calculated using a formula, for example: ,in, To compensate for speed, The reference speed can be determined based on the wire conveying speed set by the robotic arm. For graded control constants, Error compensation coefficient The average over multiple time windows (e.g., 5).
[0059] S13-5. Adjust the conveying wire speed of the robot arm within the next time window based on the compensation speed. The processing terminal sends a corresponding command to the wire conveying motor within the next time window based on the calculated compensation speed, thereby adjusting the conveying wire speed by adjusting the motor's rotational speed.
[0060] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
[0061] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.
Claims
1. A wire feeding speed compensation and accuracy calibration system for a wire binding robot, characterized in that, The system includes: multiple sensors, spaced apart along the wire conveying path of the wire binding robot; these sensors acquire and output data on wire tension, wire displacement, coordinates of each joint on the robot, and vibration data of the binding action; and a processing terminal connected to each sensor. The processing terminal is used to: obtain the cumulative deviation of wire conveying based on the wire tension and displacement; obtain the error superposition of robot joint movements based on the coordinate data of each joint; obtain the risk coefficient of wire conveying resonance caused by the robot performing the binding action based on the vibration data; and perform wire conveying superposition error analysis based on the cumulative deviation, the error superposition, and the risk coefficient, and implement graded dynamic compensation for the robot.
2. The wire feeding speed compensation and accuracy calibration system for the wire binding robot according to claim 1, characterized in that, The cumulative deviation of wire conveying is obtained based on the wire tension and the wire displacement, including: obtaining the rate of change of tension and the rate of change of displacement of the wire during conveying for at least two adjacent tracking cycles based on the wire tension and the wire displacement; obtaining the rate of change deviation coefficient characterizing the deformation of the wire conveying based on the rate of change of tension and the rate of change of displacement; and obtaining the cumulative deviation of wire conveying based on the cumulative results characterized by multiple consecutive rate of change deviation coefficients within a preset time window.
3. The wire feeding speed compensation and accuracy calibration system for the wire binding robot according to claim 2, characterized in that, The method for obtaining a rate deviation coefficient characterizing the deformation of the steel wire conveying is based on the tension change rate and the displacement change rate, including: obtaining a rate difference value based on the difference between the tension change rate and the displacement change rate, and configuring the maximum absolute value of the tension change rate and the displacement change rate as the absolute extreme value; and determining the ratio of the rate difference value to the absolute extreme value as the rate deviation coefficient.
4. The wire feeding speed compensation and accuracy calibration system for the wire binding robot according to claim 1, characterized in that, The error superposition degree of the robot's joint motion is obtained based on the coordinate data of each joint, including: obtaining the pose switching disorder degree of the robot based on the change characteristics of the coordinate data of each joint in the time dimension and the motion direction dimension; obtaining the joint switching scale of the robot based on the change characteristics of the coordinate data of each joint in the spatial dimension; and obtaining the error superposition degree of the robot's joint motion based on the pose switching disorder degree and the joint switching scale.
5. The wire feeding speed compensation and accuracy calibration system for the wire binding robot according to claim 4, characterized in that, Based on the variation characteristics of the coordinate data of each joint in the time and motion direction dimensions, the pose switching disorder of the robot is obtained, including: obtaining the coordinate change of each joint based on the difference of coordinate data in at least two adjacent tracking cycles; marking each coordinate change greater than a corresponding set change threshold as a target change, and statistically analyzing the target changes to obtain the time overlap of all joints within a preset time window; obtaining the direction switching rate of all joints based on the number of direction adjustments represented by the coordinate data of each joint within the preset time window; and obtaining the pose switching disorder of the robot based on the time overlap and the direction switching rate.
6. The wire feeding speed compensation and accuracy calibration system for the wire binding robot according to claim 4, characterized in that, Based on the spatial variation characteristics of the coordinate data of each joint, the joint switching scale of the robot is obtained, including: obtaining the angle between the direction vectors of each joint in two adjacent tracking cycles based on the coordinate data of each joint within a preset time window; obtaining the joint switching scale of the robot based on the angle between the direction vectors of all joints and the cumulative window time, wherein the cumulative window time is the product of the total number of joints and the number of time windows.
7. The wire feeding speed compensation and accuracy calibration system for the wire binding robot according to claim 1, characterized in that, The risk coefficient for wire conveying resonance caused by the robotic arm performing the binding action is obtained based on the vibration data, including: performing resonance characteristic analysis of the wire initiation conveying based on the vibration data to obtain the initiation resonance ratio; and obtaining the risk coefficient based on the initiation resonance ratio, the standard deviation of the wire feeding speed, the rising slope of the vibration data in the synchronization time window, and the mean square error of the frequency.
8. The wire feeding speed compensation and accuracy calibration system for the wire binding robot according to claim 7, characterized in that, The risk coefficient is obtained based on the starting resonance ratio, the standard deviation of the wire feeding speed, the rising slope of the vibration data within the synchronization time window, and the mean square error of the frequency. This includes: obtaining a first vibration interference coefficient based on the product of the starting resonance ratio and the rising slope; obtaining a second vibration interference coefficient based on the product of the standard deviation of the wire feeding speed and the mean square error of the frequency; and obtaining the risk coefficient based on the ratio of the first vibration interference coefficient to the second vibration interference coefficient.
9. The wire feeding speed compensation and accuracy calibration system for the wire binding robot according to claim 1, characterized in that, Based on the cumulative deviation, the superposition of errors, and the risk coefficient, an analysis of superimposed errors in wire conveying is performed, and graded dynamic compensation is implemented for the robot arm. This includes: obtaining a first error transmission coefficient based on the change in the superposition of errors over multiple preset time windows and the change in wire tension over corresponding time windows; obtaining a second error transmission coefficient based on the change in the cumulative deviation over multiple preset time windows and the standard deviation of wire tension over corresponding time windows; obtaining an error compensation coefficient based on the resonance amplification result of wire conveying represented by the first error transmission coefficient, the second error transmission coefficient, and the risk coefficient over corresponding time windows; obtaining a compensation speed for wire conveying based on the error compensation coefficient, a preset graded control constant, and the reference speed of wire conveying; and adjusting the wire conveying speed of the robot arm in the next time window based on the compensation speed.
10. The wire feeding speed compensation and accuracy calibration system for the wire binding robot according to claim 9, characterized in that, The compensation speed of the wire conveying is obtained based on the error compensation coefficient, a preset graded control constant, and the reference speed of the wire conveying, including: when the error compensation coefficient is less than a preset coefficient range, determining the graded control constant as a first constant value; when the error compensation coefficient is within the preset coefficient range, determining the graded control constant as a second constant value, the second constant value being greater than the first constant value; when the error compensation coefficient is greater than the preset coefficient range, determining the graded control constant as a third constant value, the third constant value being greater than the second constant value; obtaining the graded compensation coefficient based on the constant value determined by the graded control constant and the error compensation coefficient; and obtaining the compensation speed of the wire conveying based on the product of the reference speed and the graded compensation coefficient.