A wafer cleaning robot motion stability adaptive adjustment system and method

Abstract

The present invention relates to the field of semiconductor technology, specifically to a wafer cleaning robot motion stability adaptive adjustment system and method, the method comprising: collecting joint speed, vibration acceleration and end force data during the robot motion process, constructing a data set; using spectrum analysis and Fourier transform to extract features from the data, constructing a stability evaluation strategy, and judging the robot motion stability in real time; when an unstable state is detected, using a pre-trained clustering support vector machine model to diagnose the cause of the fault; according to the diagnosis result, dynamically adjusting the robot PID controller gain value and joint damping compensation coefficient, generating a new control law and applying it to the robot; after the control adjustment, continuously monitoring the stability until the robot recovers stability. The method of the present invention realizes automatic evaluation, fault diagnosis and control optimization of the robot stability, and improves the reliability and efficiency of the wafer cleaning equipment.

Description

Technical Field

[0001] The present invention relates to the field of semiconductor technology, and in particular to a wafer cleaning robot arm motion stability adaptive regulation system and method. Background Art

[0002] Wafer cleaning is a key link in semiconductor manufacturing, in which the motion stability of the robot arm directly affects the cleaning effect and wafer quality. The traditional control method of the wafer cleaning robot arm mainly uses a PID controller to achieve the motion control of the robot arm by manually adjusting the controller parameters. However, this control method relies on the experience and skills of the operator, lacks real-time monitoring and quantitative evaluation of the robot arm's motion state, and is difficult to adapt to the high-precision and high-stability requirements of the wafer cleaning process.

[0003] Most existing robotic arm control technologies are based on fixed adjustment strategies and lack effective mechanisms for dealing with abnormal conditions such as vibration and speed fluctuations during the movement of the robotic arm. When the robotic arm moves unstably, the existing control system cannot detect and diagnose the cause of the fault in a timely manner, resulting in a decline in wafer cleaning quality and even serious consequences such as equipment damage.

[0004] In addition, the existing robot control method lacks adaptive adjustment capabilities and cannot dynamically adjust the strategy according to the actual motion state of the robot. During the wafer cleaning process, the robot's load, motion speed and other working parameters may change. Fixed control parameters are difficult to meet the stability requirements under different working conditions, making it difficult to ensure the accuracy and stability of the robot's motion.

[0005] Therefore, there is an urgent need for a wafer cleaning robot arm motion stability control method that can monitor the robot arm motion state in real time, intelligently diagnose the cause of the fault, and adaptively adjust the adjustment strategy according to the cause of the fault.

[0006] In view of this, the present invention proposes a wafer cleaning robot arm motion stability adaptive adjustment system and method. Summary of the invention

[0007] To achieve the above-mentioned purpose, the present invention provides a system and method for adaptively adjusting the motion stability of a wafer cleaning robot arm. The specific technical scheme is as follows: A method for adaptively adjusting the motion stability of a wafer cleaning robot arm, comprising:

[0008] Collect the working data of the wafer cleaning robot, including the joint speed, vibration data and end effector force data of the robot;

[0009] Perform feature analysis on the collected working data, build a robot arm stability assessment strategy, use spectrum analysis and Fourier transform algorithm to conduct real-time assessment of the robot arm's motion state, and determine whether the robot arm's current motion is in a stable state;

[0010] Build and train a robot arm problem detection model. When the robot arm is in an unstable state, the robot arm problem detection model is used to identify the cause of the unstable robot arm failure.

[0011] Build a control algorithm for dynamically adjusting the robotic arm, and adjust the robotic arm according to the cause of the unstable fault of the robotic arm, including adjusting the gain value of the robotic arm PID controller and adjusting the damping compensation of the robotic arm joints;

[0012] After implementing control adjustments on the robotic arm, continuously monitor whether the robotic arm is in a stable state. If the robotic arm is not in a stable state, readjust the robotic arm until it is in a stable state.

[0013] Preferably, the joint speed data of the robot arm is collected, and the first The angular velocity of each joint is The sampling time is , then the angular velocity sampling value in discrete time is , is the sampling times;

[0014] Collect the vibration data of the robot arm, set in the spatial coordinate system The acceleration components in the three directions are , then the acceleration sampling value is:

[0015] ;

[0016] in, is the sampling time interval; collect the force data of the end effector of the robot arm and measure the three-dimensional force on the end of the robot arm. Suppose the three-dimensional force is ,but: ;

[0017] in, and In three-dimensional space Direction of the force vector;

[0018] Summarize the various types of collected data and build them into a data set:

[0019] ;

[0020] in, is the angular velocity vector of all joints, n is the number of joints, is the acceleration vector.

[0021] Preferably, for the dataset Extract time domain features from each signal sequence in, including the mean ,variance , RMS value and peak-to-peak ;

[0022] Use Fourier transform to convert the signal from time domain to frequency domain and extract the data set The spectrum characteristics of the discrete signal sequence , Refers to the dataset For any signal type in , the discrete Fourier transform is:

[0023] ;

[0024] in, For the frequency components, is an imaginary number, Indicated in The signal value collected at each moment, is the total number of sampling points of the signal, ;

[0025] Depend on Calculate the spectral amplitude of a signal and Phase :

[0026] ;

[0027] in, Characteristic signal In frequency The magnitude of the Characteristic signal The corresponding phase angle;

[0028] Extract the maximum value of the spectrum amplitude , Spectrum Center of Gravity Features, characterize the distribution characteristics of the signal in the frequency domain;

[0029] Construct a robot arm stability evaluation strategy, construct a robot arm stability evaluation index based on the extracted time domain and frequency domain features; define the stability index of the robot arm joint speed for:

[0030] ;

[0031] in, is the weight coefficient, and Respectively represent The mean, variance, RMS value, peak-to-peak value, maximum amplitude extraction value and spectrum centroid value of the characteristic data of each robot arm;

[0032] Similarly, calculate and define the vibration data stability index of the robot arm , and calculate and define the end force stability index of the robot arm .

[0033] Preferably, the robot arm joint speed, vibration data, and the stability evaluation threshold of the end force are set respectively, and the robot arm joint speed stability threshold is set to , the vibration data stability threshold is And the stability threshold of the end force is ;

[0034] According to the set threshold , judge in real time whether the current robot arm movement is stable:

[0035] ;

[0036] If the above three conditions are met at the same time, the current robot arm movement is considered to be in a stable state, otherwise the robot arm is considered to be in an unstable state.

[0037] Preferably, based on the collected robot arm working data, the fault cause labels corresponding to the robot arm working data are established through manual analysis to construct a training data set. ,in, is the number of samples in the training dataset, Indicates The feature vector of each sample includes the joint velocity, vibration and end force data of the robot arm. Indicates the corresponding fault cause label, is the total number of failure causes;

[0038] The clustering support vector machine C-SVM model is used as the robot arm problem detection model, using the training data set The clustering support vector machine C-SVM model is trained to obtain the optimal model parameters.

[0039] Preferably, when the robot arm is in an unstable state, that is, when it is detected in the real-time stability evaluation step Any value in does not satisfy Extract the feature vector of the robot arm working data at the current moment, and convert the feature vector Input into the trained robot arm problem detection model to obtain the predicted fault cause label ;

[0040] The prediction function of the robotic arm problem detection model built based on the C-SVM model is expressed as: ;

[0041] Each fault category There are corresponding parameters and ; Based on the predicted fault cause label , query the pre-defined correspondence table between fault causes and reasons, and obtain the fault cause of the instability of the robotic arm.

[0042] Preferably, a mapping relationship table between the fault cause and the adjustment strategy is established according to the fault cause identified by the robot arm problem detection model;

[0043] Assume the fault cause category is The corresponding adjustment strategy is , then the mapping relationship table is represented as: ;in, is the total number of failure causes;

[0044] For the PID controller of the robotic arm, the control law of the PID controller of the robotic arm is expressed as: ;in, Indicates time, is the controller output, is the tracking error, and They are proportional, integral and differential gains respectively; the gain value of the PID controller is dynamically adjusted according to the identified fault cause;

[0045] Set up The adjustment strategy for each joint is , then the corresponding PID controller gain adjustment formula is: ; in, and To adjust the strategy Calculated proportional, integral, and derivative gain adjustments; and Respectively Proportional, integral and derivative gains for each joint;

[0046] For the joint damping compensation of the robot arm, the compensation torque is expressed as: ;in, For the The damping compensation torque of each joint, is the damping coefficient, For the Angular velocity of each joint; dynamically adjust the damping coefficient of the mechanical arm joint damping compensation torque damping compensation according to the identified fault cause;

[0047] No. Adjustment strategies for each joint , the corresponding torque damping compensation coefficient adjustment formula is: ;in, To adjust the strategy The calculated torque damping compensation coefficient adjustment amount;

[0048] Apply the adjusted PID controller gain value and torque damping compensation coefficient to the robot arm control to obtain the adjusted control law: ;in, For the The controller output of each joint, For the The tracking error of each joint is calculated; the adjusted control law is applied to the robot control system to realize dynamic adjustment of the robot movement.

[0049] Preferably, after the control adjustment, the joint speed, vibration and end force data of the robot arm are continuously collected; the collected robot arm working data are feature analyzed to construct a working data feature vector; the stability index is calculated using the robot arm stability evaluation strategy; and it is determined whether the robot arm movement is stable, that is, whether the stability index meets the threshold condition;

[0050] If the robot arm is stable, the control adjustment is successful and the adjustment process ends; if the robot arm is unstable, the working data feature vector is input into the robot arm problem detection model to obtain the predicted fault cause label;

[0051] According to the fault cause label, query the mapping relationship table between the fault cause and the adjustment strategy to obtain the corresponding adjustment strategy. According to the adjustment strategy, adjust the PID controller gain value and joint damping compensation coefficient of the robot arm to obtain a new control law.

[0052] The new control law is applied to the robot control system, and the robot is repeatedly monitored and adjusted until the robot resumes stable motion. After the adjustment, If the robot arm still fails to reach a stable state after the preset maximum number of adjustments, an alarm will be triggered.

[0053] A wafer cleaning robot motion stability adaptive adjustment system, which is used to implement the wafer cleaning robot motion stability adaptive adjustment method, comprises: a data acquisition module, a stability assessment module, a fault identification module, a control adjustment module and a cycle monitoring module;

[0054] The data acquisition module is used to collect working data of the wafer cleaning robot arm, including joint speed, vibration data and end effector force data of the robot arm;

[0055] The stability assessment module is used to perform feature analysis on the collected working data, construct a stability assessment strategy for the robotic arm, and use spectrum analysis and Fourier transform algorithms to perform real-time assessment of the motion state of the robotic arm to determine whether the current motion of the robotic arm is in a stable state;

[0056] The fault identification module is used to construct and train a robot arm problem detection model. When the robot arm is in an unstable state, the cause of the instability of the robot arm is identified through the robot arm problem detection model.

[0057] The control and adjustment module is used to construct a control algorithm for dynamically adjusting the robotic arm, and to adjust the robotic arm according to the cause of the unstable fault of the robotic arm, including adjusting the gain value of the robotic arm PID controller and adjusting the damping compensation of the robotic arm joint;

[0058] The cyclic monitoring module continuously monitors whether the robotic arm is in a stable state after implementing control adjustment on the robotic arm. If the robotic arm is not in a stable state, the robotic arm is readjusted until the robotic arm is in a stable state.

[0059] An electronic device comprises: a processor and a memory, wherein the memory stores a computer program that can be called by the processor; the processor executes the method for adaptively adjusting the motion stability of a wafer cleaning robot arm by calling the computer program stored in the memory.

[0060] A computer-readable storage medium stores instructions. When the instructions are executed on a computer, the computer executes the method for adaptively adjusting the motion stability of a wafer cleaning robot arm.

[0061] Beneficial effects of the present invention: By collecting the joint speed, vibration and force data of the robotic arm, the present invention can fully understand the motion state and working performance of the robotic arm, and provide a data basis for subsequent stability evaluation and fault diagnosis.

[0062] The present invention utilizes algorithms such as spectrum analysis and Fourier transform to extract and analyze features of the collected data and construct stability evaluation indicators, which can evaluate the motion stability of the robotic arm in real time and promptly discover potential anomalies and faults.

[0063] The present invention constructs a problem detection model through a machine learning algorithm, and trains the model to identify the cause of the instability of the robotic arm. When an abnormality occurs in the robotic arm, the fault type and cause can be quickly diagnosed, providing decision support for locating and eliminating the fault.

[0064] The present invention dynamically adjusts the control algorithm of the mechanical arm, such as the gain and damping compensation of the PID controller, according to the diagnosed fault cause, so as to realize the adaptive optimization control of the movement of the mechanical arm.

[0065] After executing the control adjustment, the present invention continuously monitors the stability of the robotic arm to form a closed-loop feedback control, which can verify the effectiveness of the adjustment measures and ensure that the robotic arm is always in a stable and controlled state. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] Figure 1 A flow chart of a method for adaptively adjusting the motion stability of a wafer cleaning robot arm provided by the present invention;

[0067] Figure 2 This is a structural diagram of a wafer cleaning robot arm motion stability adaptive adjustment system provided by the present invention. DETAILED DESCRIPTION

[0068] In order to better understand the present invention, a more detailed description will be made of various aspects of the present invention with reference to the accompanying drawings. It should be understood that these detailed descriptions are only descriptions of exemplary embodiments of the present invention, and are not intended to limit the scope of the present invention in any way. Throughout the specification, the same reference numerals refer to the same elements. The expression "and / or" includes any and all combinations of one or more of the associated listed items.

[0069] In the accompanying drawings, the size, dimensions and shapes of the elements have been slightly adjusted for ease of illustration. The accompanying drawings are for illustration only and are not strictly drawn to scale. As used herein, the terms "substantially", "approximately" and similar terms are used as terms of approximation, not as terms of degree, and are intended to illustrate the inherent deviations in measurements or calculations that will be recognized by those of ordinary skill in the art. In addition, in the present invention, the order in which the steps are described does not necessarily represent the order in which these processes occur in actual operation, unless otherwise specified or can be derived from the context.

[0070] It should also be understood that expressions such as "comprises", "including", "having", "includes" and / or "comprising" are open rather than closed expressions in this specification, which indicate the presence of the stated features, elements and / or components, but do not exclude the presence of one or more other features, elements, components and / or combinations thereof. In addition, when expressions such as "at least one of..." appear after a list of listed features, they modify the entire list of features rather than just the individual elements in the list. In addition, when describing embodiments of the present invention, "may" is used to mean "one or more embodiments of the present invention". And, the term "exemplary" is intended to refer to an example or illustration.

[0071] Unless otherwise defined, all terms (including engineering terms and scientific and technological terms) used in this document have the same meaning as commonly understood by ordinary technicians in the field to which the present invention belongs. It should also be understood that unless otherwise clearly stated in the present invention, words defined in commonly used dictionaries should be interpreted as having the same meaning as their meaning in the context of the relevant technology, and should not be interpreted in an idealized or overly formal sense.

[0072] It should be noted that, in the absence of conflict, the embodiments and features of the embodiments of the present invention can be combined with each other. The present invention will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0073] Example 1

[0074] Reference Figure 1 , which is the first embodiment of the present invention, provides a method for adaptively adjusting the motion stability of a wafer cleaning robot arm.

[0075] S1: Collect the working data of the wafer cleaning robot, including the joint speed, vibration data and end effector force data of the robot.

[0076] Collect the joint velocity data of the robot arm, install encoders at each joint of the robot arm, and collect the angular velocity signals of each joint in real time; The angular velocity of each joint is The sampling time is , then the angular velocity sampling value in discrete time is ,in is the number of sampling points.

[0077] Collect vibration data of the robot arm, install acceleration sensors at key positions of the robot arm (such as wrist joints, end effectors, etc.), and collect vibration acceleration signals in real time; The acceleration component in the direction is , then the acceleration sampling value in discrete time is: ;in, is the sampling time interval; , is the number of sampling points.

[0078] Collect the force data of the end effector of the robot arm, install a force sensor on the end effector of the robot arm, and measure the three-dimensional force on the end of the robot arm in real time. Suppose the three-dimensional force is ; Then the force sampling value under discrete time is: ;

[0079] in, and In three-dimensional space Directional force data.

[0080] Aggregate the collected data into a dataset as input for subsequent analysis and modeling:

[0081] ;

[0082] in, is the angular velocity vector of all joints, n is the number of joints, is the acceleration vector.

[0083] In step S1, sensors are installed at key positions of the robot arm to collect data on joint velocity, vibration acceleration, and end force during the movement of the robot arm, and a data set is constructed to provide a data basis for subsequent stability analysis and fault diagnosis.

[0084] S2: Perform feature analysis on the collected working data, build a robot arm stability assessment strategy, use spectrum analysis and Fourier transform algorithm to conduct real-time assessment of the robot arm's motion state, and determine whether the robot arm's current motion is in a stable state.

[0085] For the dataset Extract time domain features from each signal sequence in, including the mean ,variance , RMS value and peak-to-peak .

[0086] For example, the angular velocity For example, the mean ,variance , RMS value and peak-to-peak The calculation formulas are:

[0087] ;

[0088] ;

[0089] ;

[0090] ;

[0091] The time domain characteristics of other signals can be calculated similarly. The time domain characteristics reflect the energy, fluctuation amplitude and other characteristics of the signal.

[0092] Use Fourier transform to convert the signal from time domain to frequency domain and extract the data set The spectrum characteristics of the discrete signal sequence , Refers to the dataset For any signal type in , the discrete Fourier transform (DFT) is:

[0093] ;

[0094] in, For the frequency components, is an imaginary number, Indicated in The signal value collected at each moment, is the total number of sampling points of the signal, .

[0095] Depend on The spectral amplitude of the signal can be calculated and Phase :

[0096] ;

[0097] in, Characteristic signal In frequency The magnitude of the Characteristic signal The corresponding phase angle.

[0098] Extract the maximum value of the spectrum amplitude , Spectrum Center of Gravity Features characterize the distribution characteristics of the signal in the frequency domain.

[0099] Construct a robot arm stability evaluation strategy, construct a robot arm stability evaluation index based on the extracted time domain and frequency domain features; define the stability index of the robot arm joint speed for:

[0100] ;

[0101] in is the weight coefficient; and Respectively represent Extract the maximum value of the amplitude and the centroid value of the spectrum from the characteristic data of each robot arm; The fluctuation of the angular velocity of the robot, the peak-to-peak value, the main frequency amplitude and other factors are comprehensively considered. The smaller the value, the more stable the angular velocity.

[0102] Similarly, calculate and define the vibration data stability index of the robot arm , and calculate and define the end force stability index of the robot arm .

[0103] Set the stability evaluation thresholds of the robot arm joint speed, vibration data and end force respectively, and set the robot arm joint speed stability threshold to , the vibration data stability threshold is And the stability threshold of the end force is .

[0104] According to the set threshold , judge in real time whether the current robot arm movement is stable:

[0105] ;

[0106] If the above three conditions are met at the same time, the current robot arm movement is considered to be in a stable state, otherwise the robot arm is considered to be in an unstable state.

[0107] In step S2, the time-frequency domain features of the collected robot arm working data are extracted, and a stability evaluation index is constructed. By setting a threshold, it is determined whether the robot arm movement is stable, thereby realizing a real-time evaluation of the robot arm's stability.

[0108] S3: Build and train a robot arm problem detection model. When the robot arm is in an unstable state, the robot arm problem detection model is used to identify the cause of the instability of the robot arm.

[0109] Based on the collected robot arm working data, through manual analysis, the fault cause labels corresponding to the robot arm working data are established to build a training data set ,in, is the number of samples in the training dataset, Indicates The feature vector of each sample includes the joint velocity, vibration and end force data of the robot arm. Indicates the corresponding fault cause label, is the total number of failure reasons.

[0110] The clustering support vector machine C-SVM model is used as the robot arm problem detection model; the goal of the clustering support vector machine C-SVM model is to find an optimal hyperplane , so that samples of different categories can be separated by the hyperplane, solving the multi-classification problem of the causes of unstable failure of the robotic arm; the optimization goal of the C-SVM model can be expressed as: ;

[0111] ;

[0112] in, and are the parameters of the hyperplane, is the slack variable, is the balancing factor; by solving this optimization problem, the optimal classification hyperplane is obtained.

[0113] Using the training dataset Train the selected machine learning model to obtain the optimal model parameters.

[0114] The robot arm problem detection model constructed by the cluster support vector machine C-SVM model uses the sequential minimum optimization (SMO) algorithm or the gradient descent method to solve the above optimization problem and obtain the optimal and ; During the training process, the cross-validation method is used to evaluate and tune the C-SVM model, and the C-SVM model with the best performance is selected as the final robotic arm problem detection model.

[0115] When the robot is in an unstable state, it is detected in the real-time stability assessment step. Any value in does not satisfy Extract the feature vector of the robot arm working data at the current moment , the feature vector Input into the trained robot arm problem detection model to obtain the predicted fault cause label .

[0116] The prediction function of the robotic arm problem detection model built based on the C-SVM model is expressed as: ;

[0117] Among them, each category There are corresponding parameters and ; Based on the predicted fault cause label , query the pre-defined fault cause and cause correspondence table to obtain the fault cause of the unstable robot arm; such as loose joints, motor failure, abnormal load, etc.

[0118] In step S3, a labeled training data set is constructed based on the robot arm working data, and a clustering support vector machine model is used to train the robot arm problem detection model to realize automatic diagnosis of the fault cause of the robot arm in an unstable state.

[0119] S4: Build a control algorithm for dynamically adjusting the robotic arm, and adjust the robotic arm according to the cause of the unstable fault of the robotic arm, including adjusting the gain value of the robotic arm PID controller and adjusting the damping compensation of the robotic arm joints.

[0120] According to the fault causes identified by the robot arm problem detection model, a mapping relationship table between the fault causes and the adjustment strategies is established.

[0121] Assume the fault cause category is The corresponding adjustment strategy is , then the mapping relationship table is represented as: ;in, is the total number of failure reasons.

[0122] For the PID controller of the robotic arm, the control law of the PID controller of the robotic arm is expressed as: ;in, Indicates time, is the controller output, is the tracking error, and They are proportional, integral and differential gains respectively; the gain value of the PID controller is dynamically adjusted according to the identified fault cause.

[0123] Set up The adjustment strategy for each joint is , then the corresponding PID controller gain adjustment formula is: ; in, and To adjust the strategy Calculated proportional, integral, and derivative gain adjustments; and Respectively Proportional, integral, and derivative gains for each joint.

[0124] For the joint damping compensation of the robot arm, the compensation torque is expressed as: ;in, For the The damping compensation torque of each joint, is the damping coefficient, For the The angular velocity of each joint is determined; according to the identified fault cause, the damping coefficient of the robot arm joint damping compensation torque damping compensation is dynamically adjusted.

[0125] Set up The adjustment strategy for each joint is , then the corresponding torque damping compensation coefficient adjustment formula is: ;in, To adjust the strategy The calculated torque damping compensation coefficient adjustment.

[0126] Apply the adjusted PID controller gain value and torque damping compensation coefficient to the robot arm control to obtain the adjusted control law:

[0127] ;

[0128] in, For the The controller output of each joint, For the The tracking error of each joint is reduced; the adjusted control law is applied to the robot control system to realize dynamic adjustment of the robot movement; and the stability of the robot movement is improved.

[0129] In step S4, the gain value and joint damping compensation coefficient of the robot arm PID controller are dynamically adjusted according to the diagnosed fault cause, a new control law is generated, and applied to the robot arm control system to achieve dynamic adjustment of the robot arm movement and improve stability.

[0130] S5: After implementing control adjustment on the robotic arm, continuously monitor whether the robotic arm is in a stable state. If the robotic arm is not in a stable state, readjust the robotic arm until the robotic arm is in a stable state.

[0131] After the control is adjusted, the joint speed, vibration and end force data of the robot arm are continuously collected; the collected robot arm working data are feature analyzed to construct the working data feature vector; the robot arm stability evaluation strategy is used to calculate the stability index; and it is determined whether the robot arm movement is stable, that is, whether the stability index meets the threshold condition.

[0132] If the robot arm is stable, the control adjustment is successful and the adjustment process ends; if the robot arm is unstable, the working data feature vector is input into the robot arm problem detection model to obtain the predicted fault cause label.

[0133] According to the fault cause label, the mapping relationship table between the fault cause and the adjustment strategy is queried to obtain the corresponding adjustment strategy. According to the adjustment strategy, the PID controller gain value and the joint damping compensation coefficient of the robot arm are adjusted to obtain a new control law.

[0134] The new control law is applied to the robot control system, and the robot is repeatedly tested and adjusted until the robot resumes stable motion. After the adjustment, If the robot arm still fails to reach a stable state after the preset maximum number of adjustments, an alarm will be triggered.

[0135] After the robot arm is controlled and adjusted in step S5, the stability is continuously monitored. If the robot arm is still unstable, the diagnosis and adjustment steps are repeated until the robot arm regains stability. If multiple adjustments are ineffective, an alarm is triggered to achieve closed-loop optimization of the robot arm's stability.

[0136] Example 2

[0137] Reference Figure 2 , which is the second embodiment of the present invention, provides a wafer cleaning robot arm motion stability adaptive adjustment system.

[0138] The system comprises: a data acquisition module, a stability assessment module, a fault identification module, a control and regulation module and a circulation monitoring module.

[0139] The data acquisition module is used to collect working data of the wafer cleaning robot arm, including joint speed and vibration data of the robot arm and force data of the end effector.

[0140] The stability assessment module is used to perform feature analysis on the collected working data, construct a robot arm stability assessment strategy, and use spectrum analysis and Fourier transform algorithms to perform real-time assessment of the motion state of the robot arm to determine whether the current motion of the robot arm is in a stable state.

[0141] The fault identification module is used to construct and train a robot arm problem detection model. When the robot arm is in an unstable state, the fault cause of the instability of the robot arm is identified through the robot arm problem detection model.

[0142] The control and adjustment module is used to construct a control algorithm for dynamically adjusting the robotic arm, and adjust the robotic arm according to the cause of the unstable fault of the robotic arm, including adjusting the gain value of the robotic arm PID controller and adjusting the damping compensation of the robotic arm joint.

[0143] The cyclic monitoring module continuously monitors whether the robotic arm is in a stable state after implementing control adjustment on the robotic arm. If the robotic arm is not in a stable state, the robotic arm is readjusted until the robotic arm is in a stable state.

[0144] Example 3

[0145] The present invention also provides an electronic device. The electronic device may include one or more processors and one or more memories. The memories store computer-readable codes, and when the computer-readable codes are executed by the one or more processors, the method for adaptively adjusting the motion stability of a wafer cleaning robot arm as described above may be executed.

[0146] The method or system according to the embodiment of the present invention may also be implemented with the aid of the architecture of the electronic device of the present invention.

[0147] An electronic device may include a bus, one or more CPUs, read-only memory (ROM), random access memory (RAM), a communication port for connecting to a network, input / output components, a hard disk, and the like.

[0148] A storage device in an electronic device, such as a ROM or a hard disk, can store a method for adaptively adjusting the motion stability of a wafer cleaning robot arm provided by the present invention.

[0149] A method for adaptively adjusting the motion stability of a wafer cleaning robot arm comprises: collecting working data of the wafer cleaning robot arm, including joint speed, vibration data and end effector force data of the robot arm; performing feature analysis on the collected working data, building a robot arm stability evaluation strategy, and using spectrum analysis and Fourier transform algorithms to evaluate the motion state of the robot arm in real time to determine whether the current motion of the robot arm is in a stable state; building and training a robot arm problem detection model, and when the robot arm is in an unstable state, identifying the cause of the instability of the robot arm through the robot arm problem detection model; building a control algorithm for dynamically adjusting the robot arm, and adjusting the robot arm according to the cause of the instability of the robot arm, including adjusting the gain value of the robot arm PID controller and adjusting the joint damping compensation of the robot arm; after implementing control adjustment on the robot arm, continuously monitoring whether the robot arm is in a stable state, and if the robot arm is not in a stable state, readjusting the robot arm until the robot arm is in a stable state.

[0150] Furthermore, the electronic device may also include a user interface. Of course, the architecture of the present invention is only exemplary, and when implementing different devices, one or more components in the electronic device disclosed in the present invention may be omitted according to actual needs.

[0151] Example 4

[0152] The invention also discloses a computer-readable storage medium.

[0153] The computer-readable storage medium has computer-readable instructions stored thereon.

[0154] When the computer-readable instructions are executed by the processor, a method for adaptively adjusting the motion stability of a wafer cleaning robot arm according to an embodiment of the present invention described with reference to the above drawings may be executed.

[0155] The storage medium includes, but is not limited to, for example, volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and cache memory (cache), etc. Non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc. In addition, according to an embodiment of the present invention, the process described above with reference to the flowchart may be implemented as a computer software program.

[0156] For example, the present invention provides a non-temporary machine-readable storage medium, which stores machine-readable instructions, and the machine-readable instructions can be run by a processor to execute instructions corresponding to the method steps provided by the present invention, for example: collecting working data of a wafer cleaning robot arm, including the joint speed, vibration data and end effector force data of the robot arm; performing feature analysis on the collected working data, building a robot arm stability evaluation strategy, and using spectrum analysis and Fourier transform algorithms to perform real-time evaluation of the motion state of the robot arm to determine whether the current motion of the robot arm is in a stable state; building and training a robot arm problem detection model, and when the robot arm is in an unstable state, identifying the cause of the instability of the robot arm through the robot arm problem detection model; building a control algorithm for dynamically adjusting the robot arm, and adjusting the robot arm according to the cause of the instability of the robot arm, including adjusting the gain value of the robot arm PID controller and adjusting the joint damping compensation of the robot arm; after implementing control adjustment on the robot arm, continuously monitoring whether the robot arm is in a stable state, and if the robot arm is not in a stable state, readjusting the robot arm until the robot arm is in a stable state.

[0157] When the computer program is executed by the central processing unit (CPU), the above functions defined in the method of the present invention are performed. The method, apparatus, and device of the present invention may be implemented in many ways. For example, the method, apparatus, and device of the present invention may be implemented by software, hardware, firmware, or any combination of software, hardware, and firmware.

[0158] The above sequence for the steps of the method is for illustration only, and the steps of the method of the present invention are not limited to the sequence specifically described above unless otherwise specifically stated.

[0159] In addition, in some embodiments, the present invention can also be implemented as a program recorded in a recording medium, which includes machine-readable instructions for implementing the method according to the present invention. Therefore, the present invention also covers a recording medium storing a program for executing the method according to the present invention.

[0160] In addition, the parts of the above technical solutions provided in the embodiments of the present invention that are consistent with the implementation principles of the corresponding technical solutions in the prior art are not described in detail to avoid excessive redundancy.

[0161] The specific implementation modes as described above further describe the purpose, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation mode of the present invention and is not intended to limit the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for adaptively adjusting the motion stability of a wafer cleaning robot arm, characterized in that: include: Collect the working data of the wafer cleaning robot, including the joint speed, vibration data and end effector force data of the robot; Collect the vibration data of the robot arm, set in the spatial coordinate system The acceleration components in the three directions are , then the acceleration sampling value is: ； in, is the sampling time interval; collect the force data of the end effector of the robot arm and measure the three-dimensional force on the end of the robot arm. Suppose the three-dimensional force is ,but: ； in, , and In three-dimensional space Direction of the force vector; Perform feature analysis on the collected working data, build a robot arm stability assessment strategy, use spectrum analysis and Fourier transform algorithm to conduct real-time assessment of the robot arm's motion state, and determine whether the robot arm's current motion is in a stable state; Set the robot arm joint speed, vibration data, and end force stability assessment threshold respectively, and set the robot arm joint speed stability threshold to , the vibration data stability threshold is And the stability threshold of the end force is ; According to the set threshold , judge in real time whether the current robot arm movement is stable: ； If the above three conditions are met at the same time, the current robot arm motion is considered to be in a stable state, otherwise the robot arm is considered to be in an unstable state; Build and train a robot arm problem detection model. When the robot arm is in an unstable state, the robot arm problem detection model is used to identify the cause of the unstable robot arm failure. Build a control algorithm for dynamically adjusting the robotic arm, and adjust the robotic arm according to the cause of the unstable fault of the robotic arm, including adjusting the gain value of the robotic arm PID controller and adjusting the damping compensation of the robotic arm joints; After implementing control adjustments on the robotic arm, continuously monitor whether the robotic arm is in a stable state. If the robotic arm is not in a stable state, readjust the robotic arm until it is in a stable state.

2. The method for adaptively adjusting the motion stability of a wafer cleaning robot arm according to claim 1, characterized in that: Collect the joint speed data of the robot arm, set The angular velocity of each joint is The sampling time is , then the angular velocity sampling value in discrete time is , is the sampling times; Summarize the various types of collected data and build them into a data set: ； in, is the angular velocity vector of all joints, n is the number of joints, is the acceleration vector.

3. The method for adaptively adjusting the motion stability of a wafer cleaning robot arm according to claim 2, characterized in that: For the dataset Extract time domain features from each signal sequence in, including the mean ,variance , RMS value and peak-to-peak ; Use Fourier transform to convert the signal from time domain to frequency domain and extract the data set The spectrum characteristics of the discrete signal sequence , Refers to the dataset For any signal type in , the discrete Fourier transform is: ； in, For the frequency components, is an imaginary number, Indicated in The signal value collected at each moment, is the total number of sampling points of the signal, ; Depend on Calculate the spectral amplitude of a signal and Phase : ； ； in, Characteristic signal In frequency The magnitude of the Characteristic signal The corresponding phase angle; Extract the maximum value of the spectrum amplitude , Spectrum Center of Gravity Features, characterize the distribution characteristics of the signal in the frequency domain; Construct a robot arm stability evaluation strategy, construct a robot arm stability evaluation index based on the extracted time domain and frequency domain features; define the stability index of the robot arm joint speed for: ； in, is the weight coefficient, , , , , and Respectively represent The mean, variance, RMS value, peak-to-peak value, maximum amplitude extraction value and spectrum centroid value of the characteristic data of each robot arm; Similarly, calculate and define the vibration data stability index of the robot arm , and calculate and define the end force stability index of the robot arm .

4. The method for adaptively adjusting the motion stability of a wafer cleaning robot arm according to claim 3, characterized in that: Based on the collected robot arm working data, through manual analysis, the fault cause labels corresponding to the robot arm working data are established to build a training data set ,in, is the number of samples in the training dataset, Indicates The feature vector of each sample includes the joint velocity, vibration and end force data of the robot arm. Indicates the corresponding fault cause label, is the total number of failure causes; The clustering support vector machine C-SVM model is used as the robot arm problem detection model, using the training data set The clustering support vector machine C-SVM model is trained to obtain the optimal model parameters.

5. The method for adaptively adjusting the motion stability of a wafer cleaning robot arm according to claim 4, characterized in that: When the robot is in an unstable state, it is detected in the real-time stability assessment step. Any value in does not satisfy Extract the feature vector of the robot arm working data at the current moment, and convert the feature vector Input into the trained robot arm problem detection model to obtain the predicted fault cause label ; The prediction function of the robotic arm problem detection model built based on the C-SVM model is expressed as: ； Each fault category There are corresponding parameters and ; Based on the predicted fault cause label , query the pre-defined correspondence table between fault causes and reasons, and obtain the fault cause of the instability of the robotic arm.

6. The method for adaptively adjusting the motion stability of a wafer cleaning robot arm according to claim 5, characterized in that: According to the fault causes identified by the robot arm problem detection model, a mapping relationship table between the fault causes and the adjustment strategies is established; Assume the fault cause category is The corresponding adjustment strategy is , then the mapping relationship table is represented as: ;in, is the total number of failure causes; For the PID controller of the robotic arm, the control law of the PID controller of the robotic arm is expressed as: ;in, Indicates time, is the controller output, is the tracking error, , and They are proportional, integral and differential gains respectively; the gain value of the PID controller is dynamically adjusted according to the identified fault cause; Set up The adjustment strategy for each joint is , then the corresponding PID controller gain adjustment formula is: ;in, , and To adjust the strategy Calculated proportional, integral, and derivative gain adjustments; , and Respectively Proportional, integral and derivative gains for each joint; For the joint damping compensation of the robot arm, the compensation torque is expressed as: ;in, For the The damping compensation torque of each joint, is the damping coefficient, For the Angular velocity of each joint; dynamically adjust the damping coefficient of the mechanical arm joint damping compensation torque damping compensation according to the identified fault cause; No. Adjustment strategies for each joint , the corresponding torque damping compensation coefficient adjustment formula is: ;in, To adjust the strategy The calculated torque damping compensation coefficient adjustment amount; Apply the adjusted PID controller gain value and torque damping compensation coefficient to the robot arm control to obtain the adjusted control law: ;in, For the The controller output of each joint, For the The tracking error of each joint is calculated; the adjusted control law is applied to the robot control system to realize dynamic adjustment of the robot movement.

7. The method for adaptively adjusting the motion stability of a wafer cleaning robot arm according to claim 6, characterized in that: After the control is adjusted, the joint speed, vibration and end force data of the robot arm are continuously collected; the collected robot arm working data are feature analyzed to construct the working data feature vector; the stability index is calculated using the robot arm stability evaluation strategy; whether the robot arm movement is stable, that is, whether the stability index meets the threshold condition; If the robot arm is stable, the control adjustment is successful and the adjustment process ends; If the robot arm is unstable, the working data feature vector is input into the robot arm problem detection model to obtain the predicted fault cause label; According to the fault cause label, query the mapping relationship table between the fault cause and the adjustment strategy to obtain the corresponding adjustment strategy. According to the adjustment strategy, adjust the PID controller gain value and joint damping compensation coefficient of the robot arm to obtain a new control law. The new control law is applied to the robot control system, and the robot is repeatedly monitored and adjusted until the robot resumes stable motion. After the adjustment, If the robot arm still fails to reach a stable state after the preset maximum number of adjustments, an alarm will be triggered.

8. A wafer cleaning robot motion stability adaptive adjustment system, which is used to implement a wafer cleaning robot motion stability adaptive adjustment method according to any one of claims 1 to 7, characterized in that: include: Data acquisition module, stability assessment module, fault identification module, control and regulation module and cycle monitoring module; The data acquisition module is used to collect working data of the wafer cleaning robot arm, including joint speed, vibration data and end effector force data of the robot arm; The stability assessment module is used to perform feature analysis on the collected working data, construct a stability assessment strategy for the robotic arm, and use spectrum analysis and Fourier transform algorithms to perform real-time assessment of the motion state of the robotic arm to determine whether the current motion of the robotic arm is in a stable state; The fault identification module is used to construct and train a robot arm problem detection model. When the robot arm is in an unstable state, the cause of the instability of the robot arm is identified through the robot arm problem detection model. The control and adjustment module is used to construct a control algorithm for dynamically adjusting the robotic arm, and to adjust the robotic arm according to the cause of the unstable fault of the robotic arm, including adjusting the gain value of the robotic arm PID controller and adjusting the damping compensation of the robotic arm joint; The cyclic monitoring module continuously monitors whether the robotic arm is in a stable state after implementing control adjustment on the robotic arm. If the robotic arm is not in a stable state, the robotic arm is readjusted until the robotic arm is in a stable state.

9. An electronic device, characterized in that: include: A processor and a memory, wherein the memory stores a computer program that can be called by the processor; The processor executes the method for adaptively adjusting the motion stability of a wafer cleaning robot arm as described in any one of claims 1 to 7 by calling the computer program stored in the memory.

10. A computer-readable storage medium, characterized in that: Instructions are stored, and when the instructions are executed on a computer, the computer executes a method for adaptively adjusting the motion stability of a wafer cleaning robot arm as described in any one of claims 1 to 7.

Images