Industrial robot damping control system based on posture adjustment
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- GUANGZHOU XIAOAI CLASSMATES TECHNOLOGY CO LTD
- Filing Date
- 2026-05-22
- Publication Date
- 2026-08-07
AI Technical Summary
[0004]针对上述情况，为克服现有技术的缺陷，本发明提供一种基于姿态调整的工业机器人减震控制系统，通过姿态采集模块与特征提取模块融合加速度、力觉及关节运动数据，计算出多维度震动特征，实现震动状态精准感知；利用支持向量机分类器输出震动等级，结合模糊规则库匹配阻尼及刚度修正量，实现自适应决策；再通过PID型控制律动态调整增益并输出补偿力矩，叠加至原指令实现主动减震，解决了传统系统感知滞后、控制单一及易振荡的问题，显著提升了减震效果与作业精度
（1）本发明利用姿态采集模块中的六轴惯性测量单元、六维力/力矩传感器及增量式编码器，融合采集加速度、角速度、力觉及关节运动数据，并通过特征提取模块计算出时域、频域及统计特征，实现了对机器人震动状态的多维度精准感知，解决了传统系统震动状态判断滞后且不精确的技术问题，为后续自适应减震控制提供了可靠的数据基础。
Smart Images
Figure REF-OBJ-1779416370339-000007 
Figure REF-OBJ-1779416370339-000012 
Figure REF-OBJ-1779416370339-000024
Abstract
Description
Technical Field
[0001] This invention belongs to the field of industrial robot control technology, and in particular relates to a vibration reduction control system for industrial robots based on posture adjustment. Background Technology
[0002] Industrial robots are prone to vibration during high-speed start-up and shutdown, trajectory tracking, and contact operations due to factors such as joint clearance, load changes, and external impacts. Vibration not only affects the robot's positioning accuracy and work quality but also accelerates mechanical fatigue and shortens equipment lifespan. Existing industrial robot vibration reduction control methods mainly have the following technical problems requiring improvement: First, the vibration state perception capability is insufficient. Traditional systems typically rely solely on position and velocity signals fed back by joint encoders, lacking comprehensive acquisition of multi-dimensional information such as robot body acceleration and end-effector force. This makes it difficult to accurately identify the amplitude, frequency, and distribution characteristics of vibrations, resulting in delayed and inaccurate vibration state judgment.
[0003] Second, the vibration reduction control strategy is simplistic and lacks adaptability. Existing methods mostly employ fixed-parameter PID control or simple low-pass filtering, which cannot dynamically adjust the damping coefficient, stiffness coefficient, and compensation torque according to the real-time vibration level and motion mode. This results in unstable vibration reduction performance under complex working conditions and may even cause system oscillations. Summary of the Invention
[0004] To address the above issues and overcome the shortcomings of existing technologies, this invention provides an industrial robot vibration reduction control system based on posture adjustment. By fusing acceleration, force, and joint motion data through a posture acquisition module and a feature extraction module, multi-dimensional vibration characteristics are calculated, enabling precise vibration state perception. A support vector machine classifier outputs the vibration level, which is then combined with a fuzzy rule base to match damping and stiffness correction values, achieving adaptive decision-making. Furthermore, a PID control law dynamically adjusts the gain and outputs a compensation torque, which is then superimposed on the original command to achieve active vibration reduction. This solution addresses the problems of lag in perception, simplistic control, and oscillation inherent in traditional systems, significantly improving vibration reduction performance and operational accuracy.
[0005] The technical solution adopted in this invention is as follows: an industrial robot vibration reduction control system based on posture adjustment, including a posture acquisition module, a feature extraction module, a state classification module, a vibration reduction decision module, an actuator drive module, and a feedback recording module; The attitude acquisition module acquires real-time attitude data of each joint and base of the industrial robot, outputs the original attitude vector, and sends the original attitude vector to the feature extraction module. The feature extraction module receives the original pose vector, performs time-domain feature extraction, frequency-domain feature extraction, and statistical feature extraction, outputs a feature vector, and sends the feature vector to the state classification module. The state classification module receives the feature vector, uses a support vector machine-based classifier to identify the vibration state, outputs a vibration level label, and sends the vibration level label to the vibration reduction decision module. The vibration reduction decision module receives the vibration level label, combines it with the robot's current motion command, matches the corresponding vibration reduction control strategy through a fuzzy rule base, outputs the vibration reduction control parameters, and sends the vibration reduction control parameters to the actuator drive module. The actuator drive module receives the vibration reduction control parameters, converts them into current commands for the servo driver or opening commands for the hydraulic valve, and outputs drive signals to the joint actuators of the industrial robot. The feedback recording module collects attitude data and vibration level after shock absorption in real time, outputs shock absorption effect evaluation log, and stores the shock absorption effect evaluation log in the local database.
[0006] As a preferred technical solution of this invention, the attitude acquisition module includes a six-axis inertial measurement unit, a six-dimensional force / torque sensor, and an incremental encoder. The six-axis inertial measurement unit is installed on the robot base and the housing of each joint drive motor, acquiring three-axis linear acceleration components and three-axis angular velocity components. The six-dimensional force / torque sensor is installed between the robot's end effector and the wrist joint, acquiring force components in three directions and torque components in three directions. The incremental encoder is installed on the output shaft of each joint motor, acquiring the actual position angle and angular velocity of the joint. The attitude acquisition module synchronizes all sensor data through timestamps, aligns them to the same time series, and outputs the original attitude vector.
[0007] As a preferred technical solution of this scheme, the time-domain features extracted by the feature extraction module include the mean, standard deviation, root mean square value, peak value, peak-to-peak value of the triaxial acceleration signal within each time window, as well as the cumulative position deviation and velocity change rate of the joint angle signal; the extracted frequency-domain features include the energy value of the main frequency component, the frequency band energy ratio, and the centroid frequency of the spectrum obtained after performing a fast Fourier transform on the acceleration signal; the extracted statistical features include the skewness, kurtosis, waveform factor, and impulse factor of the force signal; the feature extraction module concatenates all features to form a high-dimensional feature vector and normalizes each feature.
[0008] As a preferred technical solution of this scheme, the state classification module adopts a support vector machine with radial basis function kernel. The decision function of the support vector machine is: calculate the Gaussian kernel similarity between the input feature vector and each support vector, add a bias term after weighted summation of the similarity, and then output the category decision value through a sign function. The state classification module maps the decision value to a vibration level label, which includes no vibration, mild vibration, moderate vibration and severe vibration.
[0009] As a preferred technical solution of this scheme, the vibration reduction decision module has a pre-stored fuzzy rule library. The rule form is: if the vibration level is a specified level and the current motion mode is a specified mode, then the damping coefficient correction and stiffness coefficient correction are output. For the vibration level label, a PID type vibration reduction control law is used to calculate the joint compensation torque. The PID type vibration reduction control law is: proportional gain multiplied by the deviation between the desired joint angle and the actual joint angle, plus differential gain multiplied by the rate of change of the deviation, plus integral gain multiplied by the integral of the deviation. The values of proportional gain, integral gain and differential gain are dynamically adjusted according to the vibration level label. The vibration reduction control parameters include the damping coefficient correction, stiffness coefficient correction and compensation torque value of each joint.
[0010] As a preferred technical solution of this scheme, the actuator drive module divides the compensation torque by the motor torque constant to obtain the current command for electric servo joints; and divides the compensation torque by the product of the pump displacement coefficient and the effective area of the hydraulic cylinder to obtain the servo valve opening command for hydraulic drive joints. The actuator drive module superimposes the current command or valve opening command onto the output signal of the original motion controller and sends it to each joint driver.
[0011] As a preferred technical solution of this scheme, the feedback recording module reads the root mean square value of triaxial acceleration, the position deviation of each joint, and the vibration level label after vibration reduction at a fixed sampling period; the feedback recording module maintains a sliding window internally to store the vibration amplitude sequence before and after vibration reduction; at the end of each sampling period, the feedback recording module calculates the average vibration amplitude before and after vibration reduction within the current window, and calculates the vibration reduction effectiveness; the vibration reduction effectiveness is calculated as follows: the vibration reduction effectiveness equals the average vibration amplitude before vibration reduction minus the average vibration amplitude after vibration reduction, divided by the average vibration amplitude before vibration reduction. The amplitude of the movement is then multiplied by 100%. The feedback recording module packages the damping efficiency, vibration level label, maximum joint position deviation, timestamp, and working condition identifier into a log entry. When the accumulated number of log entries reaches a preset number, the feedback recording module triggers a storage operation to write the log entries into the local database in batches. The feedback recording module also provides a data query interface for external devices to read historical damping effect evaluation logs. When the damping efficiency is lower than a preset threshold for multiple consecutive sampling periods, the feedback recording module generates a damping effect decline warning signal and sends it to the human-machine interface of the robot controller.
[0012] The beneficial effects of the present invention after adopting the above structure are as follows: (1) This invention utilizes a six-axis inertial measurement unit, a six-dimensional force / torque sensor and an incremental encoder in the attitude acquisition module to integrate and acquire acceleration, angular velocity, force and joint motion data, and calculates time domain, frequency domain and statistical features through the feature extraction module, thereby realizing multi-dimensional and accurate perception of the robot's vibration state. This solves the technical problem of lagging and inaccurate vibration state judgment in traditional systems and provides a reliable data foundation for subsequent adaptive vibration reduction control.
[0013] (2) The present invention uses a support vector machine-based classifier in the state classification module to map the feature vectors into labels for no vibration, light vibration, moderate vibration and heavy vibration levels. Combined with the fuzzy rule library in the vibration reduction decision module, the corresponding damping coefficient correction amount and stiffness coefficient correction amount are matched according to the vibration level and motion mode. This realizes intelligent decision-making of vibration reduction control strategy, solves the technical problems of poor adaptability and unstable vibration reduction effect of traditional fixed parameter control method, and significantly improves the vibration reduction robustness under different working conditions.
[0014] (3) This invention utilizes the PID-type vibration reduction control law in the vibration reduction decision module to dynamically adjust the proportional gain, integral gain and derivative gain according to the vibration level, calculate the joint compensation torque in real time, and convert the compensation torque into a current command or valve opening command by the actuator drive module and superimpose it on the output of the original motion controller, thereby realizing active vibration reduction control. This solves the technical problem of system oscillation that is easy to be generated in high-speed start-stop and contact operations by traditional methods, effectively reduces the vibration amplitude of the robot, and improves the operation accuracy and equipment life. Attached Figure Description
[0015] The accompanying drawings are provided to further understand the present invention and form part of the specification. They are used together with the embodiments of the present invention to explain the invention and do not constitute a limitation thereof.
[0016] Figure 1 This is a system overall data flow diagram of the posture adjustment-based industrial robot vibration reduction control system proposed in this invention; Figure 2 This is a flowchart of feature extraction and vibration recognition for the posture adjustment-based industrial robot vibration reduction control system proposed in this invention. Detailed Implementation
[0017] Example 1, see Figures 1-2 The present invention provides an industrial robot vibration reduction control system based on posture adjustment, including a posture acquisition module, a feature extraction module, a state classification module, a vibration reduction decision module, an actuator drive module, and a feedback recording module. The attitude acquisition module acquires real-time attitude data of each joint and base of the industrial robot, including three-axis acceleration, three-axis angular velocity and force data of the end effector, and outputs the original attitude vector; and sends the original attitude vector to the feature extraction module. The feature extraction module receives the original pose vector, performs time-domain feature extraction, frequency-domain feature extraction, and statistical feature extraction, and outputs a feature vector; then sends the feature vector to the state classification module. The state classification module receives the feature vector, uses a support vector machine-based classifier to identify the vibration state, outputs a vibration level label, and sends the vibration level label to the vibration reduction decision module. The vibration reduction decision module receives vibration level labels, combines them with the robot's current motion commands, matches the corresponding vibration reduction control strategy using a fuzzy rule base, outputs vibration reduction control parameters, and sends the vibration reduction control parameters to the actuator drive module. The actuator drive module receives vibration damping control parameters, converts them into current commands for the servo drive or opening commands for the hydraulic valve, and outputs drive signals to the joint actuators of the industrial robot. The feedback recording module collects attitude data and vibration level after vibration reduction in real time, outputs vibration reduction effect evaluation log, and stores the vibration reduction effect evaluation log in the local database.
[0018] Example 2, based on the above examples, describes an example where the attitude acquisition module obtains real-time attitude data of each joint and base of the industrial robot, specifically including: A six-axis inertial measurement unit (IMU) is used, mounted on the robot base and the housing of the drive motors for each joint, to collect the triaxial linear acceleration components of the robot body in space. and triaxial angular velocity components It outputs inertial measurement data with a sampling frequency of no less than 100Hz; it employs a six-dimensional force / torque sensor, installed between the robot's end effector and wrist joint, to collect force components in three directions generated by the end tool's contact with the environment. and torque components in three directions It outputs force data; incremental encoders are installed on the output shafts of each joint motor to collect the actual position and angle of the joint. and angular velocity Output joint motion data.
[0019] The attitude acquisition module synchronizes all sensor data with timestamps, aligns all data to the same time series based on the robot controller clock, and outputs the original attitude vector, which contains inertial data, force data, and joint motion data.
[0020] Example 3, based on the above examples, involves a feature extraction module that receives the original pose vector, performs time-domain feature extraction, frequency-domain feature extraction, and statistical feature extraction, and outputs a feature vector, specifically including: Temporal feature extraction: For the triaxial acceleration signals within each time window, the mean, standard deviation, root mean square (RMS) value, peak value, and peak-to-peak value are calculated to extract temporal features reflecting vibration amplitude. For the joint angle signals, the cumulative position deviation and rate of change of velocity are calculated to extract features reflecting trajectory tracking error. The formula for calculating the RMS value of the acceleration signal is as follows: in, For the first The acceleration amplitude at each sampling point This represents the average acceleration within that time window. This represents the number of sampling points within the window. This formula is used to quantify the energy intensity of the vibration signal; a larger root mean square value indicates a more severe vibration amplitude.
[0021] Frequency domain feature extraction: Perform fast Fourier transform on the acceleration signal to extract the energy value of the main frequency component, the energy ratio of the frequency band (0-5Hz low frequency band, 5-20Hz mid frequency band, 20-50Hz high frequency band) and the centroid frequency of the spectrum, and extract the features that reflect the frequency distribution of vibration.
[0022] Statistical feature extraction: skewness, kurtosis, waveform factor and impulse factor are calculated for the force signal within the entire window to extract features reflecting impact and abnormal contact; all the above features are concatenated to form a high-dimensional feature vector with a dimension of 45, and each feature is normalized (zero mean unit variance normalization) to eliminate the influence of dimensions.
[0023] Example 4, based on the above examples, involves a state classification module receiving feature vectors, using a support vector machine-based classifier for vibration state identification, and outputting vibration level labels. Specifically, this includes: The state classification module pre-builds a classification model using offline training. Attitude data of the robot under different operating conditions (stable operation without load, load start-up, emergency braking, and collision contact) are collected, and vibration level labels are manually added (0-no vibration, 1-mild vibration, 2-moderate vibration, 3-severe vibration). A support vector machine with radial basis function kernel is used to learn from the training samples. The decision function is: in, The current input feature vector, For the first Support vectors, The corresponding support vectors are labeled with category labels (vibration levels are encoded using a one-to-many strategy). For Lagrange multipliers, Kernel function bandwidth parameter, For bias terms, The total number of support vectors is denoted by . This formula calculates the Gaussian kernel similarity between the input feature and all support vectors, sums them by weight, and outputs the class decision value via a sign function.
[0024] The state classification module maps the decision values to vibration level labels (0, 1, 2, 3) and outputs them to the vibration reduction decision module.
[0025] Example 5, based on the above examples, involves the vibration reduction decision module receiving vibration level labels, combining them with the robot's current motion commands, matching the corresponding vibration reduction control strategy using a fuzzy rule base, and outputting vibration reduction control parameters, specifically including: The vibration reduction decision module has a pre-stored fuzzy rule base. The rule format is "If the vibration level is L and the current motion mode is M, then output the damping coefficient correction amount". and stiffness coefficient correction amount For vibration level L, a PID-type vibration reduction control law is used to calculate the joint compensation torque. The calculation formula is: in, The deviation between the desired joint angle and the actual joint angle. , , These are the proportional, integral, and differential gain coefficients, respectively. , , The value of is dynamically adjusted by looking up a table based on the vibration level label (the higher the vibration level, the greater the proportional gain and differential gain to enhance the suppression capability). This formula is used to calculate the compensation torque that needs to be superimposed on the original motion command in real time, thereby actively attenuating the vibration.
[0026] The vibration reduction decision module outputs vibration reduction control parameters, including the damping coefficient correction, stiffness coefficient correction, and compensation torque value of each joint, and sends the vibration reduction control parameters to the actuator drive module.
[0027] Example 6, based on the above examples, involves the actuator drive module receiving vibration damping control parameters, converting them into current commands for the servo driver or opening commands for the hydraulic valve, and outputting drive signals to the joint actuators of the industrial robot. Specifically, this includes: For electric servo joints, the actuator drive module will compensate for torque. Convert to current command The relationship is transformed into ,in This represents the motor torque constant; for hydraulically driven joints, the actuator drive module converts the compensating torque into a servo valve opening command. The relationship is transformed into ,in This is the pump displacement coefficient. This represents the effective area of the hydraulic cylinder.
[0028] The actuator drive module superimposes the current command or valve opening command onto the output signal of the original motion controller and sends it to each joint driver via fieldbus (EtherCAT and CANopen) to achieve vibration reduction control.
[0029] The feedback recording module collects the root mean square value of acceleration, joint position deviation and vibration level after vibration reduction in real time, compares it with the data before vibration reduction, calculates the vibration reduction effectiveness ((vibration amplitude before vibration reduction - vibration amplitude after vibration reduction) / vibration amplitude before vibration reduction × 100%), generates a vibration reduction effect evaluation log, and stores it in the local database for subsequent controller parameter self-tuning and fault analysis.
[0030] Example 7, based on the above examples, involves a feedback recording module that collects post-damping attitude data and vibration levels in real time, outputs a damping effect evaluation log, and stores the damping effect evaluation log in a local database. Specifically, this includes: The feedback recording module reads the root mean square value of triaxial acceleration, the position deviation of each joint, and the vibration level label output by the state classification module in real time at a fixed sampling period (sampling frequency not less than 100Hz). The feedback recording module maintains a sliding window of length L (L is 1000 sampling points, corresponding to 10 seconds of data) to store the vibration amplitude sequence before and after vibration reduction.
[0031] At the end of each sampling period, the feedback recording module calculates the average vibration amplitude before damping within the current window. and the average vibration amplitude after shock absorption The vibration amplitude before damping is taken from the root mean square value of acceleration before the actuator drive module outputs the compensation torque, and the vibration amplitude after damping is taken from the root mean square value of acceleration after the actuator drive module outputs the compensation torque. The formula for calculating the damping efficiency η is: in, The root mean square average of the acceleration within the window before vibration reduction. This represents the root mean square average of the acceleration within the window after vibration reduction. The range of values is arrive , A higher value indicates a more significant shock absorption effect.
[0032] The feedback recording module packages the calculated damping efficiency, current vibration level label, maximum joint position deviation, timestamp, and working condition identifier (the current motion mode code provided by the robot controller) into a single log entry. When the accumulated log entries reach 100, the feedback recording module triggers a storage operation, writing the log entries in batches to the damping effect evaluation table in the local database.
[0033] The feedback recording module also provides a data query interface, allowing external host computers or maintenance terminals to read historical vibration damping effect evaluation logs for subsequent controller parameter self-tuning and fault analysis. When the vibration damping effectiveness rate is lower than a preset threshold (taken as 50%) for 10 consecutive sampling cycles, the feedback recording module generates a vibration damping effect decline warning signal and sends it to the robot controller's human-machine interface via the fieldbus.
[0034] Through the above design, the feedback recording module realizes real-time quantitative evaluation of vibration reduction effect, persistent storage of historical data, and abnormal early warning functions, providing data support for the long-term stable operation of the vibration reduction control system.
Claims
1. A pose adjustment based industrial robot damping control system, characterized by: The system includes an attitude acquisition module, a feature extraction module, a state classification module, a vibration reduction decision module, an actuator driving module, and a feedback recording module. The attitude acquisition module acquires real-time attitude data of each joint and base of the industrial robot, outputs the original attitude vector, and sends the original attitude vector to the feature extraction module. The feature extraction module receives the original pose vector, performs time-domain feature extraction, frequency-domain feature extraction, and statistical feature extraction, outputs a feature vector, and sends the feature vector to the state classification module. The state classification module receives the feature vector, uses a support vector machine-based classifier to identify the vibration state, outputs a vibration level label, and sends the vibration level label to the vibration reduction decision module. The vibration reduction decision module receives the vibration level label, combines it with the robot's current motion command, matches the corresponding vibration reduction control strategy through a fuzzy rule base, outputs the vibration reduction control parameters, and sends the vibration reduction control parameters to the actuator drive module. The actuator drive module receives the vibration reduction control parameters, converts them into current commands for the servo driver or opening commands for the hydraulic valve, and outputs drive signals to the joint actuators of the industrial robot. The feedback recording module collects attitude data and vibration level after shock absorption in real time, outputs shock absorption effect evaluation log, and stores the shock absorption effect evaluation log in the local database.
2. The industrial robot vibration reduction control system based on attitude adjustment according to claim 1, characterized in that: The attitude acquisition module includes: A six-axis inertial measurement unit is installed on the robot base and the housing of each joint drive motor. It collects three-axis linear acceleration components and three-axis angular velocity components and outputs inertial measurement data. A six-dimensional force / torque sensor is installed between the robot's end effector and wrist joint to collect force components in three directions and torque components in three directions, and output force data. Incremental encoders are installed on the output shafts of each joint motor to collect the actual position angle and angular velocity of the joint and output joint motion data. The attitude acquisition module synchronizes all sensor data through timestamps, aligns them to the same time series, and outputs the original attitude vector.
3. The industrial robot vibration reduction control system based on attitude adjustment according to claim 2, characterized in that: The feature extraction module extracts the following time-domain features: mean, standard deviation, root mean square value, peak value, peak-to-peak value of the triaxial acceleration signal within each time window, as well as the cumulative position deviation and velocity change rate of the joint angle signal; the extracted frequency-domain features include the energy value of the main frequency component, the frequency band energy ratio, and the centroid frequency of the spectrum obtained after performing a fast Fourier transform on the acceleration signal; and the extracted statistical features include the skewness, kurtosis, waveform factor, and impulse factor of the force sensor signal. The feature extraction module concatenates all features to form a high-dimensional feature vector and normalizes each feature.
4. The industrial robot vibration reduction control system based on attitude adjustment according to claim 3, characterized in that: The state classification module employs a support vector machine with a radial basis function kernel. The decision function of the support vector machine is as follows: calculate the Gaussian kernel similarity between the input feature vector and each support vector, perform a weighted summation of the similarities, add a bias term, and then output the class decision value through a sign function. The weights of the weighted summation are the product of the Lagrange multiplier and the corresponding support vector class label. The Gaussian kernel similarity is calculated using the kernel function bandwidth parameter. The state classification module maps the decision value to a vibration level label, which includes no vibration, mild vibration, moderate vibration, and severe vibration.
5. The industrial robot vibration reduction control system based on attitude adjustment according to claim 4, characterized in that: The vibration reduction decision module has a pre-stored fuzzy rule library. The rule format is as follows: if the vibration level is a specified level and the current motion mode is a specified mode, then the damping coefficient correction and stiffness coefficient correction are output. For the vibration level label, a PID-type vibration reduction control law is used to calculate the joint compensation torque. The PID-type vibration reduction control law is: proportional gain multiplied by the deviation between the desired joint angle and the actual joint angle, plus differential gain multiplied by the rate of change of the deviation, plus integral gain multiplied by the integral of the deviation. The values of proportional gain, integral gain, and differential gain are dynamically adjusted according to the vibration level label. The vibration reduction control parameters include the damping coefficient correction, stiffness coefficient correction, and compensation torque value of each joint.
6. The industrial robot vibration reduction control system based on attitude adjustment according to claim 5, characterized in that: For electric servo joints, the actuator drive module divides the compensation torque by the motor torque constant to obtain the current command; for hydraulically driven joints, it divides the compensation torque by the product of the pump displacement coefficient and the effective area of the hydraulic cylinder to obtain the servo valve opening command; the actuator drive module superimposes the current command or valve opening command onto the output signal of the original motion controller and sends it to each joint driver.
7. The industrial robot vibration reduction control system based on attitude adjustment according to claim 6, characterized in that: The feedback recording module collects the root mean square value of acceleration, joint position deviation and vibration level after vibration reduction in real time, compares it with the data before vibration reduction, calculates the vibration reduction effectiveness, generates a vibration reduction effect evaluation log, and stores it in the local database.
8. The industrial robot vibration reduction control system based on attitude adjustment according to claim 7, characterized in that: The feedback recording module reads the root mean square value of acceleration, joint position deviation, and vibration level label after vibration reduction at a fixed sampling period, and maintains a sliding window to store the vibration amplitude sequence before and after vibration reduction. At the end of each sampling period, it calculates the average vibration amplitude before and after vibration reduction, and then calculates the vibration reduction efficiency. The vibration reduction efficiency is calculated as the average vibration amplitude before vibration reduction minus the average vibration amplitude after vibration reduction, divided by the average vibration amplitude before vibration reduction, and then multiplied by 100%. The feedback recording module packages the vibration reduction efficiency, vibration level label, maximum joint position deviation, timestamp, and working condition identifier into log entries, and writes them to the local database in batches when the accumulated entries reach a preset number. When the vibration reduction efficiency is lower than a preset threshold for multiple consecutive sampling periods, a vibration reduction effect decline warning signal is generated and sent to the human-machine interface.