End precise control method suitable for automobile multi-axis gluing robot
By analyzing the abnormal characteristics of the robot's end-effector pose and vibration data, and optimizing the switching gain using a sliding mode controller, the problem of end-effector control instability in dynamic environments was solved, achieving precise control and improving the coating quality.
Patent Information
- Application Number
- CN202611059525.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-07-15
- Publication Date
- 2026-08-25
AI Technical Summary
Existing end-effector control methods for glue-applying robots are easily affected by mechanical factors of the robot itself in dynamic and complex real-world production environments, leading to motion uncertainty, glue-applying trajectory deviation, end-effector jitter, and glue-applying quality problems.
By acquiring robot end-effector pose and vibration data, analyzing abnormal features and joint vibration characteristics in the difference sequence, and using a sliding mode controller to adjust the switching gain, the end-effector joint angle control is optimized to achieve precise control.
It improves the accuracy of robot end-effector control, reduces problems such as glue piling, overflow, glue breakage, and uneven glue line size during the glue application process, and improves the glue application quality.
Smart Images
Figure CN122626232A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of intelligent manufacturing robot control technology, specifically to a precise end-effector control method applicable to multi-axis glue-applying robots for automobiles. Background Technology
[0002] In the automotive manufacturing industry, the adhesive application process is crucial for sealing, waterproofing, vibration reduction, and structural strength. Multi-axis industrial robots have become the mainstream equipment for automated adhesive application due to their flexibility and large workspace. However, achieving stable and high-quality adhesive application requires precise control of the end effector, which remains a core challenge.
[0003] Current end-effector control methods for dispensing robots primarily rely on the repeatability and accuracy of the robot's positioning and pre-programmed offline trajectories. However, in dynamic and complex real-world production environments, end-effectors are susceptible to mechanical factors, such as joint clearances and wear from long-term operation, which introduce significant motion uncertainties. These uncertainties translate into trajectory deviations, end-effector jitter, and even low-frequency vibrations. Especially during rapid directional changes or load variations, the stability of the robot's end-effector control further deteriorates, leading to deviations between the actual and theoretical dispensing trajectories. This results in issues such as glue buildup, overflow, glue breakage, or uneven glue line dimensions, ultimately reducing dispensing quality. Summary of the Invention
[0004] To address the aforementioned technical problems, this application provides a precise end-effector control method for automotive multi-axis adhesive applicators, thereby resolving the existing issues.
[0005] The end-efficiency precision control method for automotive multi-axis adhesive application robots in this application adopts the following technical solution: One embodiment of this application provides a method for precise end-effector control of a multi-axis adhesive applicator robot for automobiles, comprising the following steps: Acquire the end-effector pose data and vibration data of the robot during the multi-axis glue application process in automobiles, and form a difference sequence by combining the error results between the robot end-effector pose data and the preset trajectory at each time point; The first outlier is obtained by analyzing the characteristics of abrupt changes in the data in the difference sequence and using the distribution characteristics of the data fluctuation amplitude on both sides of the zero line in the difference sequence and the intensity of data oscillation in the difference sequence. By utilizing the pulse impact intensity of the end-joint vibration data and the high-frequency significance of the vibration data in the frequency domain, a second outlier is obtained. The variation of the second outlier and the overall average level are analyzed. Combined with the first outlier, a comprehensive outlier reflecting the robot's end-joint path tracking status is obtained in each glue application process. The switching gain of the sliding mode controller is adjusted based on the comprehensive abnormal values during the glue application process, thereby using the sliding mode controller to control the end joint angle of the robot during the multi-axis glue application process in automobiles.
[0006] Preferably, the analysis of data mutation characteristics in the difference sequence specifically includes: extracting all peaks and valleys in the difference sequence, obtaining mutation points in the absolute values of all peaks and valleys, calculating the ratio between the mean of all mutation points and the mean of all non-mutation points, and using it as the mutation significance value.
[0007] Preferably, the distribution characteristics of data fluctuation amplitude on both sides of the zero line in the difference sequence and the intensity of data oscillation in the difference sequence include: The first ratio is calculated as the ratio of the sum of the absolute values of all peaks and valleys greater than 0 in the difference sequence to the sum of the absolute values of all peaks and valleys in the difference sequence. The second ratio is calculated as the ratio of the sum of the absolute values of all peaks and valleys less than 0 in the difference sequence to the sum of the absolute values of all peaks and valleys in the difference sequence. The absolute value of the difference between the first ratio and the second ratio is used to characterize the distribution characteristics of the data fluctuation amplitude on both sides of the zero line in the difference sequence. Calculate the time interval between all adjacent peaks and valleys in the difference sequence, and calculate the coefficient of variation for all such time intervals to characterize the intensity of data oscillation in the difference sequence.
[0008] Preferably, the product between the significant mutation value and the absolute value of the difference is calculated, and the sum of this product and the coefficient of variation is taken as the first outlier.
[0009] Preferably, the pulse impact intensity of the end-joint vibration data and the high-frequency significance of the vibration data in the frequency domain specifically include: The ratio of the range to the root mean square value of the vibration data within each detection window is calculated as the vibration impact intensity of each detection window. The vibration data within each detection window is then subjected to frequency domain transformation. The ratio of the sum of the amplitudes corresponding to all frequencies above the centroid of the spectrum in the spectrum diagram to the sum of all amplitudes in the spectrum diagram is used as the high-frequency significance of the vibration.
[0010] Preferably, the second anomaly value is the product of the vibration impact intensity and the significance of the high-frequency vibration.
[0011] Preferably, the coordinates of the junctions between the straight line and the arc in the preset trajectory are obtained. When the three-dimensional spatial coordinates of the end pose enter the preset spatial tolerance range of the junction coordinates, the corresponding detection window is used as the direction change detection window. The spherical tolerance area formed by setting a specific distance value as the spatial detection radius with each junction coordinate as the center is used as the preset spatial tolerance range.
[0012] Preferably, the analysis of the differences in the changes of the second outlier and the overall average level specifically includes: Calculate the absolute value of the difference between the second outlier value of each direction-changing detection window and its adjacent detection windows. The sum of the mean of the absolute values of all direction-changing detection windows and the mean of the corresponding second outlier values of all detection windows is taken as the outlier value.
[0013] Preferably, the comprehensive outlier is the weighted sum of the normalized result of the first outlier and the normalized result of the outlier change value.
[0014] Preferably, the product of the normalized result of the comprehensive outlier obtained from the current glue application and the preset upper limit of switching gain is obtained. If the product is greater than or equal to the preset lower limit of switching gain, the product is used as the switching gain in the next glue application process; otherwise, the preset lower limit of switching gain is used as the switching gain in the next glue application process.
[0015] This application has at least the following beneficial effects: This application addresses motion uncertainty caused by various factors by deeply analyzing the abnormal changes in the deviation between the end effector pose and the preset adhesive application trajectory during the adhesive application process, as well as the vibration characteristics of the end effector joints. To address the issue that rapid changes in direction or load can easily affect the end effector control accuracy, it further analyzes the degree to which the joint vibration at the end effector position is affected by changes in motion state. This analysis more effectively reflects the impact of dynamic and complex actual production environments on end effector control. By comprehensively considering these characteristics, the application accurately evaluates the overall abnormal characteristics of the end effector path tracking state. Based on the evaluation results, the controller parameters for the end effector joint position are optimized, which helps improve the accuracy of robot end effector control. Attached Figure Description
[0016] Figure 1 A flowchart illustrating the steps of the precision control method for the end effector of a multi-axis adhesive applicator robot for automobiles provided in this application. Detailed Implementation
[0017] The following description, in conjunction with the accompanying drawings, details the specific scheme of the end-effector precision control method for automotive multi-axis adhesive application robots provided in this application.
[0018] This application provides an embodiment of a precision end-effector control method for a multi-axis automotive adhesive applicator. For details, please refer to [link to relevant documentation]. Figure 1 This includes the following steps: Step 1: Obtain the end-effector pose data and vibration data of the robot during the multi-axis glue application process, and form a difference sequence by combining the error results between the robot end-effector pose data and the preset trajectory at each time point.
[0019] This embodiment uses the adhesive application operation of an automobile engine end cover as an example to demonstrate the precise control of a multi-axis adhesive application robot. In the automotive manufacturing field, the adhesive application trajectory is preset, and high precision is required. The location on the automobile engine end cover where adhesive needs to be applied is the mating surface of the end cover. This mating surface contains numerous threaded holes and pin holes; therefore, the adhesive application trajectory must avoid these holes and be located inside them. This embodiment uses an offline programming method to obtain the preset adhesive application trajectory, which in this scenario mainly consists of straight lines and arcs.
[0020] During the glue application process, the robot's end-effector pose is a key parameter reflecting trajectory tracking accuracy. In this embodiment, a base coordinate system is constructed based on the robot's mounting base. The actual end-effector pose data in this base coordinate system is acquired through the robot controller's communication interface. This end-effector pose data represents the three-dimensional coordinates in the space of the end effector (glue gun). The end-effector pose acquisition frequency is set to 100 Hz. During glue application, abnormal vibrations of the robotic arm can be transmitted to the end-effector application gun, affecting its precise control. In this embodiment, vibration data of the end-effector joint position is acquired using an accelerometer. The vibration data acquisition frequency is set to 10 kHz.
[0021] Under ideal adhesive application conditions, the end effector of the adhesive application robot should move precisely along the preset adhesive application trajectory at a uniform speed to ensure the formation of a continuous, uniform sealing strip on the end cap mating surface that avoids threaded holes and pin holes. However, in actual operation, the dynamic characteristics of the robotic arm itself (such as mechanical vibration and joint clearance) and external factors (such as slight deformation of the robot mounting base and load fluctuations caused by changes in adhesive viscosity) can easily lead to unstable adhesive application or abnormal shaking at the robot's end effector, resulting in a significant deviation between the end effector pose data and the preset adhesive application trajectory.
[0022] Therefore, to monitor the motion state of the glue-applying robot, analyze the error between the robot's end-effector pose data and the preset trajectory at each moment, and form a difference sequence from the error results of the robot's end-effector pose data and the preset trajectory at each moment, preferably, in this embodiment, the projection error between the end-effector pose data and the preset trajectory in the normal plane at each moment is calculated, where a positive value indicates a deviation to one side of the normal line, and a negative value indicates a deviation to the opposite side. The obtained error results are then arranged in ascending order of time to obtain the difference sequence of the end-effector pose.
[0023] Step 2: Analyze the abrupt change characteristics of the data in the difference sequence, and use the distribution characteristics of the data fluctuation amplitude on both sides of the zero line in the difference sequence and the intensity of data oscillation in the difference sequence to obtain the first outlier.
[0024] During normal adhesive application, the difference sequence fluctuates within a small range around zero. When the end pose control is unstable, it can lead to sudden changes in the fluctuation amplitude and asymmetric deviation of the difference sequence, accompanied by unstable fluctuations in the oscillation period.
[0025] Therefore, taking the difference sequence obtained from applying adhesive to a certain workpiece as an example, preferably, this embodiment uses an adaptive multi-scale peak finding algorithm to detect peaks and outputs all peaks and valleys in the difference sequence. When a sudden change in fluctuation occurs, the absolute value of the change point will be significantly higher than other positions. Then, the absolute values of all peaks and valleys are calculated. Further, preferably, this embodiment uses the SOS detection algorithm to detect outliers in the absolute values of all peaks and valleys, outputting the anomaly probability value of each absolute value (between 0 and 1). Therefore, in this embodiment, the anomaly probability threshold is set to 0.9, and absolute values with an anomaly probability value higher than the threshold are marked as change points. Further, in this embodiment, the mean of all change points is calculated, denoted as A, which reflects the change intensity of the difference sequence; and the mean of all non-change points is calculated, denoted as B, which reflects the average level under normal fluctuations. The ratio of A to B is used as the change significance value, denoted as C, to characterize the data change characteristics in the difference sequence. The larger the C, the more significant the abnormal deviation caused by the change point is compared with the normal background fluctuation. Specifically, when no mutation point is detected, it indicates that the fluctuation of the difference sequence is relatively stable, and the value of parameter A is set to 0.
[0026] Furthermore, due to the inherent nonlinear characteristics of mechanical systems, such as differences in joint clearance and the directionality of friction, the following error of the robot's end effector becomes inconsistent. External loads also cause the deviation sequence to accumulate continuously on one side of the zero line, resulting in an asymmetric deviation in fluctuation amplitude. Under normal operating conditions, the fluctuation amplitude of the difference sequence is relatively consistent on both sides of the zero value. Therefore, the ratio between the sum of the absolute values of all peaks and valleys greater than 0 in the difference sequence and the sum of the absolute values of all peaks and valleys in the difference sequence is denoted as the first ratio; the ratio between the sum of the absolute values of all peaks and valleys less than 0 in the difference sequence and the sum of the absolute values of all peaks and valleys in the difference sequence is denoted as the second ratio. The absolute value of the difference between the first and second ratios characterizes the distribution of data fluctuation amplitude on both sides of the zero line in the difference sequence, denoted as D. The larger the D, the more uneven the distribution of fluctuation amplitude on both sides of the zero line, i.e., the more significant the asymmetric deviation characteristic. This can easily cause instability in the robot's end control, resulting in the actual application position of the sealing strip deviating from the center line of the preset trajectory. Additionally, the extruded strip may have uneven thickness across its cross-section.
[0027] Furthermore, under unstable operating conditions, the oscillation period of the difference sequence also exhibits corresponding non-uniform variations. The time intervals between all adjacent peaks and troughs in the difference sequence are then calculated, and the coefficient of variation (V) for all these time intervals is calculated to characterize the severity of data oscillations in the difference sequence. A larger V indicates more drastic fluctuations in the time intervals between adjacent peaks and troughs in the difference sequence, meaning that the oscillation period of the difference sequence varies unevenly within that time period and lacks stability.
[0028] Therefore, the first outlier reflecting the abnormal fluctuation amplitude and non-uniform oscillation period of the difference sequence is calculated using the following formula: In the formula, E is the first outlier, C is the significant value of the mutation, D is the result of taking the absolute value of the difference, and V is the coefficient of variation. The larger the obtained E, the more significant the abnormal fluctuation amplitude and non-uniform oscillation period characteristics in the difference sequence.
[0029] In the above formula, parameters C and D reflect the abrupt change and asymmetric deviation characteristics of the fluctuation amplitude of the difference sequence between the end pose and the preset adhesive application trajectory, respectively, while parameter V reflects the non-uniform characteristics of its oscillation period from the time dimension, thereby comprehensively evaluating the short-term instability characteristics of the end control.
[0030] Step 3: Using the pulse impact intensity of the end-joint vibration data and the high-frequency significance of the vibration data in the frequency domain, the second outlier is obtained. The variation of the second outlier and the overall average level are analyzed. Combined with the first outlier, a comprehensive outlier reflecting the robot's end-joint path tracking status is obtained in each glue application process.
[0031] Furthermore, during the actual glue application process, vibrations can occur due to factors such as the flexibility of the robotic arm structure, joint clearances, and servo drive response. These vibrations gradually transmit to the robot's end effector joints, causing abnormal vibrations in the glue gun. This can result in defects such as uneven glue line width, glue breaks, and glue line deviations from the predetermined trajectory. This is especially problematic on the engine end cover mating surface, where precise hole avoidance is required, and can even severely affect the sealing and bonding effect. Therefore, joint vibration characteristics are also an important factor affecting trajectory tracking errors.
[0032] Considering the strong transient nature of joint vibrations at the end-effector and their correlation with the end-effector's motion state, this embodiment sets a detection window of 0.5 seconds. Under the interference of the aforementioned factors, the vibration data of the end-effector will exhibit significant pulse characteristics and an increase in high-frequency components. Therefore, taking the i-th detection window as an example, the range and root mean square (RMS) values of the vibration data within this detection window are calculated. The ratio of the obtained range to the RMS value is taken as the vibration impact intensity of the i-th detection window, denoted as . When the vibration data includes severe impacts caused by joint clearance friction or sudden changes in instantaneous load, the range of the data will increase significantly, while the root mean square value will not fluctuate significantly and will be less affected by the impact. Therefore, the obtained... The larger the value, the higher the intensity of the impact pulse present in the joint vibration at the end position. Then, a fast discrete Fourier transform is used to obtain the spectrum of vibration data within the i-th detection window. Under normal operating conditions, the joint vibration energy at the end position is mainly concentrated at a lower fundamental frequency. When subjected to abnormal mechanical interference, other high-frequency components will appear in the spectrum. Therefore, the spectral centroid of the spectrum is first obtained, where the frequency of the spectral centroid represents the frequency with the most concentrated energy in the spectrum. Subsequently, the sum of the amplitudes corresponding to all frequencies above the spectral centroid in the spectrum is calculated, and the ratio is taken as the high-frequency significance of the vibration. The high-frequency significance of the vibration in the i-th detection window is denoted as . The result This reflects the significant high-frequency characteristics of the vibration data within the detection window.
[0033] Then, the second anomaly value reflecting the end-joint vibration within the i-th detection window is calculated, and the formula is as follows: In the formula, This represents the second outlier value of the end-joint vibration within the i-th detection window. The significance of the high-frequency vibration in the i-th detection window. Let be the vibration and impact intensity of the i-th detection window. The obtained... The larger the value, the more pronounced the pulse characteristics and high-frequency interference of the end joint vibration within the detection window. During the adhesive application process, regular or random ripples are easily generated on the surface of the sealing strip near the threaded hole and pin hole.
[0034] Meanwhile, abnormal vibrations at the end effector are often highly correlated with the robot's motion state. For example, when changing the direction of motion or passing through a specific position, the abnormal vibration characteristics will suddenly increase. However, the adhesive application path of the car engine end cover has many threaded holes and pin holes, and the adhesive application trajectory is mainly composed of alternating straight lines and arcs. When the glue gun moves to the junction of the straight line and the arc, a rapid and obvious change of direction will occur. Under the action of the robot arm's motion inertia, abnormal vibrations of the end effector joint are prone to occur. Therefore, the coordinates of the junction of the corresponding straight line and arc in the preset trajectory are obtained. When the three-dimensional spatial coordinates of the end effector pose enter the preset spatial tolerance range of the junction coordinates, the corresponding detection window is used as the change of direction detection window. In this embodiment, a spherical tolerance area is formed by setting a specific distance value as the spatial detection radius with each junction coordinate as the center, which is used as the preset spatial tolerance range. Preferably, in this embodiment, the spatial detection radius is set to 6.
[0035] The more significant the difference between the second outlier value of the directional change detection window and the second outlier value of its adjacent detection window, the more severe the influence of the motion state on the abnormal vibration at the end joint. Some detection windows have two adjacent detection windows, while others have only one or the next adjacent detection window. Furthermore, the absolute value of the difference between the second outlier values of each directional change detection window and its adjacent detection windows is calculated. The average of the absolute values obtained from all directional change detection windows is denoted as M. The larger the value of M, the more pronounced the abnormal vibration characteristics of the end joint during directional change. The sum of M and the average of the corresponding second outlier values of all detection windows is taken as the abnormal variation value, denoted as S. The obtained S reflects the comprehensive characteristics of the overall level of abnormal vibration at the end joint and the influence of changes in motion state.
[0036] In summary, based on the first outlier E and parameter S obtained from K historical glue application monitoring sessions (20 in this embodiment), the current parameters E and S are normalized using the maximum-minimum normalization method, denoted as E1 and S1 respectively. The comprehensive outlier reflecting the robot's end-effector path tracking status during the current glue application process is then calculated. In this embodiment, the specific calculation formula is as follows: In the formula, P is the comprehensive anomaly value reflecting the robot's end-effector path tracking status during the current glue application process. The normalized result of the first outlier. This is the normalized result of the outlier values. , As a weighting factor, and High-precision trajectory tracking is particularly important during the adhesive application process, and the first outlier directly reflects the deviation between the actual end-effector pose and the preset trajectory. Therefore, a larger value is set in the formula. Values, for example , Meanwhile, relying solely on position error feedback is insufficient to effectively reflect the impact of dynamic and complex actual production environments on end-effector control. Combining this with abnormal joint vibration interference helps assess the overall state of the robot's end-effector control. A larger P value indicates more significant overall abnormal characteristics of the robot's end-effector path tracking state during the adhesive application process, and a higher risk to the adhesive application quality.
[0037] Step 4: Adjust the switching gain of the sliding mode controller based on the comprehensive abnormal values during the glue application process, and then use the sliding mode controller to control the robot end joint angle during the multi-axis glue application process of the car.
[0038] This embodiment analyzes the abnormal changes in the end-effector pose and the preset adhesive application trajectory during the adhesive application process, as well as the degree to which joint vibration at the end-effector position is affected by mechanical factors and motion state, to accurately assess the comprehensive abnormal characteristics of the end-effector path tracking state. Sliding mode control (SMC) is used in the multi-axis robot control system to adjust the joint rotation angle. SMC employs an exponential reaching law to achieve precise end-effector control, and the control response rate of the end-effector joint position directly determines the accuracy of the adhesive application.
[0039] Therefore, in this embodiment, the relevant parameters of the sliding mode controller for the end-effector position are optimized based on the obtained comprehensive outlier value. Specifically, the larger the obtained P, the more unstable the robot's end-effector path tracking state is during this glue application process, requiring a larger switching gain to enable the robot to respond more quickly to external disturbances and improve control accuracy; conversely, the smaller the obtained P, the more stable the robot's end-effector path control is, allowing for a smaller switching gain to avoid excessive chattering. In this embodiment, the preset upper and lower limits of the switching gain for the sliding mode controller corresponding to the end-effector are set to 15 and 5, respectively. Based on the maximum and minimum values of the comprehensive outlier values obtained from K historical glue application processes (20 in this embodiment), the comprehensive outlier value obtained during the current glue application is normalized by the maximum-minimum value. The specific normalization process is existing technology and will not be elaborated in this embodiment.
[0040] Furthermore, the product of the normalized result of the current comprehensive outlier value obtained from the glue application and the preset upper limit of the switching gain is obtained. If the product is greater than or equal to the preset lower limit of the switching gain, the product is used as the switching gain for the next glue application process; otherwise, the preset lower limit of the switching gain is used as the switching gain for the next glue application process. In this embodiment, it is 5 to avoid excessively slow end-effector response. To avoid cold start, a preset initial switching gain (10 in this embodiment) is used for control during the first N glue application operations. Then, the sliding mode controller is used to adjust the angle of the robot's end joints to achieve precise control of the glue application robot.
[0041] The above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.
Claims
1. A precise end-effector control method applicable to multi-axis adhesive application robots for automobiles, characterized in that, Includes the following steps: Acquire the end-effector pose data and vibration data of the robot during the multi-axis glue application process in automobiles, and form a difference sequence by combining the error results between the robot end-effector pose data and the preset trajectory at each time point; The first outlier is obtained by analyzing the characteristics of abrupt changes in the data in the difference sequence and using the distribution characteristics of the data fluctuation amplitude on both sides of the zero line in the difference sequence and the intensity of data oscillation in the difference sequence. By utilizing the pulse impact intensity of the end-joint vibration data and the high-frequency significance of the vibration data in the frequency domain, a second outlier is obtained. The variation of the second outlier and the overall average level are analyzed. Combined with the first outlier, a comprehensive outlier reflecting the robot's end-joint path tracking status is obtained in each glue application process. The switching gain of the sliding mode controller is adjusted based on the comprehensive abnormal values during the glue application process, thereby using the sliding mode controller to control the end joint angle of the robot during the multi-axis glue application process in automobiles.
2. The precision end-effector control method for a multi-axis automotive adhesive applicator as described in claim 1, characterized in that, The analysis of data mutation characteristics in the difference sequence specifically includes: extracting all peaks and valleys in the difference sequence, obtaining mutation points in the absolute values of all peaks and valleys, calculating the ratio between the mean of all mutation points and the mean of all non-mutation points, and using this ratio as the mutation significance value.
3. The precision end-effector control method for a multi-axis automotive adhesive applicator as described in claim 2, characterized in that, The distribution characteristics of data fluctuation amplitude on both sides of the zero line in the difference sequence and the degree of data oscillation in the difference sequence include: The first ratio is calculated as the ratio of the sum of the absolute values of all peaks and valleys greater than 0 in the difference sequence to the sum of the absolute values of all peaks and valleys in the difference sequence. The second ratio is calculated as the ratio of the sum of the absolute values of all peaks and valleys less than 0 in the difference sequence to the sum of the absolute values of all peaks and valleys in the difference sequence. The absolute value of the difference between the first ratio and the second ratio is used to characterize the distribution characteristics of the data fluctuation amplitude on both sides of the zero line in the difference sequence. Calculate the time interval between all adjacent peaks and valleys in the difference sequence, and calculate the coefficient of variation for all such time intervals to characterize the intensity of data oscillation in the difference sequence.
4. The precision end-effector control method for automotive multi-axis adhesive application robots as described in claim 3, characterized in that, Calculate the product between the significant mutation value and the absolute value of the difference, and use the sum of this product and the coefficient of variation as the first outlier.
5. The precision end-effector control method for a multi-axis automotive adhesive applicator as described in claim 1, characterized in that, The pulse impact intensity of the end-joint vibration data and the high-frequency significance of the vibration data in the frequency domain specifically include: The ratio of the range to the root mean square value of the vibration data within each detection window is calculated as the vibration impact intensity of each detection window. The vibration data within each detection window is then subjected to frequency domain transformation. The ratio of the sum of the amplitudes corresponding to all frequencies above the centroid of the spectrum in the spectrum diagram to the sum of all amplitudes in the spectrum diagram is used as the high-frequency significance of the vibration.
6. The precision end-effector control method for a multi-axis automotive adhesive applicator as described in claim 4, characterized in that, The second anomaly is the product of the vibration impact intensity and the significance of the high-frequency vibration.
7. The precision end-effector control method for a multi-axis automotive adhesive applicator as described in claim 1, characterized in that, Obtain the coordinates of the junctions between the straight line and the arc in the preset trajectory. When the three-dimensional spatial coordinates of the end pose enter the preset spatial tolerance range of the junction coordinates, the corresponding detection window is used as the direction change detection window. The spherical tolerance area formed by setting a specific distance value as the spatial detection radius with each junction coordinate as the center is used as the preset spatial tolerance range.
8. The precision end-effector control method for a multi-axis automotive adhesive applicator as described in claim 7, characterized in that, The analysis of the differences in the second outlier and the overall average level specifically includes: Calculate the absolute value of the difference between the second outlier value of each direction-changing detection window and its adjacent detection windows. The sum of the mean of the absolute values of all direction-changing detection windows and the mean of the corresponding second outlier values of all detection windows is taken as the outlier value.
9. The precision end-effector control method for a multi-axis automotive adhesive applicator as described in claim 8, characterized in that, The comprehensive outlier is the weighted sum of the normalized result of the first outlier and the normalized result of the outlier change value.
10. The precision end-effector control method for a multi-axis automotive adhesive applicator as described in claim 1, characterized in that, Obtain the product of the normalized result of the comprehensive outlier obtained from the current glue application and the preset upper limit of switching gain. If the product is greater than or equal to the preset lower limit of switching gain, then use the product as the switching gain in the next glue application process; otherwise, use the preset lower limit of switching gain as the switching gain in the next glue application process.