A multi-axis industrial robot motion trajectory optimization method and system

Abstract

The present invention relates to the technical field of data processing, and specifically relates to a method and system for optimizing the running trajectory of a multi-axis industrial robot. For the original vibration signal of each axis in the multi-axis system, based on the independent component analysis method, the error vibration signal and the complexity factor of the motion posture of the multi-axis industrial robot are analyzed in combination with the vibration signal of the real motion. Further, the error vibration intensity coefficient of the error vibration signal is judged, and according to the error vibration intensity coefficient of the error vibration signal, it is fed back to the PID controller of the multi-axis industrial robot to adjust or correct its proportional gain component, and then it is fed back to the multi-axis linkage interpolation control algorithm to adjust the control of the corresponding axis, effectively correcting the influence factors of the non-linear characteristics on the running deviation of the multi-axis industrial robot, achieving a more accurate and reliable control of the running trajectory of the multi-axis industrial robot, realizing the optimization of the running trajectory of the multi-axis industrial robot, and improving the working efficiency of the multi-axis industrial robot.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and particularly relates to a method and system for optimizing the running trajectory of a multi-axis industrial robot. Background Art

[0002] A multi-axis industrial robot is a robot system with multiple degrees of freedom (axes), and each axis can be independently controlled for movement. Such a robot system is widely used in the industrial production field and can perform various complex tasks, such as assembly, welding, handling, machining, etc., adapting to various complex working environments and task requirements, thereby reducing labor costs and improving production efficiency and quality.

[0003] In complex working environments and task requirements, due to the influence factors of non-linear characteristics such as movement speed, inertia, friction, elasticity, etc., a certain degree of error will occur during high-speed movement of a multi-axis industrial robot. When traditional multi-axis linkage interpolation control algorithms correct errors, since the causes of the errors are non-linear, it is impossible to correct the errors in a timely and accurate manner, resulting in the continuous accumulation of these errors, and ultimately leading to the robot being unable to accurately reach the predetermined position, thereby reducing work efficiency and accuracy. Summary of the Invention

[0004] In order to solve the problem that the running trajectory of a multi-axis industrial robot cannot be accurately and reliably controlled at present, the present invention provides a method and system for optimizing the running trajectory of a multi-axis industrial robot.

[0005] A method for optimizing the running trajectory of a multi-axis industrial robot includes:

[0006] Real-time collecting the running monitoring data of the multi-axis industrial robot at different target times, where the running monitoring data includes the original vibration signals of each axis of the multi-axis industrial robot;

[0007] Obtaining the historical data set of the historical operation of the multi-axis industrial robot;

[0008] Analyzing the complexity factor of the motion posture of the multi-axis industrial robot at different target times according to the historical data set;

[0009] For different target times, analyzing the corresponding original vibration signal to obtain an error vibration signal;

[0010] Calculating the error vibration intensity coefficient at different target times according to the error vibration signal and the complexity factor;

[0011] Calculating the correction proportional gain component according to the error vibration intensity coefficients of all target times;

[0012] Optimize the running trajectory of the robot according to the proportional gain component.

[0013] The historical data set includes the historical power and historical speed of each axis. The historical power under the same motion posture is the same type of power data, and the historical speed under the same motion posture is the same type of speed data.

[0014] Analyze the complexity factor of the motion postures of the multi-axis industrial robot at different target times according to the historical data set, specifically including:

[0015] Calculate the information entropy of each axis based on the probabilities of the same type of power data and the same type of speed data of a single axis appearing in the historical data set, so as to obtain the information entropy of each axis of the industrial robot under the corresponding machine posture.

[0016] Calculate the joint entropy between each pair of axes based on the same type of power data of each pair of axes and the same type of speed data of each pair of axes, so as to obtain the joint entropy of each pair of axes of the industrial robot under the corresponding machine posture, where any two axes form a pair of axes.

[0017] Obtain the motion posture at a target time.

[0018] Calculate the complexity factor of the motion posture at the target time according to the information entropy and joint entropy corresponding to the motion posture, so as to obtain the complexity factors of the motion postures at different target times.

[0019] Among them, analyzing the corresponding original vibration signal to obtain the error vibration signal specifically includes:

[0020] Use the ICA algorithm to decompose the original vibration signal to obtain several decomposition results. Each decomposition result includes a first component signal and a second component signal.

[0021] Use the polynomial fitting method to fit the first component signal in each decomposition result into a function curve respectively, and obtain the probability distribution of all amplitude types of the first component signal in each decomposition result.

[0022] Calculate the kurtosis coefficient and skewness coefficient of the amplitude type distribution of the first component signal in each decomposition result according to the probability distribution of all amplitude types.

[0023] Construct the original vibration signal decomposition function based on the kurtosis coefficient and skewness coefficient.

[0024] Calculate the function output values of all decomposition results based on the original vibration signal decomposition function, obtain all output values of the original vibration signal decomposition function, select the decomposition result corresponding to the maximum output value as the optimal decomposition result of the original vibration signal, and the second signal component corresponding to the optimal decomposition result is the error vibration signal.

[0025] Among them, the error vibration signal includes multiple error vibration component signals;

[0026] Calculating the error vibration intensity coefficient at different target times according to the error vibration signal and the complexity factor specifically includes:

[0027] For a target time, obtain the signal maximum points of all error vibration component signals;

[0028] Calculate the error vibration delay characteristic factor generated by each signal maximum point;

[0029] Based on the error vibration delay characteristic factor at the same target time and the complexity factor of the corresponding motion posture, calculate the error vibration intensity coefficient of the vibration error signal at this target time, so as to obtain the error vibration intensity coefficients at different target times.

[0030] Among them, calculating the correction proportional gain component according to the error vibration intensity coefficients at all target times specifically includes:

[0031] Normalize the error vibration intensity coefficients at all target times;

[0032] Based on the normalized error vibration intensity coefficient and the original proportional gain component, calculate the correction proportional gain component.

[0033] A multi-axis industrial robot operation trajectory optimization system includes:

[0034] An acquisition module for real-time acquisition of the operation monitoring data of the multi-axis industrial robot at different target times, and the operation monitoring data includes the original vibration signal of each axis of the multi-axis industrial robot;

[0035] An acquisition module for obtaining the historical data set of the historical operation of the multi-axis industrial robot;

[0036] A first analysis module for analyzing the complexity factor of the motion posture of the multi-axis industrial robot at different target times according to the historical data set;

[0037] A second analysis module for analyzing the corresponding original vibration signal to obtain an error vibration signal for different target times;

[0038] A first calculation module, configured to calculate error vibration intensity coefficients at different target times according to the error vibration signal and the complexity factor;

[0039] A second calculation module, configured to calculate a correction proportional gain component according to the error vibration intensity coefficients at all target times;

[0040] An optimization module, configured to optimize the running trajectory of the robot according to the proportional gain component.

[0041] Wherein, the historical data set includes the historical power and historical speed of each axis. The historical powers in the same motion posture are of the same type of power data, and the historical speeds in the same motion posture are of the same type of speed data;

[0042] The first analysis module specifically includes:

[0043] A first calculation unit, configured to calculate the information entropy of an axis based on the probabilities of the same type of power data and the same type of speed data of a single axis appearing in the historical data set, so as to obtain the information entropy of each axis of the industrial robot in the corresponding machine posture;

[0044] A second calculation unit, configured to calculate the joint entropy between each pair of axes respectively based on the same type of power data of each pair of axes and the same type of speed data of each pair of axes, so as to obtain the joint entropy of each pair of axes of the industrial robot in the corresponding machine posture, wherein any two axes form a pair of axes;

[0045] A first acquisition unit, configured to acquire a motion posture at a target time;

[0046] A third calculation unit, configured to calculate the complexity factor of the motion posture at the target time according to the information entropy and the joint entropy corresponding to the motion posture, so as to obtain the complexity factors of the motion postures at different target times.

[0047] Wherein, the second analysis module specifically includes:

[0048] A decomposition unit, configured to decompose the original vibration signal using the ICA algorithm to obtain a plurality of decomposition results, wherein each decomposition result includes a first component signal and a second component signal;

[0049] A second acquisition unit, configured to respectively fit the first component signal in each decomposition result into a function curve using a polynomial fitting method, and obtain all the amplitude type distribution probabilities of the first component signal in each decomposition result;

[0050] A fourth calculation unit, configured to calculate the kurtosis coefficient and skewness coefficient of the amplitude type distribution of the first component signal in each decomposition result according to all the amplitude type distribution probabilities;

[0051] A construction unit, configured to construct an original vibration signal decomposition function based on the kurtosis coefficient and the skewness coefficient;

[0052] A third acquisition unit, configured to calculate function output values of all decomposition results based on the original vibration signal decomposition function, obtain all output values of the original vibration signal decomposition function, and select the decomposition result corresponding to the maximum output value as the optimal decomposition result of the original vibration signal, where the second signal component corresponding to the optimal decomposition result is the error vibration signal.

[0053] Wherein, the error vibration signal includes multiple error vibration component signals;

[0054] The first calculation module specifically includes:

[0055] A fourth acquisition unit, configured to obtain signal maximum points of all error vibration component signals at a target moment;

[0056] A fifth calculation unit, configured to calculate error vibration delay characteristic factors generated by each signal maximum point;

[0057] A sixth calculation unit, configured to calculate an error vibration intensity coefficient of the vibration error signal at the target moment based on the error vibration delay characteristic factor and the complexity factor of the corresponding motion posture at the same target moment, so as to obtain error vibration intensity coefficients at different target moments.

[0058] Wherein, the second calculation module specifically includes:

[0059] A normalization unit, configured to perform normalization processing on error vibration intensity coefficients at all target moments;

[0060] A fifth calculation unit, configured to calculate a corrected proportional gain component based on the normalized error vibration intensity coefficient and the original proportional gain component.

[0061] The present invention has the following beneficial effects: For the original vibration signal of each axis in a multi-axis system, the present invention combines the error vibration signal and the complexity factor of the motion posture based on the independent component analysis method, further judges the error vibration intensity coefficient of the error vibration signal, and feeds back the error vibration intensity coefficient of the error vibration signal to the PID controller of the multi-axis industrial robot to adjust or correct its proportional gain component, thereby feeding back to the multi-axis linkage interpolation control algorithm to adjust the control of the corresponding axis, effectively correcting the influence factors of the nonlinear characteristics on the operation deviation of the multi-axis industrial robot, achieving more accurate and reliable control of the operation trajectory of the multi-axis industrial robot, realizing the optimization of the operation trajectory of the multi-axis industrial robot, and improving the working efficiency of the multi-axis industrial robot. Description of the Drawings

[0062] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the following will briefly introduce the accompanying drawings required in the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other accompanying drawings can be obtained based on these drawings.

[0063] Figure 1 Schematic flow chart of a method for optimizing the running trajectory of a multi-axis industrial robot provided by an embodiment of the present invention;

[0064] Figure 2 Schematic diagram of the first sub-process of a method for optimizing the running trajectory of a multi-axis industrial robot provided by an embodiment of the present invention;

[0065] Figure 3 Schematic diagram of the second sub-process of a method for optimizing the running trajectory of a multi-axis industrial robot provided by an embodiment of the present invention;

[0066] Figure 4 Schematic diagram of the third sub-process of a method for optimizing the running trajectory of a multi-axis industrial robot provided by an embodiment of the present invention;

[0067] Figure 5 Schematic diagram of the fourth sub-process of a method for optimizing the running trajectory of a multi-axis industrial robot provided by an embodiment of the present invention;

[0068] Figure 6 Schematic diagram of a system for optimizing the running trajectory of a multi-axis industrial robot provided by an embodiment of the present invention;

[0069] Figure 7 Schematic diagram of the first sub-system of a system for optimizing the running trajectory of a multi-axis industrial robot provided by an embodiment of the present invention;

[0070] Figure 8 Schematic diagram of the second sub-system of a system for optimizing the running trajectory of a multi-axis industrial robot provided by an embodiment of the present invention;

[0071] Figure 9 Schematic diagram of the third sub-system of a system for optimizing the running trajectory of a multi-axis industrial robot provided by an embodiment of the present invention;

[0072] Figure 10 Schematic diagram of the fourth sub-system of a system for optimizing the running trajectory of a multi-axis industrial robot provided by an embodiment of the present invention. Detailed implementation manners

[0073] In order to further elaborate on the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following specifically describes, in conjunction with the accompanying drawings and preferred embodiments, a multi-axis industrial robot running trajectory optimization method and system proposed according to the present invention, including its specific implementation manners, structures, features and effects. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.

[0074] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which the present invention belongs.

[0075] The following specifically describes the specific solution of a multi-axis industrial robot running trajectory optimization method and system provided by the present invention in conjunction with the accompanying drawings.

[0076] For a multi-axis industrial robot system, each axis in its multi-axis system can work independently or collaboratively. Therefore, each axis has an independent motor, driver and sensor. Regarding the independent motor, driver and sensor of each axis as a sub-controller of this multi-axis system, which is responsible for managing and executing the motion control tasks related to this axis. The central controller is the main control unit of the multi-axis industrial robot system, responsible for coordinating the motions of each axis. This control unit receives high-level instructions and decomposes them into control commands for individual axes. The multi-axis linkage interpolation control algorithm coordinates the motions between axes to minimize the accumulation of out-of-synchronization and errors between axes as much as possible to achieve precise motion of the robot. However, in an actual system, due to the influence of dynamic characteristics (such as inertia, friction, elasticity, etc.) and non-linear effects (such as friction force, coupling effect, etc.), certain errors will still occur. Therefore, it is necessary to optimize the feedback logic by combining a PID control sensor to more precisely monitor and correct errors in the robot system in real time, and achieve higher-quality and more reliable motion control.

[0077] Please refer to Figure 1 , which shows a flowchart of a multi-axis industrial robot running trajectory optimization method provided by an embodiment of the present invention. The method for detecting the connecting wire of the indoor unit of a fixed-frequency air conditioner includes steps S101 - S107:

[0078] S101. Real-time collect the running monitoring data of the multi-axis industrial robot at different target moments, and the running monitoring data includes the original vibration signals of each axis of the multi-axis industrial robot.

[0079] Install an integrated sensor in the multi-axis industrial robot system to collect the operation monitoring data of the motor shafts during the real-time movement of the robot, including the original vibration signals of each axis in the multi-axis motors, and perform preprocessing, including but not limited to smoothing, denoising, etc., to ensure the accuracy of subsequent analysis.

[0080] S102. Obtain the historical data set of the historical operation of the multi-axis industrial robot.

[0081] At the same time, obtain the input power and speed of each axis in the multi-axis system of the robot, and the historical feature data under different postures. These historical feature data form the historical data set.

[0082] S103. Analyze the complexity factor of the movement postures of the multi-axis industrial robot at different target times according to the historical data set.

[0083] Errors in the movement posture may lead to a decrease in the accuracy during its positioning and task execution, thus affecting its accuracy. Especially for applications that require high-precision operations (such as assembly, welding, etc.), the posture errors at different times may have different impacts on the final quality effect of the product.

[0084] When the industrial robot needs to perform complex posture transformations, more precise control is usually required to ensure the desired posture is achieved. This means that even small posture errors may have a greater impact on task execution, so more frequent and refined error corrections are needed.

[0085] Specifically, the complexity of the movement postures at different times can be quantified as the coordination between each axis in the multi-axis system. When the coordination is greater, the posture complexity is higher. At any time, the posture of the robot is composed of the power and speed of each axis, and the coordination between the axes is crucial for achieving a specific posture. In this embodiment, the complexity of the operation postures at different times is quantitatively characterized by analyzing the performance of the movement postures at different times in the historical data.

[0086] The historical data set includes the historical power and historical speed of each axis. The historical power under the same movement posture is the same type of power data, and the historical speed under the same movement posture is the same type of speed data.

[0087] Further, in one embodiment, as Figure 2 shown, step S103 specifically includes the following steps:

[0088] S1031. Calculate the information entropy of each axis based on the probability of the same type of power data and the same type of speed data of a single axis appearing in the historical data set, so as to obtain the information entropy of each axis of the industrial robot under the corresponding machine posture.

[0089] S1032. Calculate the joint entropy between each pair of axes based on the same-type power data and the same-type speed data of each pair of axes, so as to obtain the joint entropy of each pair of axes of the industrial robot in the corresponding machine posture, where any two axes form a pair of axes.

[0090] S1033. Obtain the motion posture at a target moment.

[0091] S1034. Calculate the complexity factor of the motion posture at the target moment according to the information entropy and the joint entropy corresponding to the motion posture, so as to obtain the complexity factors of the motion postures at different target moments.

[0092] For the power and speed historical data of each axis, obtain the probability that the power and speed data in the same motion posture in its historical data appear in the overall historical data set, and then obtain the information entropy of the power and speed data of each axis. At the same time, for the historical data between each pair of axes in the multi-axis system, calculate its joint entropy. Subsequently, the information entropy is used to correct the joint entropy to obtain the final complexity factor of the motion posture of the overall multi-axis system.

[0093] Specifically, the calculation formula for the posture complexity factor of the industrial robot at different target moments is:

[0094]

[0095] In the formula, i represents the motion posture of the target robot at the i-th moment, and also represents the i-th motion posture. N represents the total number of axes in the multi-axis system of the industrial robot, and n represents the n-th axis in the multi-axis system of the industrial robot. represents the information entropy of the n-th axis in the historical data in the posture same as the i-th motion posture. H represents the total number set of pairs formed by the n-th axis in the multi-axis system with any axis in the system. h represents the h-th pair of axis combinations in the H set. represents the joint entropy of the historical input data set of the h-th axis combination in the historical data in the posture same as the i-th motion posture. R i represents the complexity factor of the i-th motion posture.

[0096] In summary, It represents the product of the information entropy of the historical data set of the nth axis and the joint entropy of the historical data set of the hth axis combination in the same posture as the ith motion posture in the historical data. The correlation between it and other axes in the multi-axis system is corrected by the degree of disorder of the historical input data of a single axis to obtain the corrected correlation degree of each axis with other axes in the system. The larger this value is, the more concentrated the distribution range of the historical data of the nth axis is, indicating that the contribution degree of this axis in forming the robot motion posture is larger because its motion is more stable and controllable. Therefore, the weight of the correlation degree of this axis with other axes in the coordination of the overall multi-axis system is increased;

[0097] It represents the weighted average of the corrected correlation degree factors between all single axes and other axes in the multi-axis system of the industrial robot's target moment motion posture. The larger this value is, the stronger the correlation between the multi-axis systems in the industrial robot's target moment motion posture, indicating that the coordination factor between each axis in this multi-axis system is larger. Furthermore, the higher the posture complexity of the industrial robot in the ith motion posture, then the i larger the value of R.

[0098] S104. For different target moments, analyze the corresponding original vibration signals to obtain error vibration signals.

[0099] Independent Component Analysis (ICA) method is a common signal processing algorithm, which can be used to decompose multiple mixed signals and can effectively separate each component signal of the original signal. Each component corresponds to an independent source in the system.

[0100] In a multi-axis industrial robot system, the vibration signals generated during high-speed motion can be regarded as composite signals mixed by multiple sources, which contain both the electrical axis vibration signals generated by real motion and the error vibration signals generated by factors such as friction and inertia in the system (resulting from the mechanical structure of the robot, friction between connecting parts, and inertia of components). There are errors in using the vibration signal analysis results as feedback conditions for the multi-axis linkage interpolation control algorithm in the control system, resulting in a decrease in the accuracy of the robot's behavior trajectory.

[0101] Use an integrated sensor to obtain the vibration signals generated by the multi-axis system of the multi-axis industrial robot at different motion speeds, and use the ICA algorithm to decompose the original vibration signals to remove the low-frequency signals caused by non-high-speed motion, that is, vibration error signals.

[0102] In one embodiment, as Figure 3 shown, step S104 specifically includes the following steps:

[0103] S1041. Decompose the original vibration signal using the ICA algorithm to obtain several decomposition results. Each decomposition result includes a first component signal and a second component signal.

[0104] It should be noted that the ICA algorithm set in this application can obtain several signal decomposition results. There are two component signals in each decomposition result. The signal with a larger decomposition integral is regarded as the first component signal, and the remaining decomposed signals are regarded as the second component signals. In a multi-axis industrial robot system, the normal vibration signal should have a certain periodicity because when the industrial robot performs various operations, its mechanical components will generate regular vibrations, and the vibration signal of the real multi-axis system of the industrial robot is often non-Gaussian due to the influence of multiple vibration modes.

[0105] S1042. Use the polynomial fitting method to fit the first component signal in each decomposition result into a function curve, and obtain all the amplitude type distribution probabilities of the first component signal in each decomposition result.

[0106] S1043. Calculate the kurtosis coefficient and skewness coefficient of the amplitude type distribution of the first component signal in each decomposition result according to all the amplitude type distribution probabilities.

[0107] Specifically, for the first component signal obtained in any decomposition result, use the polynomial fitting method to fit it into a function curve, and at the same time obtain all the amplitude type distribution probabilities in the first component signal. Then, according to the known technology, the kurtosis coefficient and skewness coefficient of the amplitude type distribution of the first component signal can be obtained for subsequent analysis and screening of the decomposition results.

[0108] S1044. Construct an original vibration signal decomposition function based on the kurtosis coefficient and skewness coefficient.

[0109] Specifically, the formula of the original vibration signal decomposition function is:

[0110]

[0111] In the formula, K i represents the kurtosis coefficient of all amplitude type distributions in the first component signal of the multi-axis system in the i-th motion posture, U i represents the skewness coefficient of all amplitude type distributions in the first component signal of the multi-axis system in the i-th motion posture, T i represents the total number of signal sampling points in the first component signal of the multi-axis system in the i-th motion posture, t i represents the t-th signal sampling point in the first component signal of the multi-axis system in the i-th motion posture, f i(t) represents the fitting function of the first component signal of the multi-axis system in the i-th motion posture, f i (t + γ) represents the fitting time-delay function of the first component signal of the multi-axis system in the i-th motion posture, γ represents the time-delay amount, H i (t) represents the decomposition function of the original vibration signal of the multi-axis system in the i-th motion posture of the system. That is, t represents time, H i (t) represents the decomposition function of the original vibration signal corresponding to the motion posture with respect to time.

[0112] In summary, (K i ×U i ) represents the product of the kurtosis coefficient and the skewness coefficient of all amplitude types in the first component signal of the multi-axis system in the i-th motion posture. The larger this value, the sharper and more irregular the distribution of the signal amplitude, indicating that the amplitude distribution of the signal deviates more from the Gaussian distribution, that is, the greater the non-Gaussianity; represents the fitting function and the fitting time-delay function of the first component signal of the multi-axis system in the i-th motion posture. Calculate the product within the time range from (t = 1) to (t = T), and integrate all these products. This value is a specific value from 0 to 1. The closer it is to 1, the stronger the periodicity of the target signal. (It should be noted that since the period value, that is, the time-delay amount, of each signal cannot be determined here, γ here is an iterative value starting from 0, increasing by 1 each time, gradually approaching the true period and time-delay amount, so as to find the overall maximum value to end).

[0113] represents the sum of the non-Gaussianity and periodicity of the first component signal of the multi-axis system in the i-th motion posture. In the ICA analysis method, the worse the Gaussianity of the component signal, the better the decomposition result. At the same time, considering that the true vibration signal of the robot motion has periodicity compared with the error vibration signal, the Gaussianity of the component signal is corrected. The larger this value, the less Gaussian the amplitude distribution of the first component signal is and the more periodic it is. The greater the possibility that this first component signal is the true electric axis vibration signal during the robot motion, then H i (t) has a larger value.

[0114] S1045. Calculate the function output values of all decomposition results based on the original vibration signal decomposition function, obtain all output values of the original vibration signal decomposition function, and select the decomposition result corresponding to the maximum output value as the optimal decomposition result of the original vibration signal. The second signal component corresponding to the optimal decomposition result is the error vibration signal.

[0115] The function output values of all decomposition results are calculated through the above formula, and the decomposition result corresponding to the maximum value of the output value H(t) is selected as the optimal result. In this optimal result, its first component signal is regarded as the true electrical axis vibration signal during the movement of the robot, and the second component signal is regarded as the error vibration signal.

[0116] S105. Calculate the error vibration intensity coefficient at different target times according to the error vibration signal and the complexity factor.

[0117] The error vibration signal includes multiple error vibration component signals. The signal strength of the error vibration signal is judged according to the obtained error vibration signal. The strength of the error vibration signal determines the regulation coefficient of the control system. The specific signal strength of the error vibration signal can be represented by the characteristic that the amplitude of its delayed vibration signal is different from that of the normal vibration signal with a higher strength.

[0118] In one embodiment, as Figure 4 shown, step S105 specifically includes steps S1051 - S1053:

[0119] S1051. For a target time, obtain the signal maximum points of all error vibration component signals.

[0120] First, obtain the signal maximum points of all error vibration component signals. For all the signal maximum points in the obtained error vibration signal, taking any signal maximum point as the starting point, define a window area, and the size of the window area is set to 20 signal sampling points to facilitate the further analysis of the oscillation signal.

[0121] S1052. Calculate the error vibration delay characteristic factor generated by each signal maximum point.

[0122] S1053. Calculate the error vibration intensity coefficient of the vibration error signal at this target time based on the error vibration delay characteristic factor at the same target time and the complexity factor of the corresponding motion posture, so as to obtain the error vibration intensity coefficients at different target times.

[0123] The error vibration delay characteristic factor generated by the signal maximum point may represent the influence of factors such as friction and inertia at the signal maximum point on the error vibration of the multi-axis robot system. Combine the error vibration delay characteristic factor and the complexity factor of the motion posture to construct a specific formula for calculating the error vibration intensity coefficient of the error vibration signal. The formula is as follows:

[0124]

[0125] In the formula, R i represents the complexity factor at the i-th target time (i.e., the i-th motion posture), Mi represents the set of all maximum points in the multi-axis system error vibration signal in the \(i\)-th motion posture, \(m\) i represents the \(m\)-th maximum point in the multi-axis system error vibration signal in the \(i\)-th motion posture i th maximum point, \(E\) mi represents the slope at the \(m\)-th maximum point i th maximum point, \(A\) i represents the set of all sampling points in the window region corresponding to the \(m\)-th signal maximum point in the multi-axis system error vibration signal in the \(i\)-th motion posture. \(a\) represents the \(a\)-th signal maximum point in the \(A\)-th set i i represents the probability of occurrence of the signal amplitude corresponding to the \(a\)-th signal maximum point in the window region corresponding to the \(m\)-th signal maximum point in the multi-axis system error vibration signal in the \(i\)-th motion posture in the overall error vibration signal, \(Q\) i i represents the error vibration intensity coefficient of the multi-axis system error vibration signal in the \(i\)-th motion posture.

[0126] In summary, represents the product of the probability of occurrence of the signal amplitude corresponding to the \(a\)-th signal maximum point in the window region corresponding to the \(m\)-th signal maximum point in the multi-axis system error vibration signal in the \(i\)-th motion posture in the overall error vibration signal. The product result is the error vibration delay characteristic factor. The smaller this value, the lower the probability of occurrence of the signal amplitude in the subsequent window region at the \(m\)-th signal maximum point in the overall error vibration signal, and the less common the signal change in this window region. This may indicate that factors such as friction and inertia have a greater impact on the error vibration of the multi-axis robot system at this point; i i represents first correcting the slope by the error vibration delay characteristic factor of the maximum point, and then taking the weighted average of the corrected slopes of all maximum points in the error vibration signal (i.e., the characteristic factors of all maximum points in the signal). The larger this value, the more significant or intense the signal change rate at the maximum point in the entire error vibration signal, and a larger delay vibration is generated, indicating that the error vibration generated by the robot system during movement or operation is more obvious and intense.

[0127] represents the product of the posture complexity factor in the \(i\)-th motion posture and the error vibration delay characteristic factors of all maximum points in the signal. The larger this value, the higher the complexity of the robot's motion posture at the current moment, and when correcting the intensity coefficient of the error vibration signal, it is necessary to more effectively suppress the influence of errors on the robot's running trajectory. Then \(Q\) i ​​​​​The larger the value.

[0128] S106. Calculate the correction proportional gain component according to the error vibration intensity coefficients at all target times.

[0129] Since the larger the error vibration intensity coefficient of the error vibration signal, it means that there are more obvious vibration errors on each axis when the multi-axis industrial robot system is running at high speed. Therefore, in the multi-axis linkage interpolation control algorithm, the proportional gain coefficient P of the PID can be adjusted so that the control algorithm can stably output the regulation amount under the operating state affected by this non-linear characteristic.

[0130] In one embodiment, as Figure 5 shown, step S106 specifically includes the following steps:

[0131] S1061. Normalize the error vibration intensity coefficients at all target times.

[0132] S1062. Calculate the correction proportional gain component based on the normalized error vibration intensity coefficient and the original proportional gain component. Specifically, construct a correction adjustment proportional gain formula:

[0133] K p′ =(1 + Q i ′)K p

[0134] In the formula, Q i ′ represents the error vibration intensity coefficient of the multi-axis system error vibration signal of the industrial robot after normalization at the i-th motion posture (i.e., the i-th target time), K p represents the proportional gain component output before regulation, and K p′ represents the proportional gain component output after being corrected by the intensity coefficient of the error vibration signal, that is, the correction proportional gain component.

[0135] S107. Optimize the running trajectory of the robot according to the correction proportional gain component.

[0136] Finally, use the obtained correction proportional gain component to feedback to the multi-axis linkage interpolation control algorithm, so that it can better control the power and needle feeding speed of each axis in the multi-axis system, timely correct the error accumulation generated during the operation of the industrial robot, and improve the operation accuracy of the multi-axis industrial robot.

[0137] The present invention provides a multi-axis industrial robot running trajectory optimization system. The multi-axis industrial robot running trajectory optimization system of the present invention corresponds to a multi-axis industrial robot running trajectory optimization method. For the details not described in the multi-axis industrial robot running trajectory optimization system of the present invention, reference can be made to the specific embodiments of the multi-axis industrial robot running trajectory optimization method.

[0138] As Figure 6 shown, the multi-axis industrial robot trajectory optimization system specifically includes:

[0139] An acquisition module 101, configured to collect in real time the operation monitoring data of the multi-axis industrial robot at different target times, where the operation monitoring data includes the original vibration signals of each axis of the multi-axis industrial robot.

[0140] An obtaining module 102, configured to obtain the historical data set of the historical operation of the multi-axis industrial robot.

[0141] A first analysis module 103, configured to analyze the complexity factor of the motion postures of the multi-axis industrial robot at different target times according to the historical data set.

[0142] A second analysis module 104, configured to analyze the corresponding original vibration signal for different target times to obtain an error vibration signal.

[0143] A first calculation module 105, configured to calculate the error vibration intensity coefficient at different target times according to the error vibration signal and the complexity factor.

[0144] A second calculation module 106, configured to calculate a correction proportional gain component according to the error vibration intensity coefficients of all target times.

[0145] An optimization module 107, configured to optimize the operation trajectory of the robot according to the proportional gain component.

[0146] In an embodiment, the historical data set includes the historical power and historical speed of each axis. The historical power in the same motion posture is the same type of power data, and the historical speed in the same motion posture is the same type of speed data.

[0147] In an embodiment, as Figure 7 shown, the first analysis module 103 specifically includes:

[0148] A first calculation unit 1031, configured to calculate the information entropy of each axis of the industrial robot in the corresponding machine posture by calculating the probability of the same type of power data and the same type of speed data of a single axis appearing in the historical data set.

[0149] A second calculation unit 1032, configured to calculate the joint entropy between each pair of axes based on the same type of power data of each pair of axes and the same type of speed data of each pair of axes, so as to obtain the joint entropy of each pair of axes of the industrial robot in the corresponding machine posture, where any two axes form a pair of axes.

[0150] A first obtaining unit 1033, configured to obtain the motion posture at a target time.

[0151] A third calculation unit 1034, configured to calculate a complexity factor of the motion posture at the target moment according to the information entropy and joint entropy corresponding to the motion posture, so as to obtain complexity factors of motion postures at different target moments.

[0152] In one embodiment, as Figure 8 shown, the second analysis module 104 specifically includes:

[0153] A decomposition unit 1041, configured to decompose the original vibration signal using an ICA algorithm to obtain a plurality of decomposition results, where each decomposition result includes a first component signal and a second component signal.

[0154] A second acquisition unit 1042, configured to respectively fit the first component signal in each decomposition result into a function curve using a polynomial fitting method, and obtain all amplitude type distribution probabilities of the first component signal in each decomposition result.

[0155] A fourth calculation unit 1043, configured to calculate a kurtosis coefficient and a skewness coefficient of the amplitude type distribution of the first component signal in each decomposition result according to the all amplitude type distribution probabilities.

[0156] A construction unit 1044, configured to construct an original vibration signal decomposition function based on the kurtosis coefficient and the skewness coefficient.

[0157] A third acquisition unit 1045, configured to calculate function output values of all decomposition results based on the original vibration signal decomposition function, obtain all output values of the original vibration signal decomposition function, and select the decomposition result corresponding to the maximum output value as the optimal decomposition result of the original vibration signal, where the second signal component corresponding to the optimal decomposition result is the error vibration signal.

[0158] The error vibration signal includes a plurality of error vibration component signals.

[0159] In one embodiment, as Figure 9 shown, the first calculation module 105 specifically includes:

[0160] A fourth acquisition unit 1051, configured to acquire signal maximum points of all error vibration component signals for a target moment;

[0161] A fifth calculation unit 1052, configured to calculate an error vibration delay characteristic factor generated by each signal maximum point;

[0162] The sixth calculation unit 1053 is configured to calculate an error vibration intensity coefficient of the vibration error signal at the target moment based on the error vibration delay characteristic factor and the complexity factor of the corresponding motion posture at the same target moment, so as to obtain the error vibration intensity coefficients at different target moments.

[0163] In one embodiment, as Figure 10 shown, the second calculation module 106 specifically includes:

[0164] A normalization unit 1061 is configured to normalize the error vibration intensity coefficients at all target moments.

[0165] A seventh calculation unit 1062 is configured to calculate a corrected proportional gain component based on the normalized error vibration intensity coefficient and the original proportional gain component.

[0166] A multi-axis industrial robot operating trajectory optimization method and system provided by the present invention, for the original vibration signal of each axis in the multi-axis system, based on the independent component analysis method, combined with the periodic characteristic of the vibration signal of the real motion, obtains the error vibration component signal of the optimal decomposition result. Further, the error vibration intensity coefficient of the error vibration component signal is judged, and according to the vibration intensity coefficient of the error vibration component signal, it is fed back to the PID controller to adjust its corrected proportional gain component, and then fed back to the multi-axis linkage interpolation control algorithm to adjust the control of the corresponding axis, so as to achieve more accurate and reliable control of the operating trajectory of the multi-axis industrial robot, and realize the optimization of the operating trajectory of the multi-axis industrial robot.

[0167] It should be noted that: the above-mentioned sequence of the embodiments of the present invention is only for description, and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0168] Each embodiment in this specification is described in a progressive manner. The same and similar parts between the embodiments can be referred to each other. The key points of each embodiment are the differences from other embodiments.

Claims

1. A method for optimizing the running trajectory of a multi-axis industrial robot, characterized in that, The method includes: Collecting in real time the operation monitoring data of a multi-axis industrial robot at different target times, where the operation monitoring data includes the original vibration signals of each axis of the multi-axis industrial robot; Obtaining the historical data set of the historical operation of the multi-axis industrial robot; Analyzing the complexity factor of the motion posture of the multi-axis industrial robot at different target times according to the historical data set; For different target times, analyzing the corresponding original vibration signal to obtain an error vibration signal; Calculating the error vibration intensity coefficient at different target times according to the error vibration signal and the complexity factor; Calculating the correction proportional gain component according to the error vibration intensity coefficients of all target times; Optimizing the operation trajectory of the robot according to the correction proportional gain component; The historical data set includes the historical power and historical speed of each axis. The historical power under the same motion posture is the same type of power data, and the historical speed under the same motion posture is the same type of speed data; The analyzing the complexity factor of the motion posture of the multi-axis industrial robot at different target times according to the historical data set specifically includes: Calculating the information entropy of each axis based on the probability of the same type of power data and the same type of speed data of a single axis appearing in the historical data set, so as to obtain the information entropy of each axis of the industrial robot under the corresponding machine posture; Calculating the joint entropy between each pair of axes based on the same type of power data of each pair of axes and the same type of speed data of each pair of axes respectively, so as to obtain the joint entropy of each pair of axes of the industrial robot under the corresponding machine posture, where any two axes form a pair of axes; Obtaining the motion posture at a target time; Calculating the complexity factor of the motion posture at the target time according to the information entropy and joint entropy corresponding to the motion posture, so as to obtain the complexity factors of the motion postures at different target times; The error vibration signal includes multiple error vibration component signals; The calculating the error vibration intensity coefficient at different target times according to the error vibration signal and the complexity factor specifically includes: For a target time, obtaining the signal maximum points of all error vibration component signals; Calculating the error vibration delay characteristic factor generated by each signal maximum point; Calculating the error vibration intensity coefficient of the vibration error signal at the target time based on the error vibration delay characteristic factor at the same target time and the complexity factor of the corresponding motion posture, so as to obtain the error vibration intensity coefficients at different target times.

2. The multi-axis industrial robot running trajectory optimization method according to claim 1, wherein, The analyzing the corresponding original vibration signal to obtain an error vibration signal specifically includes: Using the ICA algorithm to decompose the original vibration signal to obtain several decomposition results, where each decomposition result includes a first component signal and a second component signal; Using the polynomial fitting method to fit the first component signal in each decomposition result into a function curve respectively, and obtaining all the amplitude type distribution probabilities of the first component signal in each decomposition result; Calculating the kurtosis coefficient and skewness coefficient of the amplitude type distribution of the first component signal in each decomposition result according to all the amplitude type distribution probabilities; Construct an original vibration signal decomposition function based on the kurtosis coefficient and skewness coefficient; Calculate the function output values of all decomposition results based on the original vibration signal decomposition function, obtain all output values of the original vibration signal decomposition function, and select the decomposition result corresponding to the maximum output value as the optimal decomposition result of the original vibration signal. The second signal component corresponding to the optimal decomposition result is the error vibration signal.

3. The multi-axis industrial robot running trajectory optimization method according to claim 1, characterized in that The calculating of the correction proportional gain component according to the error vibration intensity coefficients at all target times specifically includes: Normalize the error vibration intensity coefficients at all target times; Calculate the correction proportional gain component based on the normalized error vibration intensity coefficients and the original proportional gain component.

4. A multi-axis industrial robot running trajectory optimization system, characterized in that, The system includes: An acquisition module, configured to acquire in real time the operation monitoring data of the multi-axis industrial robot at different target times, where the operation monitoring data includes the original vibration signal of each axis of the multi-axis industrial robot; An obtaining module, configured to obtain the historical data set of the historical operation of the multi-axis industrial robot; A first analysis module, configured to analyze the complexity factor of the motion posture of the multi-axis industrial robot at different target times according to the historical data set; A second analysis module, configured to analyze the corresponding original vibration signal to obtain an error vibration signal for different target times; A first calculation module, configured to calculate the error vibration intensity coefficient at different target times according to the error vibration signal and the complexity factor; A second calculation module, configured to calculate the correction proportional gain component according to the error vibration intensity coefficients at all target times; An optimization module, configured to optimize the operation trajectory of the robot according to the proportional gain component; The historical data set includes the historical power and historical speed of each axis. The historical power in the same motion posture is the same type of power data, and the historical speed in the same motion posture is the same type of speed data; The first analysis module specifically includes: A first calculation unit, configured to calculate the information entropy of a single axis based on the probability of occurrence of the same type of power data and the same type of speed data of the single axis in the historical data set, so as to obtain the information entropy of each axis of the industrial robot in the corresponding machine posture; A second calculation unit, configured to calculate the joint entropy between each pair of axes based on the same type of power data of each pair of axes and the same type of speed data of each pair of axes, so as to obtain the joint entropy of each pair of axes of the industrial robot in the corresponding machine posture, where any two axes form a pair of axes; A first obtaining unit, configured to obtain the motion posture at a target time; A third calculation unit, configured to calculate the complexity factor of the motion posture at the target time according to the information entropy and the joint entropy corresponding to the motion posture, so as to obtain the complexity factors of the motion postures at different target times; The error vibration signal includes multiple error vibration component signals; The first calculation module specifically includes: A fourth obtaining unit, configured to obtain the signal maximum points of all error vibration component signals for a target time; A fifth calculation unit, configured to calculate the error vibration delay characteristic factor generated by each signal maximum point; A sixth computing unit, configured to calculate an error vibration intensity coefficient of the vibration error signal at the target moment based on an error vibration delay characteristic factor and a complexity factor of the corresponding motion posture at the same target moment, so as to obtain error vibration intensity coefficients at different target moments.

5. The multi-axis industrial robot running trajectory optimization system according to claim 4, wherein, The second analysis module specifically includes: A decomposition unit, configured to decompose the original vibration signal by using an ICA algorithm to obtain a plurality of decomposition results, where each decomposition result includes a first component signal and a second component signal; A second acquisition unit, configured to respectively fit the first component signal in each decomposition result into a function curve by using a polynomial fitting method, and obtain all amplitude type distribution probabilities of the first component signal in each decomposition result; A fourth computing unit, configured to calculate a kurtosis coefficient and a skewness coefficient of the amplitude type distribution of the first component signal in each decomposition result according to all the amplitude type distribution probabilities; A construction unit, configured to construct an original vibration signal decomposition function based on the kurtosis coefficient and the skewness coefficient; A third acquisition unit, configured to calculate function output values of all decomposition results based on the original vibration signal decomposition function, obtain all output values of the original vibration signal decomposition function, and select the decomposition result corresponding to the maximum output value as the optimal decomposition result of the original vibration signal, where the second signal component corresponding to the optimal decomposition result is the error vibration signal.

6. The multi-axis industrial robot running trajectory optimization system according to claim 4, characterized in that The second computing module specifically includes: A normalization unit, configured to perform normalization processing on the error vibration intensity coefficients at all target moments; A seventh computing unit, configured to calculate a corrected proportional gain component based on the normalized error vibration intensity coefficient and the original proportional gain component.

Images