An intelligent assembly system for mechanical automation

Abstract

This invention relates to the field of intelligent assembly technology and discloses an intelligent assembly system for mechanical automation. This system addresses the problem of gradual accumulation of assembly errors during mechanical assembly. The system includes: collecting speed rationality assessment parameters for mechanical assembly; evaluating the speed adjustment drive index; adjusting the current assembly speed if necessary to obtain the actual assembly speed; collecting sensor data from the workpiece to assess the qualification of the assembly action; transmitting the sensor monitoring results to an early warning module if the assembly action is deemed unqualified; and collecting image data of the assembled workpiece using an image acquisition device to determine if the assembled workpiece meets visual inspection requirements. If the assembled workpiece does not meet the visual inspection requirements, the visual inspection results are transmitted to the early warning module, effectively improving the accuracy of mechanical assembly.

Description

Technical Field

[0001] This invention relates to the field of intelligent assembly technology, and more specifically to an intelligent assembly system for mechanical automation. Background Technology

[0002] Mechanical automated assembly systems, as crucial supporting equipment in modern manufacturing, are widely used in various fields such as automotive, electronics, precision instruments, and intelligent manufacturing. These systems typically rely on multiple functional modules, including robotic arm execution, workpiece supply, visual inspection, control scheduling, and human-machine interaction, to achieve fully automated operation from parts transfer and positioning to precision assembly. Through the high-precision execution of the robotic arm and real-time detection by sensors, the assembly system can ensure assembly efficiency and consistency in mass production scenarios, providing a key technological foundation for enterprises to improve production capacity and reduce labor costs.

[0003] During assembly tasks, to achieve closed-loop control from workpiece supply, robotic arm operation, position detection, assembly completion, and quality verification, existing systems need to have the ability to coordinate and respond to assembly sequence, assembly cycle time, and process parameters, thereby ensuring high-precision docking between parts. With the continuous development of concepts such as "real-time detection," "closed-loop control," and "dynamic scheduling," mechanical automation assembly systems are gradually evolving towards having process monitoring, real-time feedback, and state self-adaptation capabilities.

[0004] In recent years, existing technologies have generally adopted assembly modes based on fixed cycle times or preset speeds. This means that during assembly, robotic arm operations are performed according to a pre-set speed and sequence, and after assembly, an inspection module corrects for errors. Some methods improve assembly accuracy by analyzing historical assembly data to identify deviation trends and using standardized parameter tables to adjust the speed.

[0005] However, the above-mentioned technologies have at least the following technical problems:

[0006] In actual assembly processes, due to limitations in the robotic arm's operating state, fixed assembly speeds or post-assembly corrections are often insufficient to respond promptly to uncertainties during the assembly process. Relying on a uniform calibration process after assembly completion can easily lead to the gradual accumulation of assembly errors during the execution phase, causing compensation delays and consequently affecting the accuracy of the assembly results and the stability of the system operation. Summary of the Invention

[0007] In order to overcome the above-mentioned defects of the prior art, the present invention provides an intelligent assembly system for mechanical automation to solve the problems existing in the background art.

[0008] To achieve the above objectives, the present invention provides the following technical solution:

[0009] An intelligent assembly system for mechanical automation, the system comprising: a workpiece supply and positioning module, configured to obtain required workpiece information, convey the workpiece according to the workpiece information, and collect workpiece size and position parameters in real time, and use a positioning device to park the workpiece at the assembly station according to the workpiece size and position parameters; a robotic arm docking module, configured to complete the movement and docking of the workpiece at the assembly station according to a control instruction; an assembly speed adjustment module, configured to collect speed rationality evaluation parameters for mechanical assembly, the speed rationality evaluation parameters including the position coordinates, three-axis acceleration, and angle of the end effector of the robotic arm, evaluate a speed adjustment driving index according to the speed rationality evaluation parameters, determine whether the current assembly speed needs to be adjusted according to the speed adjustment driving index, and if it is determined that the current assembly speed needs to be adjusted, adjust the current assembly speed according to the speed adjustment driving index to obtain an actual assembly speed; a workpiece assembly module, configured to assemble the docked workpiece by the end effector of the robotic arm according to the actual assembly speed; a sensing and monitoring module, during the assembly process, collect workpiece sensor data in real time through a sensor group, the sensor data including assembly torque, clamping pressure, and displacement change, evaluate an assembly action qualification degree according to the sensor data, determine whether the assembly action is qualified according to the assembly action qualification degree, and if it is determined that the assembly action is unqualified, transmit the sensing and monitoring result to an early warning module; a vision detection module, if it is determined that the assembly action is qualified, collect image data of the assembled workpiece through an image acquisition device, the image data including appearance features and position coordinates, obtain a workpiece assembly standard model, compare the image data of the assembled workpiece with the workpiece assembly standard model, and determine whether the assembled workpiece meets the vision detection requirements, and if the assembled workpiece does not meet the vision detection requirements, transmit the vision detection result to the early warning module; a data management and storage module, if the assembled workpiece meets the vision detection requirements, store and manage the workpiece information, sensor data, and image data during the assembly process; an early warning module, if receiving the sensing and monitoring result, issue an early warning for unqualified assembly actions; if receiving the vision detection result, issue an early warning for unqualified assembly images; the steps for obtaining the speed adjustment driving index are: set a detection time period, obtain the position coordinates of the end effector of the robotic arm within the historical detection time period, and evaluate a regression positioning stability influence coefficient according to the position coordinates of the end effector of the robotic arm; obtain the three-axis acceleration of the end effector of the robotic arm within the historical detection time period, and evaluate a vibration disturbance sensitivity influence coefficient according to the three-axis acceleration of the end effector of the robotic arm; obtain the angle of the end effector of the robotic arm within the historical detection time period, and evaluate a driving stability margin influence coefficient according to the angle of the end effector of the robotic arm;The regression positioning stability influence coefficient, vibration disturbance sensitivity influence coefficient, and drive stability margin influence coefficient are normalized. The speed regulation drive index is then evaluated based on the normalized regression positioning stability influence coefficient, vibration disturbance sensitivity influence coefficient, and drive stability margin influence coefficient. The specific steps for obtaining this index are as follows: In the formula, This is expressed as the speed regulation drive index. This is expressed as the normalized regression positioning stability impact coefficient. This is expressed as the normalized vibration disturbance sensitivity influence coefficient. This is expressed as the normalized driving stability margin influence coefficient. , , These are the weighting coefficients of the normalized regression positioning stability influence coefficient, the normalized vibration disturbance sensitivity influence coefficient, and the normalized drive stability margin influence coefficient.

[0010] Preferably, the steps for obtaining the regression positioning stability influence coefficient are as follows: First, obtain the target position coordinates of the robotic arm end effector for each regression positioning operation within the historical detection period, forming a reference point sequence, with the data in the reference point sequence recorded as reference points. Second, obtain the actual position coordinates of the robotic arm end effector for each regression positioning operation within the historical detection period, forming an actual regression positioning sequence, with the data in the actual regression positioning sequence recorded as actual points. Third, for each regression, calculate the three-dimensional Euclidean distance between the corresponding reference point and the actual point to obtain the positioning deviation value. Fourth, iterate through each regression to obtain a positioning deviation sequence, extract the maximum value in the positioning deviation sequence, and record it as the maximum deviation value. Fifth, calculate the standard deviation of the data in the positioning deviation sequence to obtain the deviation fluctuation, and sum the maximum deviation value and the deviation fluctuation to obtain the regression positioning stability influence coefficient.

[0011] Preferably, the step of obtaining the vibration disturbance sensitivity influence coefficient is as follows: A sampling frequency is set; based on the sampling frequency, the three-axis vibration acceleration data of the robotic arm end effector in a stable standby state without performing assembly operations is obtained and recorded as the reference vibration sequence, with the three axes being the X-axis, Y-axis, and Z-axis; based on the sampling frequency, the three-axis acceleration data of the robotic arm end effector during the actual assembly process within a historical detection period is obtained and recorded as the actual vibration sequence; the root mean square (RMS) value is calculated for each direction of the actual vibration sequence and the reference vibration sequence, resulting in the actual RMS value of the X-axis, the actual RMS value of the Y-axis, the actual RMS value of the Z-axis, the reference RMS value of the X-axis, the reference RMS value of the Y-axis, and the reference RMS value of the Z-axis; the ratio of the actual RMS value to the reference RMS value in each direction is calculated. The vibration amplitude ratios in the X, Y, and Z axes were obtained. A short-time Fourier transform was performed on the triaxial acceleration data in the actual vibration sequence to extract the triaxial high-frequency disturbance energy, including X-axis, Y-axis, and Z-axis high-frequency disturbance energy. The triaxial reference high-frequency energy was obtained, and the ratios of the triaxial high-frequency disturbance energy to the corresponding triaxial reference high-frequency energy were calculated to obtain the X-axis energy surge ratio, Y-axis energy surge ratio, and Z-axis energy surge ratio. The vibration disturbance sensitivity influence coefficient was calculated based on these ratios.

[0012] Preferably, the steps for obtaining the drive stability margin influence coefficient are as follows: Obtain the control command angle sequence and actual feedback angle sequence of the robotic arm end effector within a historical detection period; for each command, calculate the difference between the control command angle value and the actual feedback angle value, and take the absolute value to obtain the control delay error; iterate through all commands within the historical detection period to obtain the error sequence within the historical detection period; perform first-order differencing on the error sequence, calculate the error change value between adjacent commands, construct an error change sequence, and based on the error change sequence, calculate the square value of the error change value for each command, and sum and average all square values ​​to obtain the average value of the square of the error fluctuation slope; set a dynamic threshold, obtain continuous segments in the error sequence that are greater than the dynamic threshold, and count the time required for each segment to fall back from the maximum error to below the dynamic threshold, denoted as the recovery time; calculate the average recovery time by averaging all recovery times in the error sequence; multiply the average value of the square of the error fluctuation slope with the average recovery time, and perform square root processing on the product result to obtain the drive stability margin influence coefficient.

[0013] Preferably, the step of determining whether the current assembly speed needs to be adjusted based on the speed adjustment drive index is as follows: compare the speed adjustment drive index with the adjustment threshold; if the speed adjustment drive index is greater than the adjustment threshold, then the current assembly speed needs to be adjusted; if the speed adjustment drive index is less than or equal to the adjustment threshold, then the current assembly speed does not need to be adjusted.

[0014] Preferably, the step of obtaining the actual assembly speed is as follows: calculate the ratio of the adjustment threshold to the speed adjustment drive index to obtain the speed adjustment factor; multiply the speed adjustment factor with the current assembly speed to obtain the actual assembly speed.

[0015] Preferably, the steps for obtaining the assembly action qualification are as follows: during the assembly process, the assembly torque, clamping pressure, and displacement change of each sampling point are collected in real time; the standard values ​​of the assembly torque, clamping pressure, and displacement change are obtained; the torque deviation is calculated based on the assembly torque and the standard value of the assembly torque at each sampling point; the clamping pressure deviation is calculated based on the clamping pressure and the standard value of the clamping pressure at each sampling point; the displacement change deviation is calculated based on the displacement change and the standard value of the displacement change at each sampling point; and the assembly action qualification is calculated based on the torque deviation, clamping pressure deviation, and displacement change standard value.

[0016] Preferably, the step of determining whether an assembly action is qualified based on the assembly action qualification degree is as follows: compare the assembly action qualification degree with the action qualification threshold; if the assembly action qualification degree is greater than or equal to the action qualification threshold, the assembly action is determined to be qualified; if the assembly action qualification degree is less than the action qualification threshold, the assembly action is determined to be unqualified.

[0017] Preferably, the step of comparing the image data of the assembled workpiece with the standard assembly model of the workpiece to determine whether the assembled workpiece meets the visual inspection requirements is as follows: Acquire the image data of the assembled workpiece; extract the edge contours and key structural feature points of the current workpiece image using edge detection and contour extraction algorithms, denoted as the current feature set; extract the edge contours and key feature points of the standard assembly model of the workpiece, denoted as the standard feature set; compare each key feature point in the current feature set with each key feature point in the standard feature set one by one, calculate the Euclidean distance between each corresponding key feature point, and calculate the mean to obtain the feature matching deviation. The system obtains the center coordinates of the current workpiece image and the center coordinates of the workpiece assembly standard model, denoted as the current center coordinates and the standard center coordinates, respectively. It calculates the Euclidean distance between the current center coordinates and the standard center coordinates to obtain the image coordinate deviation value. The system then multiplies the feature matching deviation value and the image coordinate deviation value, takes the square root, and uses the reciprocal as the image pass rate. The system compares the image pass rate with the image pass rate threshold. If the image pass rate is greater than or equal to the image pass rate threshold, the assembled workpiece is determined to meet the visual inspection requirements. If the image pass rate is less than the image pass rate threshold, the assembled workpiece is determined to not meet the visual inspection requirements.

[0018] The technical effects and advantages of this invention are as follows:

[0019] The system collects speed rationality assessment parameters for mechanical assembly, evaluates the speed adjustment drive index, and adjusts the current assembly speed if necessary to obtain the actual assembly speed. It also collects sensor data from the workpiece to assess the qualification of the assembly action. If the assembly action is deemed unqualified, the sensor monitoring results are transmitted to the early warning module. If the assembly action is deemed qualified, image data of the assembled workpiece is collected using an image acquisition device to determine if it meets visual inspection requirements. If the assembled workpiece does not meet visual inspection requirements, the visual inspection results are transmitted to the early warning module, effectively improving the accuracy of mechanical assembly. Attached Figure Description

[0020] Figure 1 This is a structural diagram of an intelligent assembly system for mechanical automation provided in an embodiment of this application. Detailed Implementation

[0021] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. In addition, the forms of the various structures described in the following embodiments are merely illustrative. The intelligent assembly system for mechanical automation involved in the present invention is not limited to the structures described in the following embodiments. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] This invention provides an intelligent assembly system for mechanical automation, such as... Figure 1 As shown, the system includes:

[0023] The workpiece supply and positioning module is used to acquire workpiece information required for the assembly task. Based on the workpiece information, the workpiece is transported from the raw material warehouse through an automated conveying device. The workpiece size and position parameters are collected in real time. The positioning device is used to perform precise calibration based on the workpiece size and position parameters, and the workpiece is placed at the assembly station to ensure that the workpiece posture meets the assembly requirements, thereby providing standardized input conditions for subsequent assembly operations.

[0024] The robotic arm docking module is used to move and dock workpieces in the assembly station according to control commands;

[0025] The assembly speed adjustment module is used to collect speed rationality evaluation parameters for mechanical assembly. The speed rationality evaluation parameters include the position coordinates, three-axis acceleration, and angle of the end effector of the robotic arm. The speed adjustment drive index is obtained based on the speed rationality evaluation parameters. The speed adjustment drive index is used to determine whether the current assembly speed needs to be adjusted. If the current assembly speed needs to be adjusted, the current assembly speed is adjusted according to the speed adjustment drive index to obtain the actual assembly speed.

[0026] In this embodiment, it should be specifically explained that the step of obtaining the speed regulation drive index is as follows:

[0027] Set a detection time period, obtain the position coordinates of the robotic arm end effector within the historical detection time period, and evaluate the regression positioning stability influence coefficient based on the position coordinates of the robotic arm end effector. It should be noted that the detection time period can be adjusted according to the actual situation, for example, the detection time period can be 1 day or 1 week.

[0028] The three-axis acceleration of the robotic arm end effector during the historical detection period is obtained, and the vibration disturbance sensitivity influence coefficient is evaluated based on the three-axis acceleration of the robotic arm end effector.

[0029] The angle of the end effector of the robotic arm during the historical detection period is obtained, and the drive stability margin influence coefficient is evaluated based on the angle of the end effector of the robotic arm.

[0030] The regression positioning stability influence coefficient, vibration disturbance sensitivity influence coefficient, and drive stability margin influence coefficient are normalized. Specifically, vector normalization is used. In this embodiment, these three coefficients together form a vector. The norm value is obtained by calculating the square root of the sum of their squares, and then each coefficient is divided by this norm value to complete the normalization process. Vector normalization is used to ensure comparability and consistency among the three dynamic composite indices when calculating the speed regulation drive index. Since vector normalization is existing technology, its specific algorithm steps are not detailed in this embodiment. The speed regulation drive index is evaluated based on the normalized regression positioning stability influence coefficient, vibration disturbance sensitivity influence coefficient, and drive stability margin influence coefficient. The specific steps for obtaining this index are as follows:

[0031] ;

[0032] In the formula, This is expressed as the speed regulation drive index. The coefficient represents the normalized regression positioning stability influence coefficient. A larger coefficient indicates greater fluctuations or offsets in the positioning coordinates of the end effector during regression or repetitive positioning operations by the robotic arm in the assembly process, reflecting poor regression positioning stability at the current assembly speed. Insufficient positioning accuracy at higher speeds can easily lead to increased assembly errors, workpiece docking failures, or actuator interference. Therefore, when the regression positioning stability influence coefficient increases, the speed adjustment drive index should be increased accordingly to trigger the system to reduce the current assembly speed, thereby enhancing the stability and positioning accuracy of the assembly process. This is represented as the normalized vibration disturbance sensitivity influence coefficient. A larger vibration disturbance sensitivity influence coefficient indicates a more sensitive assembly system to external vibrations or internal driving disturbances, making it more prone to execution errors, assembly deviations, or structural instability. Since excessive assembly speed may exacerbate the system's response to vibration disturbances, leading to a decrease in the pass rate, when the vibration disturbance sensitivity influence coefficient increases, the speed adjustment drive index should be increased accordingly. This prompts the system to slow down the assembly speed to reduce the assembly quality risks caused by vibration interference, thereby improving the stability and reliability of the assembly process. This represents the normalized drive stability margin influence coefficient. A larger coefficient indicates a smaller stability margin for the current assembly actuator, meaning that under high-speed operation or rapid response conditions, unstable phenomena such as decreased drive accuracy, position drift, or execution lag are more likely to occur. To avoid assembly errors or malfunctions due to insufficient drive system stability margin, the speed adjustment drive index should be significantly increased when this coefficient rises. This prompts the assembly speed to adjust downwards, thereby reducing the execution load, ensuring the drive system operates within a controllable range, and improving the accuracy and reliability of the assembly process. , , These represent the weighting coefficients of the normalized regression positioning stability influence coefficient, the normalized vibration disturbance sensitivity influence coefficient, and the normalized drive stability margin influence coefficient. , , Obtained through the analytic hierarchy process, and The Analytic Hierarchy Process (AHP) is a weighting method based on a multi-factor judgment structure, suitable for solving the problem of directly quantifying the influence of multiple evaluation factors on the target result. By constructing a multi-level hierarchical structure including a target layer, a criterion layer, and an indicator layer, the velocity regulation driving index is used as the target layer, and the regression positioning stability influence coefficient, vibration disturbance sensitivity influence coefficient, and driving stability margin influence coefficient are used as indicator layer factors. Based on expert experience, historical assembly effects, or known assembly data, each factor is compared pairwise to construct a judgment matrix. The consistency ratio is calculated and normalization is performed, thereby obtaining the weight coefficients of each influence coefficient in the construction of the velocity regulation driving index.

[0033] In this embodiment, it should be specifically explained that the steps for obtaining the regression positioning stability influence coefficient are as follows:

[0034] Obtain the target position coordinates of the robotic arm end effector for each regression positioning operation within the historical detection period, and form a reference point sequence. The data in the reference point sequence are recorded as reference points.

[0035] Obtain the actual position coordinates of the robotic arm end effector for each regression positioning operation within the historical detection period, forming an actual regression positioning sequence. The data in the actual regression positioning sequence are recorded as actual points.

[0036] For each regression, the three-dimensional Euclidean distance between the corresponding reference point and the actual point is calculated to obtain the positioning deviation value;

[0037] By iterating through each regression, a positioning deviation sequence is obtained. The maximum value in the positioning deviation sequence is extracted and recorded as the maximum deviation value.

[0038] The standard deviation of the data in the positioning deviation sequence is calculated to obtain the deviation fluctuation. The maximum deviation value and the deviation fluctuation are summed to obtain the regression positioning stability influence coefficient.

[0039] By analyzing the positioning actions repeatedly performed by the robotic arm in multiple historical tasks, the deviation between the target position and the actual position is extracted. A stability assessment index is then constructed by combining the maximum deviation and the standard deviation of the fluctuation. This index comprehensively reflects the worst-case performance and stability of repeated positioning, effectively identifying the combined effects of high-frequency micro-deviations and low-frequency severe deviations, ensuring that the speed adjustment logic is more targeted and safer.

[0040] In this embodiment, it should be specifically explained that the steps for obtaining the vibration disturbance sensitivity influence coefficient are as follows:

[0041] Set the sampling frequency, and obtain the three-axis vibration acceleration data of the robotic arm end effector in a stable standby state when no assembly operation is performed, based on the sampling frequency. Record it as the reference vibration sequence, with the three axes being the X-axis, Y-axis and Z-axis.

[0042] Based on the sampling frequency, the three-axis acceleration data of the end effector of the robotic arm during the actual assembly process in the historical detection period are obtained and recorded as the actual vibration sequence.

[0043] The root mean square (RMS) values ​​are calculated for the actual vibration sequence and the reference vibration sequence in each direction, resulting in the actual RMS values ​​for the X-axis, Y-axis, and Z-axis, as well as the reference RMS values ​​for the X-axis, Y-axis, and Z-axis.

[0044] The ratios of the actual root mean square value to the reference root mean square value in each direction are calculated to obtain the vibration change amplitude ratios in the X-axis, Y-axis, and Z-axis directions, which are used to characterize the degree of vibration response change of the assembly system in the three axes during the current testing cycle.

[0045] Short-time Fourier transform is performed on the triaxial acceleration data in the actual vibration sequence to extract the triaxial high-frequency disturbance energy, which includes X-axis high-frequency disturbance energy, Y-axis high-frequency disturbance energy and Z-axis high-frequency disturbance energy.

[0046] The short-time Fourier transform (SFT) is a time-frequency analysis method used to analyze the frequency variations of non-stationary signals. The basic principle is to divide the original signal into time windows and perform a Fourier transform within each short time interval to obtain the signal's spectral information at different time points.

[0047] It should be noted that the method of extracting triaxial high-frequency disturbance energy by performing short-time Fourier transform on triaxial acceleration data in actual vibration sequences is an existing technology, and this embodiment will not provide a detailed description of its specific steps.

[0048] The triaxial reference high-frequency energy is obtained, and the ratio of the triaxial high-frequency disturbance energy to the corresponding triaxial reference high-frequency energy is calculated to obtain the X-axis energy surge ratio, Y-axis energy surge ratio, and Z-axis energy surge ratio. By calculating the ratio of the triaxial high-frequency disturbance energy to the corresponding reference energy, the abnormal amplification degree of vibration energy in each axis direction can be identified, thereby accurately judging the multidimensional impact of vibration disturbance on system stability.

[0049] The triaxial reference high-frequency energy refers to the statistical reference value of the high-frequency energy in the triaxial acceleration signal extracted by short-time Fourier transform when the assembly is in a stable state. It can be obtained by statistical calculation of historical stable assembly data.

[0050] The vibration disturbance sensitivity influence coefficient is calculated based on the vibration amplitude ratios in the X-axis, Y-axis, and Z-axis directions, as well as the X-axis energy surge ratio, Y-axis energy surge ratio, and Z-axis energy surge ratio. The specific steps for obtaining this coefficient are as follows:

[0051] ;

[0052] In the formula, It is expressed as the vibration disturbance sensitivity influence coefficient. Expressed as the ratio of vibration amplitude changes in the X-axis direction, This is expressed as the ratio of vibration amplitude changes in the Y-axis direction. Expressed as the ratio of vibration amplitude changes in the Z-axis direction, Expressed as the X-axis energy surge ratio, Expressed as the Y-axis energy surge ratio, It is expressed as the Z-axis energy surge ratio.

[0053] The vibration amplitude ratio reflects the system's response intensity to instantaneous disturbances, while the energy surge ratio characterizes the energy concentration effect of the system under high-frequency disturbances. Together, they measure the overall sensitivity of the system to vibration in complex assembly environments. By summing the squares of the six indicators and taking the root, not only can sign interference be avoided, but the impact of abnormal disturbances can also be effectively amplified. This is beneficial for accurately identifying the vibration disturbance risk level of the current assembly system and provides a reliable basis for speed regulation strategies.

[0054] In this embodiment, it should be specifically explained that the steps for obtaining the driving stability margin influence coefficient are as follows:

[0055] Obtain the control command angle sequence and actual feedback angle sequence of the robotic arm end effector during the historical detection period, and ensure that the command and feedback data are aligned with a unified timestamp for subsequent stability margin response analysis;

[0056] The control command angle sequence refers to the set of angle commands issued by the control system to drive the movement of the end effector of the robotic arm within a historical detection period. It records the target angle value set for each control action and reflects the execution path expected by the system.

[0057] The actual feedback angle sequence refers to the actual motion angle values ​​that the end effector collects and reports in real time through the angle sensor within a corresponding time period. It reflects the actual execution of the robotic arm and is used to compare control accuracy and response stability.

[0058] For each instruction, the difference between the control instruction angle value and the actual feedback angle value is calculated, and the absolute value is taken to obtain the control delay error. By traversing all instructions within the historical detection period, the error sequence within the historical detection period is obtained, which reflects the delay stability between the actuator control signal and the actual action.

[0059] The error sequence is first-order differencing is performed to calculate the error change value between adjacent instructions, and an error change sequence is constructed. Based on the error change sequence, the square value of the error change value for each instruction is calculated, and all square values ​​are accumulated and averaged to obtain the average value of the square of the error fluctuation slope, which is used to evaluate the severity of the control error change over time.

[0060] Set a dynamic threshold, obtain continuous segments in the error sequence that are greater than the dynamic threshold, and count the time required for each segment to fall back from the maximum error to below the dynamic threshold. Record this as the recovery time. Calculate the average recovery time by averaging all recovery times in the error sequence.

[0061] Dynamic threshold refers to an error judgment threshold that is dynamically calculated based on the overall fluctuation characteristics of the control error sequence, obtained through an adaptive threshold method, and combined with the current assembly state or operating conditions. It is used to identify significant error deviation segments.

[0062] The average value of the square of the error fluctuation slope is multiplied by the average recovery time, and the square root of the product is applied to obtain the driving stability margin influence coefficient.

[0063] The average of the squared slope of the error fluctuation reflects the severity of the control error change with the command, while the mean recovery time measures the system's ability to recover from a high-error state to a steady state. The product of these two factors, after square root processing, can effectively reveal the potential instability risks the system may face under high-speed or high-frequency control conditions. This facilitates dynamic evaluation of actuator stability during actual assembly, thus providing a reliable basis for the adaptive adjustment of the speed regulation drive index.

[0064] In this embodiment, it should be specifically explained that the step of determining whether the current assembly speed needs to be adjusted based on the speed adjustment drive index is as follows:

[0065] The speed adjustment driving index is compared with the adjustment threshold. If the speed adjustment driving index is greater than the adjustment threshold, the current assembly speed is determined to need adjustment; if the speed adjustment driving index is less than or equal to the adjustment threshold, the current assembly speed is determined not to need adjustment. It should be noted that the adjustment threshold is obtained through an adaptive threshold method, which is a method that dynamically calculates the adjustment threshold based on historical assembly data and the current operating state, used to replace the insufficient adaptability of fixed empirical thresholds. By maintaining the speed adjustment driving index sequence within recent assembly cycles in real time, and combining the assembly action qualification rate and assembly image detection results for each cycle, a mapping relationship between assembly state and index changes is constructed. Based on this, statistical analysis methods are used to extract the boundary index value that can effectively distinguish between "adjustment-required state" and "adjustment-free state," which serves as the speed adjustment judgment threshold under the current assembly environment.

[0066] In this embodiment, it should be specifically explained that the steps for obtaining the actual assembly speed are as follows:

[0067] The speed adjustment factor is obtained by calculating the ratio of the adjustment threshold to the speed adjustment drive index.

[0068] The actual assembly speed is obtained by multiplying the speed adjustment factor by the current assembly speed.

[0069] The higher the speed regulation drive index, the worse the stability of the assembly system. If the index exceeds the adjustment threshold, it is determined that the current assembly speed is too fast. In order to avoid positioning deviation, increased vibration or drive instability caused by excessive speed, the speed should be reduced to ensure the safe and reliable assembly process.

[0070] The workpiece assembly module is used to assemble docked workpieces according to the actual assembly speed using end effectors of a robotic arm (including but not limited to grippers, suction devices, and screw tightening devices).

[0071] During the assembly process, the sensor monitoring module collects sensor data of the workpiece in real time through the sensor group. The sensor data includes assembly torque, clamping pressure and displacement changes. The assembly action qualification is evaluated based on the sensor data. The assembly action qualification is used to determine whether the assembly action is qualified. If the assembly action is determined to be unqualified, the subsequent operation is immediately interrupted and the sensor monitoring results are transmitted to the early warning module.

[0072] It should be noted that the sensor group includes force sensors, torque sensors, and position sensors, etc.

[0073] It should be noted that the sensor data is collected based on a preset assembly reference point, which corresponds to a key part of the target workpiece structure used for motion evaluation, to ensure that the collected data has a unified spatial reference and judgment basis.

[0074] In this embodiment, it should be specifically explained that the steps for obtaining the assembly action qualification are as follows:

[0075] During the assembly process, the assembly torque, clamping pressure, and displacement change of each sampling point are collected in real time at each sampling point.

[0076] Obtain the standard values ​​of assembly torque, clamping pressure, and displacement change. Calculate the torque deviation based on the assembly torque at each sampling point and the standard value of the assembly torque. The specific steps are as follows:

[0077] ;

[0078] In the formula, This is expressed as torque deviation, where n is the number of sampling points. Let the assembly torque at the i-th sampling point be denoted as . This is expressed as the standard value of the assembly torque;

[0079] The clamping pressure deviation is calculated based on the clamping pressure at each sampling point and the standard clamping pressure value. The specific steps for obtaining this deviation are as follows:

[0080] ;

[0081] In the formula, This is expressed as clamping pressure deviation. Let be the clamping pressure at the i-th sampling point. This is expressed as the standard value of clamping pressure;

[0082] The displacement deviation is calculated based on the displacement change and the standard value of displacement change at each sampling point. The specific steps for obtaining this deviation are as follows:

[0083] ;

[0084] In the formula, This is expressed as the displacement variation deviation. This is represented as the displacement change at the i-th sampling point. Represented as the standard value of displacement change;

[0085] The assembly operation qualification is calculated based on the torque deviation, clamping pressure deviation, and displacement change standard value. The specific steps are as follows:

[0086] ;

[0087] In the formula, This represents the qualification rate of the assembly process. This is expressed as torque deviation. This is expressed as clamping pressure deviation. It is expressed as the displacement change deviation. The cube root operation can balance the mutual influence of the three deviations and avoid the calculation imbalance caused by the single extreme value; taking the reciprocal structure ensures that the smaller the deviation, the higher the qualification.

[0088] In this embodiment, it should be specifically explained that the steps for determining whether an assembly action is qualified based on the assembly action qualification rate are as follows:

[0089] The assembly action pass rate is compared with the action pass threshold. If the assembly action pass rate is greater than or equal to the action pass threshold, the assembly action is judged to be qualified; if the assembly action pass rate is less than the action pass threshold, the assembly action is judged to be unqualified. The action pass threshold can be obtained by the adaptive threshold method.

[0090] If the assembly action is deemed qualified, the visual inspection module acquires image data of the assembled workpiece through an image acquisition device. The image data includes appearance features and position coordinates. It obtains a standard model of workpiece assembly and compares the image data of the assembled workpiece with the standard model of workpiece assembly to determine whether the assembled workpiece meets the visual inspection requirements. If the assembled workpiece does not meet the visual inspection requirements, the visual inspection result is transmitted to the early warning module.

[0091] The workpiece assembly standard model is a pre-established benchmark visual model of the completed workpiece assembly state. It is used to characterize the geometric structure, key feature point positions, edge contours, component docking relationships, and other image information that the target workpiece should possess after assembly. This standard model can be constructed by image acquisition and processing of historical qualified assembly samples, or by rendering and converting a CAD 3D model. After feature calibration and error tolerance setting, it forms a standard reference template that can be used for comparison of assembly results.

[0092] In this embodiment, it should be specifically explained that the step of comparing the image data of the assembled workpiece with the standard model of the workpiece assembly to determine whether the assembled workpiece meets the visual inspection requirements is as follows:

[0093] Acquire image data of assembled workpieces, and extract edge contours and key structural feature points from the current workpiece image using edge detection and contour extraction algorithms. This set is denoted as the current feature set. Extract the edge contours and key feature points of the workpiece assembly standard model, and denote them as the standard feature set. ;

[0094] It should be noted that the use of edge detection and contour extraction algorithms to extract the edge contours and key structural feature points of the current workpiece image is an existing technology, and this embodiment will not provide a detailed description of its specific steps.

[0095] Each key feature point in the current feature set is compared with each key feature point in the standard feature set one by one. The Euclidean distance between each corresponding key feature point is calculated and the mean is calculated to obtain the feature matching deviation value.

[0096] Obtain the center coordinates of the current workpiece image and the center coordinates of the workpiece assembly standard model, and denot them as the current center coordinates and the standard center coordinates, respectively. Calculate the Euclidean distance between the current center coordinates and the standard center coordinates to obtain the image coordinate deviation value.

[0097] The square root of the product of the feature matching deviation value and the image coordinate deviation value is taken, and the reciprocal of the product is used as the image qualification score.

[0098] The image passability is compared with the image passability threshold. If the image passability is greater than or equal to the image passability threshold, the assembled workpiece is determined to meet the visual inspection requirements. If the image passability is less than the image passability threshold, the assembled workpiece is determined to not meet the visual inspection requirements. The image passability threshold can be obtained by the adaptive threshold method.

[0099] Edge detection algorithms are used to extract regions in an image that exhibit significant grayscale changes, typically representing object contours or structural boundaries. In this embodiment, methods such as the Sobel or Canny operators are employed to identify the location distribution of edge points by calculating the gradient changes of image pixels in the horizontal and vertical directions, thereby achieving preliminary extraction of the key geometric contours of the assembled workpiece.

[0100] Contour extraction algorithms are used to track and connect continuous edge points based on edge detection results, forming closed or open geometric contour curve structures. In this embodiment, a contour tracking algorithm is combined to extract the complete structure of the workpiece boundary.

[0101] The data management and storage module stores and manages the workpiece information, sensor data, and image data during the assembly process if the assembled workpiece meets the visual inspection requirements. It establishes an assembly history record through a database, supports data traceability and statistical analysis, and provides a data foundation for subsequent optimization of assembly processes and improvement of control strategies.

[0102] The early warning module will issue an assembly action non-compliance warning if it receives sensor monitoring results, indicating that there is an abnormality in the assembly action of the current workpiece; if it receives visual inspection results, it will issue an assembly image non-compliance warning, indicating that there is an abnormality in the assembly image of the current workpiece.

[0103] By integrating the anomaly feedback results from the sensor monitoring module and the vision inspection module, real-time classification, identification, and response to assembly action anomalies and assembly image anomalies are achieved. On one hand, if the sensor monitoring results indicate unqualified assembly actions, the early warning module can immediately issue an assembly action anomaly warning to prevent abnormal operations from continuing. On the other hand, if the vision inspection results indicate unqualified images, it can also promptly issue image anomaly warnings to prevent workpieces with structural deviations or positioning errors from entering subsequent processes. This module implements a dual-check mechanism for the physical assembly process and the vision inspection stage, significantly improving the system's response speed to abnormal events, processing accuracy, and overall robustness of the assembly process.

[0104] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

[0105] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. An intelligent assembly system for mechanical automation, characterized in that, The system includes; The workpiece supply and positioning module is used to acquire the required workpiece information, transport the workpiece according to the workpiece information, collect the workpiece size and position parameters in real time, and use the positioning device to place the workpiece at the assembly station according to the workpiece size and position parameters. The robotic arm docking module is used to move and dock workpieces in the assembly station according to control commands; The assembly speed adjustment module is used to collect speed rationality evaluation parameters for mechanical assembly. The speed rationality evaluation parameters include the position coordinates, three-axis acceleration, and angle of the end effector of the robotic arm. The speed adjustment drive index is obtained based on the speed rationality evaluation parameters. The speed adjustment drive index is used to determine whether the current assembly speed needs to be adjusted. If the current assembly speed needs to be adjusted, the current assembly speed is adjusted according to the speed adjustment drive index to obtain the actual assembly speed. The workpiece assembly module is used to assemble docked workpieces according to the actual assembly speed via the end effector of the robotic arm; During the assembly process, the sensor monitoring module collects sensor data of the workpiece in real time through the sensor group. The sensor data includes assembly torque, clamping pressure and displacement changes. The assembly action qualification is evaluated based on the sensor data. The assembly action qualification is used to determine whether the assembly action is qualified. If the assembly action is determined to be unqualified, the sensor monitoring result is transmitted to the early warning module. If the assembly action is deemed qualified, the visual inspection module acquires image data of the assembled workpiece through an image acquisition device. The image data includes appearance features and position coordinates. It obtains a standard model of workpiece assembly and compares the image data of the assembled workpiece with the standard model of workpiece assembly to determine whether the assembled workpiece meets the visual inspection requirements. If the assembled workpiece does not meet the visual inspection requirements, the visual inspection result is transmitted to the early warning module. The data management and storage module stores and manages the workpiece information, sensor data, and image data during the assembly process if the assembled workpiece meets the visual inspection requirements. If the early warning module receives sensor monitoring results, it will issue an early warning for unqualified assembly actions. If a visual inspection result is received, an assembly image non-compliance warning will be issued; The steps for obtaining the speed regulation drive index are as follows: Set a detection time period, obtain the position coordinates of the robot arm end effector within the historical detection time period, and evaluate the regression positioning stability influence coefficient based on the position coordinates of the robot arm end effector. The three-axis acceleration of the robotic arm end effector during the historical detection period is obtained, and the vibration disturbance sensitivity influence coefficient is evaluated based on the three-axis acceleration of the robotic arm end effector. The angle of the end effector of the robotic arm during the historical detection period is obtained, and the drive stability margin influence coefficient is evaluated based on the angle of the end effector of the robotic arm. The regression positioning stability influence coefficient, vibration disturbance sensitivity influence coefficient, and drive stability margin influence coefficient are normalized. The speed regulation drive index is then evaluated based on these normalized coefficients. The specific steps for obtaining the index are as follows: ； In the formula, This is expressed as the speed regulation drive index. This is expressed as the normalized regression positioning stability impact coefficient. This is expressed as the normalized vibration disturbance sensitivity influence coefficient. This is expressed as the normalized driving stability margin influence coefficient. , , These are the weighting coefficients of the normalized regression positioning stability influence coefficient, the normalized vibration disturbance sensitivity influence coefficient, and the normalized drive stability margin influence coefficient.

2. The intelligent assembly system for mechanical automation according to claim 1, characterized in that, The steps for obtaining the regression positioning stability influence coefficient are as follows: Obtain the target position coordinates of the robotic arm end effector for each regression positioning operation within the historical detection period, and form a reference point sequence. The data in the reference point sequence are recorded as reference points. The actual position coordinates of the robotic arm end effector for each regression positioning operation within the historical detection period are obtained to form an actual regression positioning sequence. The data in the actual regression positioning sequence are recorded as actual points. For each regression, the three-dimensional Euclidean distance between the corresponding reference point and the actual point is calculated to obtain the positioning deviation value; By iterating through each regression, a positioning deviation sequence is obtained. The maximum value in the positioning deviation sequence is extracted and recorded as the maximum deviation value. The standard deviation of the data in the positioning deviation sequence is calculated to obtain the deviation fluctuation. The maximum deviation value and the deviation fluctuation are summed to obtain the regression positioning stability influence coefficient.

3. The intelligent assembly system for mechanical automation according to claim 1, characterized in that, The steps for obtaining the vibration disturbance sensitivity influence coefficient are as follows: Set the sampling frequency, and obtain the three-axis vibration acceleration data of the robotic arm end effector in a stable standby state when no assembly operation is performed, based on the sampling frequency. Record it as the reference vibration sequence, with the three axes being the X-axis, Y-axis and Z-axis. Based on the sampling frequency, the three-axis acceleration data of the end effector of the robotic arm during the actual assembly process within the historical detection time period are obtained and recorded as the actual vibration sequence. The root mean square (RMS) values ​​are calculated for the actual vibration sequence and the reference vibration sequence in each direction, resulting in the actual RMS values ​​for the X-axis, Y-axis, and Z-axis, as well as the reference RMS values ​​for the X-axis, Y-axis, and Z-axis. The ratios of the actual root mean square value to the reference root mean square value in each direction are calculated to obtain the vibration amplitude ratios in the X-axis, Y-axis, and Z-axis directions. Short-time Fourier transform is performed on the triaxial acceleration data in the actual vibration sequence to extract the triaxial high-frequency disturbance energy, which includes X-axis high-frequency disturbance energy, Y-axis high-frequency disturbance energy and Z-axis high-frequency disturbance energy. The triaxial reference high-frequency energy is obtained, and the ratio of the triaxial high-frequency disturbance energy to the corresponding triaxial reference high-frequency energy is calculated to obtain the X-axis energy surge ratio, Y-axis energy surge ratio and Z-axis energy surge ratio. The vibration disturbance sensitivity influence coefficient is calculated based on the vibration amplitude ratios in the X-axis, Y-axis, and Z-axis directions, as well as the X-axis energy surge ratio, Y-axis energy surge ratio, and Z-axis energy surge ratio.

4. The intelligent assembly system for mechanical automation according to claim 1, characterized in that: The steps for obtaining the driving stability margin influence coefficient are as follows: Obtain the control command angle sequence and actual feedback angle sequence of the robotic arm end effector within the historical detection period; For each instruction, the difference between the control instruction angle value and the actual feedback angle value is calculated, and the absolute value is taken to obtain the control delay error. All instructions within the historical detection time period are traversed to obtain the error sequence within the historical detection time period. Perform first-order difference on the error sequence, calculate the error change value between adjacent instructions, construct the error change sequence, calculate the square value of the error change value for each instruction based on the error change sequence, and sum and average all the square values ​​to obtain the average value of the square of the error fluctuation slope. Set a dynamic threshold, obtain continuous segments in the error sequence that are greater than the dynamic threshold, and count the time required for each segment to fall back from the maximum error to below the dynamic threshold. Record this as the recovery time. Calculate the average recovery time by averaging all recovery times in the error sequence. The average value of the square of the error fluctuation slope is multiplied by the average recovery time, and the square root of the product is applied to obtain the driving stability margin influence coefficient.

5. The intelligent assembly system for mechanical automation according to claim 1, characterized in that: The step of determining whether the current assembly speed needs adjustment based on the speed adjustment drive index is as follows: The speed adjustment drive index is compared with the adjustment threshold. If the speed adjustment drive index is greater than the adjustment threshold, it is determined that the current assembly speed needs to be adjusted. If the speed adjustment drive index is less than or equal to the adjustment threshold, then the current assembly speed is determined to be unnecessary to adjust.

6. The intelligent assembly system for mechanical automation according to claim 5, characterized in that: The steps for obtaining the actual assembly speed are as follows: The speed adjustment factor is obtained by calculating the ratio of the adjustment threshold to the speed adjustment drive index. The actual assembly speed is obtained by multiplying the speed adjustment factor by the current assembly speed.

7. The intelligent assembly system for mechanical automation according to claim 1, characterized in that: The steps for obtaining the assembly action pass rate are as follows: During the assembly process, the assembly torque, clamping pressure, and displacement change of each sampling point are collected in real time at each sampling point. Obtain the standard values ​​of assembly torque, clamping pressure, and displacement change, and calculate the torque deviation based on the assembly torque and the standard value of assembly torque at each sampling point; The clamping pressure deviation is calculated based on the clamping pressure at each sampling point and the standard value of clamping pressure. The displacement deviation is calculated based on the displacement change and the standard value of displacement change at each sampling point. The assembly operation qualification rate is calculated based on the torque deviation, clamping pressure deviation, and displacement change standard value.

8. The intelligent assembly system for mechanical automation according to claim 1, characterized in that: The steps for determining whether an assembly action is qualified based on the qualification of the assembly action are as follows: The assembly action pass rate is compared with the action pass threshold. If the assembly action pass rate is greater than or equal to the action pass threshold, the assembly action is judged to be qualified; if the assembly action pass rate is less than the action pass threshold, the assembly action is judged to be unqualified.

9. The intelligent assembly system for mechanical automation according to claim 1, characterized in that: The step of comparing the image data of the assembled workpiece with the standard model of the workpiece assembly to determine whether the assembled workpiece meets the visual inspection requirements is as follows: The image data of the assembled workpiece is acquired. The edge contour and key structural feature points of the current workpiece image are extracted through edge detection and contour extraction algorithms and recorded as the current feature set. The edge contour and key feature points of the workpiece assembly standard model are extracted and recorded as the standard feature set. Each key feature point in the current feature set is compared with each key feature point in the standard feature set one by one. The Euclidean distance between each corresponding key feature point is calculated and the mean is calculated to obtain the feature matching deviation value. Obtain the center coordinates of the current workpiece image and the center coordinates of the workpiece assembly standard model, and denot them as the current center coordinates and the standard center coordinates, respectively. Calculate the Euclidean distance between the current center coordinates and the standard center coordinates to obtain the image coordinate deviation value. The square root of the product of the feature matching deviation value and the image coordinate deviation value is taken, and the reciprocal of the product is used as the image qualification score. The image pass rate is compared with the image pass rate threshold. If the image pass rate is greater than or equal to the image pass rate threshold, the assembled workpiece is determined to meet the visual inspection requirements; if the image pass rate is less than the image pass rate threshold, the assembled workpiece is determined to not meet the visual inspection requirements.

Images