Mechanical arm tail end active vibration suppression method and system based on predictive compensation

Abstract

The invention relates to the technical field of mechanical arm vibration suppression, in particular to a mechanical arm tail end active vibration suppression method and system based on predictive compensation, and the method comprises the following steps: S1, collecting the three-axis acceleration, angular velocity and vibration spectrum data of a mechanical arm in real time through a high-frequency IMU sensor; monitoring a rolling angle, a pitch angle and a yaw angle of a base attitude through an auxiliary IMU sensor, and fusing double IMU data through a Kalman filtering algorithm; s2, generating a pre-compensation torque based on the fused data, optimizing a model gain coefficient in combination with a real-time deviation between a frequency spectrum actually measured by a high-frequency IMU sensor and a model prediction frequency spectrum, and correcting and outputting the pre-compensation torque in real time; and S3, the corrected compensation torque is injected into a feed-forward channel of a joint motor control loop, vibration suppression torque is output to a joint execution mechanism through a high-gain controller, and residual vibration at the tail end of the mechanical arm is counteracted in real time. According to the invention, an active vibration suppression system can be constructed to solve the problem of real-time vibration suppression under multi-source disturbance coupling.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of mechanical arm vibration suppression, and in particular to a mechanical arm end active vibration suppression method and system based on predictive compensation. BACKGROUND

[0002] As a core execution mechanism of industrial automation, the end positioning accuracy of a mechanical arm directly affects the process quality of machining, assembly and the like. Under high-speed motion or direction changing conditions, the mechanical arm will cause three types of vibrations due to its inherent structural characteristics: 1. Joint transmission flexible vibration: Joint transmission flexible vibration is caused by the elastic deformation characteristics of a harmonic reducer. As a core transmission component of the mechanical arm, the flexspline of the harmonic reducer will periodically elastically deform under high-speed start-stop conditions, resulting in a micron-level angle lag between the output shaft and the input shaft of the joint. This lag energy is transmitted to the end through the connecting rod, forming low-frequency residual vibration and affecting the end positioning accuracy of the mechanical arm. 2. Mechanical resonance: Mechanical resonance is caused by the frequency domain coupling of lightweight structure and motion excitation. Modern mechanical arms widely use carbon fiber composite materials to reduce weight, but the inherent modal density of thin-walled structures significantly increases in the 80-150Hz frequency band. When the acceleration in the trajectory planning suddenly changes or the external excitation frequency approaches the natural frequency, the system kinetic energy is converted into deformation potential energy, forming a standing wave at a specific node, with an amplitude amplification of more than 3 times, resulting in displacement deviation in the vibration energy concentration area of the mechanical arm end. 3. Multi-source coupled vibration: Multi-source coupled vibration is the nonlinear superposition result of base disturbance, load mutation and thermal deformation. Base disturbance is transmitted through the kinematic chain. According to the DH coordinate system conversion model, the base attitude angle deviation is amplified through the Jacobian matrix, and the end pose error can reach 3mm. Load mutation causes joint torque transients, breaking the dynamic balance and producing 2-10Hz low-frequency shaking with a duration of up to 400ms. Thermal deformation is caused by the thermal expansion characteristics of the material. After continuous operation for 4 hours, the connecting rod is thermally elongated by 0.1mm, changing the inherent frequency of the structure and making the original vibration suppression strategy invalid. After coupling, the vibration energy will be greatly increased, further reducing the positioning accuracy.

[0003] By searching, Chinese patent publication number CN106737857B discloses a mechanical arm end vibration suppression method, including the following steps: collecting the acceleration and / or amplitude of the mechanical arm end; when the acceleration is greater than or equal to the critical acceleration, slowing down the movement of the mechanical arm until the acceleration is a safe acceleration; and / or, when the amplitude is greater than or equal to the critical amplitude, slowing down the movement of the mechanical arm until the amplitude is a safe amplitude; when the acceleration is greater than or equal to the alarm acceleration, stopping the movement of the mechanical arm and sounding an alarm; and / or, when the amplitude is greater than or equal to the critical amplitude, stopping the movement of the mechanical arm and sounding an alarm. By using the above method, the vibration of the end of the multi-joint mechanical arm can be effectively and quantitatively suppressed, ensuring that the vibration of the mechanical arm end is within a reasonable range, effectively protecting the detection quality of the X-ray machine and the safety of the equipment.

[0004] By searching, Chinese patent publication number CN114347018B discloses a mechanical arm disturbance compensation method based on a wavelet neural network, mainly divided into two parts: disturbance signal prediction and feedforward feedback compensation. The disturbance signal prediction part aims to improve the prediction accuracy of the disturbance by using a wavelet neural network analysis of the time-varying near-periodic disturbance signal online prediction model, which improves the prediction accuracy of the disturbance; the feedforward feedback compensation part aims to improve the positioning accuracy of the robot end by using a feedforward feedback joint compensation control method, which calculates the joint compensation angle and adds it to the feedforward control system to improve the compensation effect and thus improve the positioning accuracy of the mechanical arm end.

[0005] Although the prior art 1 (CN106737857B) and the prior art 2 (CN114347018B) solve the mechanical arm vibration problem on one side, but under high speed or variable load working conditions, the above prior art still has the following technical defects: 1. Vibration monitoring and compensation are disconnected: the passive vibration suppression strategy of the prior art 1 relies on the end acceleration and amplitude threshold to trigger deceleration or stop, and the feedback delay causes the suppression of >100Hz high frequency vibration to fail; the prior art 2 predicts the base disturbance through a wavelet neural network, but the base IMU cannot capture the high frequency residual vibration at the end, and the joint angle compensation needs to be converted through the Jacobian matrix, introducing a 3-5ms calculation delay, which is difficult to match the phase requirement of the vibration energy concentration area, thus failing to solve the real-time vibration suppression problem; 2. Insufficient adaptability of the dynamics model: the offline trained neural network weight of the prior art 2 is fixed and cannot be corrected in real time, which causes the stiffness matrix to drift due to sudden load changes or the size of the connecting rod to change due to thermal deformation, resulting in a mismatch between the compensation command and the real dynamics parameters, thus failing to solve the real-time vibration suppression problem; 3. The vibration suppression execution real-time defect: the prior art lacks a direct countermechanism for high-frequency vibration, such as the vibration suppression of the prior art 1 which relies on the overall speed reduction of the mechanical arm, and the joint angle feedforward of the prior art 2 which needs to pass through the motor PI control loop and lacks sufficient response bandwidth, thus being unable to cover the three times frequency domain of the vibration base frequency (such as 150Hz resonance which requires a control bandwidth of more than 450Hz), and thus being unable to solve the real-time vibration suppression problem. SUMMARY

[0006] In order to construct an active vibration suppression system to solve the real-time vibration suppression problem under the coupling of the above-mentioned multiple source disturbances, the application provides a mechanical arm end active vibration suppression method and system based on predictive compensation.

[0007] In the first aspect, the application provides a mechanical arm end active vibration suppression method based on predictive compensation, which adopts the following technical scheme: a mechanical arm end active vibration suppression method based on predictive compensation, comprising the following steps: S1, vibration data cooperative acquisition and fusion: install a high-frequency IMU sensor on the flange at the end of the mechanical arm to real-time collect three-axis acceleration, angular velocity and vibration frequency spectrum data of the mechanical arm; at the same time, install an auxiliary IMU sensor on the base of the mechanical arm to synchronously monitor the roll angle, pitch angle and yaw angle of the base attitude; then fuse the high-frequency IMU sensor data and the auxiliary IMU sensor data through Kalman filtering algorithm to improve the signal-to-noise ratio and realize time and space synchronization; S2, structure flexible vibration suppression: based on the vibration data fused in step S1, generate a pre-compensation torque through a rigid-flexible coupled dynamics model of the mechanical arm, and combine the real-time deviation of the measured frequency spectrum of the high-frequency IMU sensor and the model predicted frequency spectrum to dynamically optimize the gain coefficient in the rigid-flexible coupled dynamics model through Lyapunov stability algorithm, real-time correct the output accuracy of the pre-compensation torque, and output the corrected compensation torque; S3, vibration suppression instruction execution: inject the corrected compensation torque output in step S2 into the feedforward channel of the joint motor control loop, output the vibration suppression torque to the joint actuator through a high-gain controller with a response bandwidth covering the three times frequency domain of the vibration base frequency, and real-time cancel the residual vibration at the end of the mechanical arm.

[0008] Optionally, the double-IMU data fusion method of step S1 comprises the following steps: S11, based on the precise time synchronization protocol, realize the microsecond-level time alignment of the high-frequency IMU sensor and the auxiliary IMU sensor, and establish the time domain basis for the subsequent frequency domain processing; S12, using the synchronized data in step S11, perform wavelet threshold denoising processing on the original data collected by the high-frequency IMU sensor to eliminate high-frequency interference components above 200Hz, and simultaneously separate the effective vibration signal and the environmental temperature drift noise by using the principal component analysis algorithm to output the pure vibration signal; S13, calculate the real-time vibration energy value based on the pure vibration signal output in step S12, dynamically allocate data weight according to the energy threshold, and the specific method is as follows: When the vibration energy is greater than 0.1 times the gravity acceleration, the high-frequency IMU sensor data is used as the dominant output; When the vibration energy is less than or equal to 0.1 times the gravity acceleration, switch to the auxiliary IMU sensor data for compensation output.

[0009] Optionally, the pre-compensation torque generation of step S2 is step S21, which includes the following steps: S211, based on the vibration data fused in step S1, perform a harmonic reducer step loading experiment in a constant temperature environment of 25℃ to establish a three-dimensional mapping relationship between joint stiffness, load and speed; when the joint speed is greater than 1m / s, the following dynamic adjustment is executed based on the basic mapping relationship: According to the speed increment gradient, the stiffness compensation coefficient is improved to enhance the joint anti-deformation ability to suppress high-frequency vibration; According to the inverse function of joint speed, the pre-tightening force output value is reduced to avoid the risk of resonance caused by over-constraint; Refresh the pre-tightening force reference value every 100ms, fuse the stiffness compensation coefficient and the pre-tightening force output value to generate real-time stiffness compensation parameters; S212, based on the real-time stiffness compensation parameters generated in step S21, identify the resonance frequency band of 80 to 150Hz through finite element modal analysis, and locate the vibration energy concentration area of the robot arm; based on the diagnosis result of this resonance frequency band, inject band-stop filter instructions into the pre-compensation torque to suppress the transmission of resonance energy to the end of the robot arm; S213, dynamically couple the real-time stiffness compensation parameters of step S21 and the band-stop filter instructions of step S22 to form a composite vibration suppression strategy of time-domain stiffness enhancement and frequency-domain energy blocking, and verify the convergence of the composite vibration suppression strategy through Lyapunov stability condition to ensure that the torque command input by the gain correction module meets the global asymptotic stability, and then optimize the composite vibration suppression strategy based on the verification result, and the specific method is as follows: When high-frequency vibration dominates, strengthen the stiffness compensation weight; When the resonance frequency band is concentrated, increase the band-stop filter weight; S214, encapsulate the optimized composite vibration suppression strategy as a pre-compensation torque command package, and transmit it to the gain correction module through a standard industrial bus protocol.

[0010] Optionally, step S2 further includes step S22 to suppress trajectory discontinuous vibration, which includes the following steps: S221, re-plan the target motion trajectory of the robot arm using a quintic polynomial, limit the acceleration to within 300m / s³ through a time scaling algorithm, and eliminate the acceleration step at the trajectory corner; S222, a sinusoidal function transition segment is inserted at the trajectory corner to make the acceleration curve smooth and gradually change, thereby directly reducing the impact vibration component in the high-frequency IMU sensor data of the input step S1; S223, the optimized trajectory instruction is input to the vibration data cooperative acquisition module of step S1 as a motion reference for high-frequency IMU sensor data acquisition, and simultaneously links the structure flexible vibration suppression module of step S2 to dynamically generate a pre-compensation torque based on the trajectory characteristics. The specific method is as follows: The stiffness compensation weight is strengthened in the high-speed section to enhance the joint deformation resistance; The band-stop filtering instruction strength is improved in the corner section to block the resonance energy transmission.

[0011] Optionally, the step S2 further includes a step S23 of suppressing base disturbance transmission, which includes the following steps: S231, the roll angle, pitch angle and yaw angle deviation of the base attitude are measured in real time by the auxiliary IMU sensor of step S1; S232, a kinematics conversion model of the mechanical arm is established based on space vector chain multiplication, the end pose offset caused by the base attitude angle deviation of step S231 is mapped into a joint space compensation torque through the Jacobian matrix transpose, and the kinematics transmission error of the mechanical arm end pose caused by the base disturbance is eliminated; S233, the joint space compensation torque of step S232 is superimposed on the corrected compensation torque output by step S2 in proportion, wherein the gain coefficient of step S2 is in positive proportion to the attitude angle deviation of step S231; then the superimposed total compensation torque is input to the high-gain controller of step S3 in real time to offset the end drift caused by the base disturbance through the feedforward channel of step S3, thereby achieving base disturbance suppression.

[0012] Optionally, the step S2 further includes a step S24 of suppressing load mutation vibration, which includes the following steps: S241, the load change rate is detected in real time by the torque sensor installed at the output end of the joint harmonic reducer, and when the load change rate is detected to be greater than 50 N·m / ms, the stiffness compensation coefficient of step S211 is improved in proportion to the load mutation strength based on the three-dimensional mapping relationship between the joint stiffness and the load and the speed established in step S211, and the stiffness matrix value in the rigid-flexible coupled dynamics model of step S2 is updated synchronously, so that the rigid-flexible coupled dynamics model matches the mutation working condition in real time and eliminates the low-frequency resonance caused by dynamics mismatch; S242, the composite vibration suppression strategy generation module of step S213 is regenerated based on the stiffness matrix value updated in step S241 to generate a pre-compensation torque, the high-gain controller of step S3 is used to spread the spectrum of the pre-compensation torque in the frequency domain to cover the low-frequency shaking frequency band, and then the spread spectrum vibration suppression torque is injected into the joint actuator of step S3 through the feedforward channel of step S3 to real-time offset the low-frequency shaking caused by the load mutation and suppress the residual amplitude.

[0013] Optionally, the step S2 further comprises a step S25 of kinematic model thermal deformation correction, which comprises the following steps: S251, a network of temperature sensors is deployed on the joints and links of the mechanical arm to monitor the temperature change of -10℃ to 60℃ in real time, the thermal deformation amount of the links of the mechanical arm is calculated based on the thermal expansion coefficient of the material, the influence of thermal deformation on the vibration data acquisition accuracy of step S1 is analyzed, and the end positioning error caused by thermal drift is quantified; S252, the link size parameters in the rigid-flexible coupled dynamics model of step S2 are adjusted in proportion according to the thermal deformation amount, the geometric parameters in the three-dimensional mapping relationship between joint stiffness and load and speed established in step S211 are updated synchronously, the kinematic model mismatch caused by thermal deformation is eliminated, and the model prediction spectrum shift is suppressed; S253, the core parameters of the rigid-flexible coupled kinematic model database of step S2 are automatically calibrated every 30 minutes, the composite vibration suppression strategy generation module of step S213 is regenerated based on the updated rigid-flexible coupled kinematic model of step S2 to generate a pre-compensation torque, and the high-gain controller of step S3 is used to real-time offset the end low-frequency drift caused by thermal deformation.

[0014] Optionally, after step S3, step S4 of vibration suppression effect closed-loop verification is performed, which comprises the following steps: S41, based on the high-frequency IMU sensor data of step S1 and the pure vibration signal output by step S12, the acceleration amplitude before and after vibration suppression is extracted to calculate the vibration suppression rate, when the vibration suppression rate is less than 70%, step S42 is performed, and when the vibration suppression rate is greater than or equal to 90%, step S43 is performed; S42, the sampling frequency of the high-frequency IMU sensor and the auxiliary IMU sensor of step S1 is increased, the microsecond-level time synchronization accuracy of the high-frequency IMU sensor and the auxiliary IMU sensor is enhanced, the dynamic adjustment frequency of the stiffness compensation weight and the band rejection filter strength of step S213 is increased to improve the convergence speed of the composite vibration suppression strategy, and the trajectory re-planning module of step S221 is activated to reduce the joint speed of the mechanical arm to a safety threshold; S43, the torque output of the joint motor is reduced through the standard industrial bus protocol of step S214, the system total power consumption is reduced while maintaining the vibration suppression rate greater than 85%.

[0015] In a second aspect, the application provides a mechanical arm end active vibration suppression system, which adopts the following technical scheme: a mechanical arm end active vibration suppression system, comprising: A multi-sensor fusion module, which comprises a high-frequency IMU sensor, an auxiliary IMU sensor and a joint strain sheet network, the high-frequency IMU sensor is installed on the mechanical arm end flange, and is used for real-time collection of three-axis acceleration, angular velocity and vibration frequency spectrum data; the auxiliary IMU sensor is installed on the mechanical arm base, and is used for synchronous monitoring of roll angle, pitch angle and yaw angle; the joint strain sheet network is used for forming a dual-band monitoring network, and is used for covering 0-500Hz vibration frequency domain; A double-loop control module, which comprises a prediction unit and a correction unit, the prediction unit is used for storing a rigid-flexible coupling dynamics model of the mechanical arm, and outputs a pre-compensation torque instruction; the correction unit dynamically optimizes gain coefficients in the rigid-flexible coupling dynamics model through Lyapunov stability algorithm; An execution module, which comprises a high-gain controller, the high-gain controller is used for outputting a vibration suppression torque to a joint execution mechanism, and a response bandwidth covers a vibration base frequency three times frequency domain.

[0016] In summary, the application has the following beneficial technical effects: The application constructs a high-precision vibration monitoring network through double-IMU collaborative collection and Kalman filter fusion, accurately captures three-axis acceleration, angular velocity and 0-500Hz vibration frequency spectrum at the mechanical arm end, and eliminates base disturbance interference at the same time; based on a rigid-flexible coupling dynamics model and Lyapunov real-time optimization, a pre-compensation torque is dynamically generated and output precision is corrected, effectively inhibiting joint flexible deformation and mechanical resonance; through a high-gain feedforward controller, a vibration suppression torque is directly injected into a joint execution mechanism, real-time residual vibration in a 2-150Hz frequency band is offset, stability and processing quality under high-speed running conditions are significantly improved, thereby an active vibration suppression system is constructed, to solve the real-time vibration suppression problem under the coupling of multiple source disturbances; In the multi-source perception layer, a high-precision vibration monitoring network is constructed through double-IMU collaborative collection and Kalman filter fusion, a microsecond-level time synchronization protocol eliminates phase deviation caused by transmission delay, sym8 wavelet basis 5-layer decomposition and soft threshold processing are combined to suppress high-frequency interference above 200Hz, PCA principal component analysis separates environmental temperature drift noise, and outputs a pure vibration signal with residual noise less than 0.05g; a dynamic weight distribution mechanism based on vibration energy threshold retains high-frequency vibration details at the end under high-energy conditions, and suppresses base micro-vibration interference under low-energy conditions, providing a high-reliability data basis for vibration suppression decision-making; In the dynamic modeling layer, the harmonic reducer step loading experiment establishes the three-dimensional mapping of joint stiffness and load, speed, and the stiffness compensation coefficient is improved according to the speed increment gradient and the pre-tightening force output is dynamically reduced, and the real-time stiffness compensation parameters are refreshed every 100ms to enhance the high-speed anti-deformation capability; The finite element modal analysis accurately locates the resonance frequency band of 80-150Hz, and injects band-stop filtering instructions to block the resonance energy transmission; Lyapunov stability verifies the dynamic optimization of time domain stiffness weight and frequency domain filtering strength, forming a globally stable composite vibration suppression strategy; For trajectory discontinuous vibration, the quintic polynomial re-planning eliminates the acceleration step, and the sine transition segment is inserted to realize the smooth transition of acceleration, and the impact vibration energy is reduced; The base disturbance suppression derives the end position offset through the space vector chain model, and the Jacobian transpose mapping generates joint compensation torque, and the dynamic gain superposition injects the total vibration suppression torque; The load mutation response detects the change rate in real time through the torque sensor, improves the stiffness compensation coefficient according to the mutation strength, and suppresses the low-frequency shaking through frequency domain spread spectrum processing; The thermal deformation correction calculates the link deformation based on the material thermal expansion coefficient, dynamically adjusts the model geometric parameters, and periodically calibrates to ensure long-term high matching degree of the model; In the execution optimization layer, the high-gain feedforward controller directly injects the vibration suppression torque into the joint actuator, covering the full frequency band vibration suppression of 2-150Hz, and the vibration suppression effect is verified through the vibration suppression rate. When the vibration suppression rate is less than 70%, the sampling frequency is increased to 2kHz, the synchronization accuracy is strengthened to 0.5μs, the strategy optimization frequency is increased to 20 times / s, and the joint speed is reduced to 0.8m / s; When the vibration suppression rate is greater than or equal to 90%, the joint torque is reduced by 10%, the power consumption is reduced by 15%, and the vibration suppression rate is maintained at more than 85%. BRIEF DESCRIPTION OF DRAWINGS

[0017] Figure 1 is a flowchart of the active vibration suppression method of the embodiment of the present application; Figure 2 is a flowchart of step S1 of the active vibration suppression method of the embodiment of the present application; Figure 3 is a flowchart of step S2 of the active vibration suppression method of the embodiment of the present application; Figure 4 is a flowchart of step S21 of the active vibration suppression method of the embodiment of the present application; Figure 5 is a flowchart of step S22 of the active vibration suppression method of the embodiment of the present application; Figure 6 is a flowchart of step S23 of the active vibration suppression method of the embodiment of the present application; Figure 7 is a flowchart of step S24 of the active vibration suppression method of the embodiment of the present application; Figure 8 is a flowchart of step S25 of the active vibration suppression method of the embodiment of the present application; Figure 9 This is a flowchart of step S4 of the active vibration suppression method in this application embodiment; Figure 10 This is a diagram illustrating the composition of the active vibration damping system according to an embodiment of this application; Figure 11 This is a flowchart of the multi-sensor fusion module in the active vibration suppression system of this application embodiment; Figure 12 This is a flowchart of the dual-loop control module in the active vibration damping system of this application embodiment; Figure 13 This is a flowchart of the execution module in the active vibration damping system of this application embodiment. Detailed Implementation

[0018] The following is in conjunction with the appendix Figures 1-13 This application will be described in further detail.

[0019] This application discloses an active vibration suppression method for the end effector of a robotic arm based on predictive compensation. For example... Figure 1 As shown, an active vibration suppression method for the end effector of a robotic arm based on predictive compensation includes the following steps: S1. Collaborative acquisition and fusion of vibration data: A high-frequency IMU sensor is installed on the end flange of the robotic arm to collect the three-axis acceleration, angular velocity, and vibration spectrum data of the robotic arm in real time. The sampling frequency must be greater than 1kHz. At the same time, an auxiliary IMU sensor is installed on the base of the robotic arm to monitor the roll angle, pitch angle, and yaw angle of the base attitude. Then, the high-frequency IMU sensor data and the auxiliary IMU sensor data are fused by the Kalman filter algorithm to improve the signal-to-noise ratio and achieve spatiotemporal synchronization. S2. Structural Flexible Vibration Suppression: Based on the vibration data fused in step S1, a pre-compensation torque is generated through the rigid-flexible coupling dynamic model of the robotic arm. The real-time deviation between the measured spectrum of the high-frequency IMU sensor and the predicted spectrum of the model is combined with the Lyapunov stability algorithm to dynamically optimize the gain coefficient in the rigid-flexible coupling dynamic model, correct the output accuracy of the pre-compensation torque in real time, and output the corrected compensation torque. S3. Vibration suppression command execution: The corrected compensation torque output in step S2 is injected into the feedforward channel of the joint motor control circuit to avoid the phase delay of traditional PID control. A high-gain controller with a response bandwidth covering three times the frequency domain of the vibration fundamental frequency (response delay ≤ 0.5ms) outputs the vibration suppression torque to the joint actuator to cancel the residual vibration in the 2-150Hz frequency band at the end of the robotic arm in real time.

[0020] The method fuses the data collected by the dual IMUs and Kalman filter to construct a high-precision vibration monitoring network, accurately captures the three-axis acceleration, angular velocity and 0-500Hz vibration spectrum of the end of the mechanical arm, and eliminates the disturbance interference of the base at the same time; based on the rigid-flexible coupling dynamics model and Lyapunov real-time optimization, the pre-compensation torque is dynamically generated and the output accuracy is corrected, which effectively suppresses the joint flexible deformation and mechanical resonance; the vibration suppression torque is directly injected into the joint actuator through a high-gain feedforward controller, which real-time cancels the residual vibration in the frequency band of 2-150Hz, and significantly improves the stability and processing quality under high-speed running conditions.

[0021] As shown in Figure 2 The dual-IMU data fusion method of step S1 includes the following steps: S11, based on the precise time synchronization protocol, the microsecond-level time alignment of the high-frequency IMU sensor and the auxiliary IMU sensor is realized, the sampling time is marked by a hardware timestamp, the phase offset caused by transmission delay is eliminated, the time domain reference is established for subsequent frequency domain analysis, and the synchronized data stream is transmitted to the data processing unit at a sampling rate of 1kHz, ensuring time domain consistency; S12, based on the synchronized data of step S11, the original data collected by the high-frequency IMU sensor is decomposed by 5 layers using sym8 wavelet basis, the vibration components of different frequency bands are separated, the soft threshold processing is applied to the high-frequency components above 200Hz to eliminate electromagnetic interference and high-frequency noise, the denoised signal retaining the effective vibration characteristics is retained, the PCA decomposition is performed on the denoised data, the first three principal components with energy ratio exceeding 95% are extracted as effective vibration signals, the environmental temperature drift noise is separated, the three-axis acceleration and angular velocity signals are reconstructed based on the principal components, the residual noise is less than 0.05 times the gravity acceleration, and then the pure vibration signal is output; S13, the three-axis acceleration components of the pure vibration signal output by step S12 are squared and summed, and then the square root of the sum is taken to obtain the real-time vibration energy scalar, and then the data weight is dynamically allocated according to the energy threshold, the specific method is as follows: When the vibration energy is greater than 0.1 times the gravity acceleration, the high-frequency IMU sensor data is used as the main output, the high-frequency vibration details at the end are retained, i.e. the high-frequency IMU sensor data accounts for 80% and the auxiliary IMU sensor data accounts for 20%; When the vibration energy is less than or equal to 0.1 times the gravity acceleration, the auxiliary IMU sensor data is switched to compensation output to suppress the base micro-vibration interference, i.e. the high-frequency IMU sensor data accounts for 60% and the auxiliary IMU sensor data accounts for 40%; S14, the dynamic weight allocation result is fused by a Kalman filter to output the final vibration suppression reference signal.

[0022] The fusion method of step S14 is as follows: S141, based on the time correlation of the vibration signal, predict the vibration state at the next time, for example, if there is an upward acceleration trend, predict that the next time will still maintain the inertia of upward movement; S142, as the prediction step increases, reduce the state confidence to simulate the prediction uncertainty, and the attenuation amplitude is calibrated by the sensor static test; S143, compare the predicted state with the weighted fused observation data to calculate the difference value; S144, if the difference value exceeds the sensor calibration error (such as >0.005g), increase the correction strength (up to 90%) to quickly narrow the predicted value and the observed value; if the difference value is small (<0.001g), reduce the correction strength (minimum to 10%) to avoid over-tuning oscillation; S145, superimpose the corrected residual error on the predicted state to output the optimized vibration signal true value; S146, according to the correction effect, improve the state confidence, and provide a more reliable starting point for the next round of prediction.

[0023] Step S1 eliminates the phase offset caused by transmission delay through microsecond-level time synchronization protocol, ensures the time domain consistency of double-IMU data, and establishes a high-precision reference for frequency domain analysis; 5-layer decomposition and soft threshold processing are adopted based on sym8 wavelet basis to effectively suppress high-frequency interference above 200Hz; combined with PCA principal component analysis, the first three principal components are extracted to separate environmental temperature drift noise, and the output residual noise is less than 0.05g; based on the vibration energy threshold dynamic weight distribution mechanism, in high-energy working conditions, the high-frequency IMU sensor data is used to dominate and retain the end vibration details, and in low-energy working conditions, the auxiliary IMU sensor data is used to suppress the base micro-vibration interference; finally, the vibration suppression reference signal is output through Kalman filter fusion, which significantly improves the vibration monitoring precision and working condition adaptability.

[0024] As shown in Figure 3 and Figure 4 , the pre-compensation torque generation of step S2 is step S21, which includes the following steps: S211, based on the vibration data fused in step S1, perform harmonic reducer step loading experiment in constant temperature environment of 25°C, load gradient increases within 0-20kg, step length 2kg, speed range covers 0.1-3m / s, establish three-dimensional mapping relationship of joint stiffness and load, speed; when the joint speed is greater than 1m / s, based on the basic mapping relationship, the following dynamic adjustment is executed synchronously: a, according to the speed increment gradient, increase the stiffness compensation coefficient (the stiffness compensation coefficient increases by 15% for every 0.5m / s increase in speed), enhance the joint anti-deformation ability to suppress high-frequency vibration, and the calculation formula of the stiffness compensation amount brought by speed increment is:

[0025] wherein, is the stiffness compensation amount brought by the velocity increment, v is the real-time joint velocity, is the joint stiffness base value output by the step S211 step loading test; b. According to the inverse function of joint velocity, the pre-tightening force output value is reduced to avoid the risk of resonance caused by over-constraint, and the calculation formula of the pre-tightening force output value is:

[0026] wherein, is the real-time pre-tightening force output value, is the reference pre-tightening force (the calibration value when the low speed is less than or equal to 1 m / s), and 0.2 is the speed influence factor (to suppress the resonance caused by over-constraint).

[0027] c. The pre-tightening force reference value is refreshed every 100 ms, the stiffness compensation coefficient and the pre-tightening force output value are fused to generate a real-time stiffness compensation parameter, which supports high-frequency vibration suppression, and the calculation formula of the real-time stiffness compensation parameter is:

[0028] wherein, is the real-time stiffness compensation parameter.

[0029] S212, based on the real-time stiffness compensation parameter generated in step S21, the resonance frequency band of 80-150 Hz is identified through finite element modal analysis, and the vibration energy concentration area of the mechanical arm is located; based on the diagnosis result of this resonance frequency band, a band-stop filtering instruction is injected into the pre-compensation torque to suppress the transmission of resonance energy to the end of the mechanical arm; S213, dynamically coupling the real-time stiffness compensation parameter of step S21 and the band-stop filtering instruction of step S22 to form a composite vibration suppression strategy of time-domain stiffness enhancement and frequency-domain energy blocking, and verifying the convergence of the composite vibration suppression strategy through Lyapunov stability condition to ensure that the torque instruction input by the gain correction module meets the global asymptotic stability, and then optimizing the composite vibration suppression strategy based on the verification result, the specific method is as follows: When the component greater than 100 Hz accounts for more than 60%, the stiffness compensation weight is strengthened to 1.2-1.5 times; When the energy of 80-150 Hz is greater than 70% of the total energy, the band-stop filtering weight is increased to 1.3 times; S214, the optimized composite vibration suppression strategy is packaged as a pre-compensation torque instruction package and transmitted to the gain correction module through a standard industrial bus protocol.

[0030] Step S21 constructs a high-precision joint stiffness and load, speed three-dimensional mapping through harmonic reducer step loading experiment, improves the stiffness compensation coefficient according to the speed increment gradient, and dynamically reduces the pre-tightening force output. Real-time stiffness compensation parameters are refreshed every 100 ms to significantly enhance the anti-deformation ability in high-speed working conditions; based on finite element modal analysis, the resonance frequency band of 80-150 Hz is accurately positioned, and the resonance energy transmission is blocked by injecting band-stop filtering instructions; the Lyapunov stability is verified to dynamically optimize the time-domain stiffness weight and the frequency-domain filtering strength, forming a globally stable composite vibration suppression strategy; finally, the instructions are packaged and transmitted to the gain correction module through the standard industrial bus protocol to realize high-reliability and low-delay closed-loop control of the pre-compensation torque, solve the multi-frequency vibration coupling problem in high-speed and high-load working conditions, and suppress the residual vibration at the end.

[0031] As shown in Figure 3 and Figure 5 , step S2 further includes step S22 for suppressing trajectory discontinuous vibration, which includes the following steps: S221, a quintic polynomial is used to re-plan the target motion trajectory of the robot arm, and the time scaling algorithm is used to limit the acceleration within 300 m / s³ to eliminate the acceleration step at the trajectory corner; S222, a sinusoidal function transition section is inserted at the trajectory corner to make the acceleration curve smooth and gradually change, thereby directly reducing the impact vibration component in the high-frequency IMU sensor data input into step S1; S223, the optimized trajectory instruction is input into the vibration data cooperative acquisition module of step S1 as the motion reference for high-frequency IMU sensor data acquisition, and the structure flexibility vibration suppression module of step S2 is linked to dynamically generate a pre-compensation torque based on the trajectory characteristics. The specific method is as follows: The stiffness compensation weight is strengthened in the high-speed section to enhance the joint anti-deformation ability; The band-stop filtering instruction strength is increased in the corner section to block the resonance energy transmission of 80-150 Hz.

[0032] Step S22 eliminates the acceleration step at the corner through quintic polynomial trajectory re-planning, inserts a sinusoidal transition section to make the acceleration curve smooth and gradually change, and directly suppresses the impact vibration component greater than 100 Hz in the high-frequency IMU sensor data. The optimized trajectory is input into the vibration acquisition module as the motion reference, and the vibration suppression module is linked to strengthen the stiffness compensation weight in the high-speed section to improve the joint anti-deformation ability, and to increase the band-stop filtering strength in the corner section to block the resonance energy transmission, thereby significantly reducing the impact vibration and resonance risk caused by trajectory discontinuity, and suppressing the residual amplitude at the end.

[0033] As shown in Figure 3 and Figure 6 , step S2 further includes step S23 for suppressing base disturbance transmission, which includes the following steps: S231, measure the roll angle, pitch angle and yaw angle deviation of the base attitude in real time through the auxiliary IMU sensor of step S1; S232, establish a kinematics conversion model of the mechanical arm based on space vector chain multiplication, and map the end position pose offset caused by the base attitude angle deviation of step S231 into joint space compensation torque through Jacobian matrix transpose, so as to eliminate the kinematics transmission error of the mechanical arm end position caused by the base disturbance; S233, superimpose the joint space compensation torque of step S232 to the corrected compensation torque output by step S2 in proportion, wherein the gain coefficient of step S2 is in positive proportion to the attitude angle deviation of step S231; then input the superimposed total compensation torque into the high gain controller of step S3, and real-time offset the end drift caused by the base disturbance through the feedforward channel of step S3, so as to achieve the base disturbance suppression.

[0034] Step S23 captures the roll angle, pitch angle and yaw angle deviation in real time through the high-precision attitude monitoring of the base IMU, derives the end position pose offset based on the space vector chain kinematics model, and generates the joint compensation torque by mapping with Jacobian matrix transpose, so as to accurately eliminate the kinematics transmission error caused by the base disturbance; through the dynamic gain superposition mechanism, the compensation torque is injected into the total vibration suppression torque in proportion, and the high gain feedforward control is combined to real-time offset the end drift, so as to improve the positioning accuracy under the base disturbance condition.

[0035] As shown in Figure 3 and Figure 7 , step S2 further includes step S24 for suppressing load mutation vibration, which includes the following steps: S241, real-time detect the load change rate through the torque sensor installed at the output end of the joint harmonic reducer, when the load change rate is detected to be greater than 50 N·m / ms, based on the three-dimensional mapping relationship between joint stiffness, load and speed established in step S211, the stiffness compensation coefficient of step S211 is improved in proportion to the load mutation intensity, and the stiffness matrix value in the rigid-flexible coupled dynamics model of step S2 is updated synchronously, so that the rigid-flexible coupled dynamics model matches the mutation working condition in real time, and the low-frequency resonance caused by the dynamics mismatch is eliminated; S242, link the composite vibration suppression strategy generation module of step S213, regenerate the pre-compensation torque based on the updated stiffness matrix value of step S241, perform frequency domain spread processing on the pre-compensation torque through the high gain controller of step S3, expand the vibration suppression frequency band to 2-10 Hz, cover the low-frequency swing caused by the load mutation, and then inject the spreaded vibration suppression torque into the joint actuator of step S3 through the feedforward channel of step S3, real-time offset the low-frequency swing caused by the load mutation, and suppress the residual amplitude.

[0036] Step S24 detects the load change rate in real time through a high-precision torque sensor, dynamically improves the stiffness compensation coefficient based on the three-dimensional stiffness mapping relationship, synchronously updates the stiffness matrix of the rigid-flexible coupling dynamics model, and eliminates the 6-10Hz low-frequency resonance caused by the load mutation; the linkage composite vibration suppression strategy module regenerates the pre-compensation torque, generates the vibration suppression torque through the high-gain controller frequency domain spread spectrum, and injects the joint actuator through the feedforward channel in real time to suppress the low-frequency shaking amplitude caused by the load mutation, and significantly improves the end stability in the heavy load grabbing working condition.

[0037] As shown in Figure 3 and Figure 8 , step S2 further includes step S25 kinematics model thermal deformation correction, which includes the following steps: S251, a network of temperature sensors is deployed at the joints and links of the mechanical arm, real-time monitoring of -10℃ to 60℃ temperature change, while based on the material thermal expansion coefficient to calculate the thermal deformation of the mechanical arm link, analyze the influence of thermal deformation on the vibration data acquisition accuracy of step S1, and quantify the end positioning error caused by thermal drift; S252, according to the thermal deformation amount, the link size parameters in the rigid-flexible coupling dynamics model of step S2 are adjusted in proportion, and the geometric parameters in the three-dimensional mapping relationship of joint stiffness and load, speed established in step S211 are synchronously updated, to eliminate the kinematics model mismatch caused by thermal deformation and suppress the model prediction spectrum shift; S253, the core parameters of the rigid-flexible coupling kinematics model database of step S2 are automatically calibrated every 30 minutes, the composite vibration suppression strategy generation module of step S213 is linked, the pre-compensation torque is regenerated based on the updated rigid-flexible coupling kinematics model of step S2, and the end low-frequency drift caused by thermal deformation is compensated in real time through the high-gain controller of step S3.

[0038] Step S25 monitors the temperature change of -10℃ to 60℃ in real time through the temperature sensor network, accurately calculates the thermal deformation of the mechanical arm link based on the material thermal expansion coefficient, quantifies the end positioning error caused by thermal drift, dynamically adjusts the link size parameters of the rigid-flexible coupling dynamics model according to the deformation amount, and synchronously updates the geometric parameters of the joint stiffness and load, speed mapping relationship, to eliminate the model mismatch caused by thermal deformation; through the 30-minute periodic calibration mechanism to refresh the model database core parameters, the composite vibration suppression strategy module regenerates the pre-compensation torque, and the high-gain controller band-pass filters the real-time compensation of the end low-frequency drift, to improve the positioning accuracy of the mechanical arm in the 8-hour continuous operation working condition.

[0039] As shown in Figure 1 and Figure 9 , step S3 is followed by step S4 vibration suppression effect closed-loop verification, which includes the following steps: S41, based on the high-frequency IMU sensor data of step S1 and the pure vibration signal output in step S12, the acceleration amplitude before and after vibration suppression is extracted to calculate the vibration suppression rate, when the vibration suppression rate is less than 70%, step S42 is executed, when the vibration suppression rate is greater than or equal to 90%, step S43 is executed; S42, the sampling frequency of the high-frequency IMU sensor and the auxiliary IMU sensor of step S1 is increased to 2kHz, and the microsecond-level time synchronization accuracy of the high-frequency IMU sensor and the auxiliary IMU sensor is enhanced to 0.5μs; the dynamic adjustment frequency of the stiffness compensation weight and the band-stop filter strength of step S213 is increased to 20 times per second to improve the convergence speed of the composite vibration suppression strategy; the trajectory re-planning module of step S221 is activated, and the joint speed of the robot arm is reduced to the safety threshold of 0.8m / s; S43, the torque output of the joint motor is reduced by 10% through the standard industrial bus protocol of step S214, while maintaining the vibration suppression rate greater than 85%, the total power consumption of the system is reduced by 15%.

[0040] Step S4 quantitatively evaluates the real-time diagnosis of vibration suppression efficiency through vibration suppression rate, when the vibration suppression rate is less than 70%, the sampling frequency is increased to 2kHz, the synchronization accuracy is enhanced to 0.5μs, the high-frequency vibration capture capability is enhanced, the strategy optimization frequency is increased to 20 times per second, the convergence speed of the composite vibration suppression strategy is accelerated, the joint speed is reduced to 0.8m / s, and the impact energy is suppressed; when the vibration suppression rate is greater than or equal to 90%, the joint torque output is reduced by 10% through the standard industrial bus protocol, the system power consumption is reduced by 15%, while maintaining the vibration suppression rate greater than 85%, the dynamic balance of vibration suppression efficiency and energy consumption is realized, and the end positioning accuracy in all working conditions is improved.

[0041] The present application significantly improves the end positioning accuracy and stability of the robot arm in high-speed and high-load working conditions through multi-source perception, dynamic modeling and execution optimization, as follows: In the multi-source perception layer, a high-precision vibration monitoring network is constructed by dual-IMU cooperative acquisition and Kalman filter fusion, the microsecond-level time synchronization protocol eliminates the phase shift caused by transmission delay, combined with sym8 wavelet basis 5-layer decomposition and soft threshold processing to suppress high-frequency interference above 200Hz, PCA principal component analysis separates environmental temperature drift noise, and outputs a pure vibration signal with residual noise less than 0.05g; based on the dynamic weight distribution mechanism of vibration energy threshold, the end high-frequency vibration details are preserved in high-energy working conditions, and the base micro-vibration interference is suppressed in low-energy working conditions, providing a high-reliability data basis for vibration suppression decision.

[0042] In the dynamic modeling layer, the harmonic reducer step loading experiment establishes a three-dimensional mapping of joint stiffness and load, speed, and the stiffness compensation coefficient is improved according to the speed increment gradient and the pre-tightening force output is dynamically reduced, and the real-time stiffness compensation parameters are refreshed every 100ms to enhance the high-speed anti-deformation capability; The finite element modal analysis accurately locates the resonance frequency band of 80-150Hz, and injects band-stop filtering instructions to block the resonance energy transmission; Lyapunov stability verifies the dynamic optimization of time domain stiffness weight and frequency domain filtering strength, forming a globally stable composite vibration suppression strategy; For trajectory discontinuous vibration, the quintic polynomial re-planning eliminates the acceleration step, and the sine transition segment is inserted to realize the smooth transition of acceleration, reducing the impact vibration energy; The base disturbance suppression derives the end position offset through the space vector chain model, and the Jacobian transpose mapping generates joint compensation torque, and the dynamic gain superposition injects the total vibration suppression torque; The load mutation response detects the change rate in real time through the torque sensor, improves the stiffness compensation coefficient according to the mutation strength, and suppresses the low-frequency shaking through frequency domain spread spectrum processing; The thermal deformation correction calculates the link deformation based on the material thermal expansion coefficient, dynamically adjusts the model geometric parameters, and periodically calibrates to ensure long-term high matching degree of the model.

[0043] In the execution optimization layer, the high-gain feedforward controller directly injects the vibration suppression torque into the joint actuator, covering the full frequency band vibration suppression of 2-150Hz, and the vibration suppression effect is closed-loop verified by the vibration suppression rate to diagnose the efficiency in real time. When the vibration suppression rate is less than 70%, the sampling frequency is increased to 2kHz, the synchronization accuracy is strengthened to 0.5μs, the strategy optimization frequency is increased to 20 times per second, and the joint speed is reduced to 0.8m / s; When the vibration suppression rate is greater than or equal to 90%, the joint torque is reduced by 10%, the power consumption is reduced by 15%, and the vibration suppression rate is maintained at more than 85%.

[0044] In summary, the application constructs an active vibration suppression system, solves the real-time vibration suppression problem under the coupling of multiple source disturbances in the prior art, provides core technical support for high-precision operation scenarios such as precision manufacturing and semiconductor assembly, and the specific effects are as follows: 1. Precision improvement: the end positioning accuracy of the robot arm is ±0.03mm, the thermal drift suppression rate is ≥90%, and the positioning drift is <0.1mm for 8 hours of continuous operation; 2. Vibration suppression efficiency: the residual vibration suppression rate is ≥92% in the frequency band of 2-150Hz, among which the resonance energy at 150Hz is reduced by 40dB, and the amplitude of 10Hz low-frequency shaking is reduced from 0.25mm to 0.02mm; 3. Dynamic response: the delay from vibration occurrence to vibration suppression execution is <5ms, and the convergence speed of the composite strategy is improved by 60%; 4. Energy efficiency balance: when the vibration suppression rate is greater than or equal to 90%, the power consumption is reduced by 15%, and the vibration suppression rate is maintained above 85%; 5. Working condition adaptability: covering multiple disturbance scenarios such as high speed, heavy load, trajectory corner, base disturbance, etc., improving the stability of all working conditions. Embodiment

[0045] The embodiment discloses a mechanical arm end active vibration suppression system. Figure 10 As shown in the figure, a mechanical arm end active vibration suppression system comprises a multi-sensor fusion module, a double-loop control module and an execution module.

[0046] As shown in the figure, the multi-sensor fusion module comprises a high-frequency IMU sensor, an auxiliary IMU sensor and a joint strain sheet network, the high-frequency IMU sensor is installed on the mechanical arm end flange; the auxiliary IMU sensor is installed on the mechanical arm base; the joint strain sheet network is installed on the harmonic reducer output end in a circumferential uniform distribution. Figure 11

[0047] The multi-sensor fusion module accurately captures three-axis acceleration, angular velocity and 0-500Hz vibration frequency spectrum data of the mechanical arm end through the high-frequency IMU sensor, and monitors the high-frequency micro-vibration component greater than 100Hz in real time; the auxiliary IMU sensor synchronously monitors the base attitude disturbance data in combination with a temperature drift compensation mechanism, and eliminates the base drift interference; the joint strain sheet network detects ±0.15mm elastic deformation in a circumferential uniform distribution and sets a 50N·m / ms load mutation trigger threshold, and realizes high-sensitivity deformation sensing. The three form a double-band monitoring network through microsecond-level time synchronization based on the PTP protocol: the 0-10Hz low-frequency band is covered by the joint strain sheet network and the auxiliary IMU sensor, and the 80-500Hz high-frequency band is dominated by the high-frequency IMU sensor, which improves the signal-to-noise ratio and reduces residual noise, provides vibration sensing basis of full-band, high-precision and low-delay for vibration suppression control, and significantly enhances the working condition adaptability and vibration suppression decision reliability.

[0048] As shown in the figure, the double-loop control module comprises a prediction unit and a correction unit, the prediction unit is used for storing a rigid-flexible coupling dynamics model of the mechanical arm, and integrating a stiffness and load, speed three-dimensional mapping database and a finite element modal analysis engine. Figure 12

[0049] ​​The double-loop control module generates joint stiffness compensation coefficients in real time through a stiffness and load, speed three-dimensional mapping database of the prediction unit, dynamically refreshes parameters every 100 ms, and significantly enhances the joint anti-deformation ability under high-speed working conditions; the finite element modal analysis engine accurately identifies the resonance frequency band of 80-150 Hz, generates band rejection filtering instructions, and the stopband attenuation is ≥40 dB, effectively blocking the resonance energy transmission; the correction unit uses Lyapunov stability algorithm to construct energy function, ensures global asymptotic stability, and dynamically optimizes the weight of composite vibration suppression strategy according to the vibration working condition: when high-frequency vibration dominates (energy ratio > 60% above 100Hz), increase the stiffness compensation weight to 1.5 times, enhance the joint anti-high-frequency impact ability; when the resonance frequency band is concentrated (energy ratio > 70% between 80-150Hz), increase the band rejection filtering strength to 1.3 times, and improve the resonance energy suppression rate.

[0050] As shown in Figure 13 The execution module includes a high-gain controller and a joint execution mechanism, the high-gain controller responds to a bandwidth covering three times the vibration base frequency domain, and injects vibration suppression torque into the joint motor feedforward channel; the joint execution mechanism adopts a harmonic reducer vibration suppression structure, which includes a double-row cross-roller bearing integrated piezoelectric ceramic actuator and a pre-tightening force dynamic regulation mechanism.

[0051] The execution module directly injects vibration suppression torque into the joint motor feedforward channel through the high-gain controller, avoiding the phase delay problem of traditional PID closed loop, and the current loop control accuracy reaches ±0.5%, significantly improving the real-time and accuracy of vibration suppression instructions; the joint execution mechanism adopts an innovative harmonic reducer vibration suppression structure, which directly offsets the elastic deformation amount detected by the joint strain gauge network (response delay <0.2ms) through a double-row cross-roller bearing integrated piezoelectric ceramic actuator, realizes real-time compensation of micron-level deformation, improves the attenuation of 150Hz resonance energy, maintains the reference pre-tightening force under low-speed working conditions (≤1m / s), ensures the smoothness of joint operation, and increases the pre-tightening force output by 40% under high-speed working conditions (>1m / s), enhances the joint anti-deformation ability, and suppresses the inertial impact caused by high-speed start-stop. The two technologies cooperatively cover the residual vibration suppression of 2-150Hz full frequency band, the amplitude is more stable, and through the instantaneous torque injection of the feedforward channel, the response lag problem of traditional vibration suppression systems under high-speed variable load working conditions is solved, and the end positioning accuracy of the robot arm is improved.

[0052] The implementation principle of the end of the mechanical arm active vibration suppression system in this embodiment is as follows: a multi-sensor fusion module realizes full-band vibration sensing, that is, a high-frequency IMU sensor quasi-captures 0-500Hz vibration spectrum at the end, monitors high-frequency micro-vibration components greater than 100Hz in real time, an auxiliary IMU sensor eliminates base drift interference in combination with a temperature drift compensation mechanism, a joint strain sheet network detects elastic deformation and sets a load mutation trigger threshold, and the three form a dual-band monitoring network through microsecond-level time synchronization, thereby providing a high-precision and low-delay sensing basis for vibration suppression control; a double-loop control module generates joint stiffness compensation coefficients in real time based on a three-dimensional mapping database of stiffness and load and speed, enhances high-speed deformation resistance, a finite element modal analysis engine accurately identifies a resonance frequency band of 80-150Hz, generates a band rejection filter instruction, a correction unit constructs an energy function through Lyapunov stability algorithm to ensure global asymptotic stability, and dynamically optimizes vibration suppression strategy weights to improve resonance suppression rate; a high-gain controller of an execution module injects vibration suppression torque into a feedforward channel to avoid PID phase delay, and a double-row cross-roller bearing integrated piezoelectric ceramic actuator of a joint execution mechanism directly offsets elastic deformation, a pre-tightening force dynamic control mechanism maintains a reference pre-tightening force in low-speed working conditions and increases output by 40% in high-speed working conditions to enhance deformation resistance, and the two technologies cooperatively cover 2-150Hz full-band vibration suppression. In summary, the application constructs an active vibration suppression system to solve the real-time vibration suppression problem under multi-source disturbance coupling in the prior art.

[0053] The above are preferred embodiments of the application, and do not limit the protection scope of the application, so: any equivalent changes made on the basis of the structure, shape, principle of the application should be covered within the protection scope of the application.

Claims

1. A method for actively damping vibrations at the end of a robot arm based on predictive compensation, characterized in that: The method comprises the following steps: S1, vibration data cooperative acquisition and fusion: a high-frequency IMU sensor is installed on the flange at the end of the mechanical arm to collect real-time three-axis acceleration, angular velocity and vibration frequency spectrum data of the mechanical arm; at the same time, an auxiliary IMU sensor is installed on the base of the mechanical arm to synchronously monitor the roll angle, pitch angle and yaw angle of the base attitude; then, the high-frequency IMU sensor data and the auxiliary IMU sensor data are fused through a Kalman filtering algorithm to improve the signal-to-noise ratio and realize time and space synchronization; S2, structure flexible vibration suppression: based on the vibration data fused in step S1, a pre-compensation torque is generated through a rigid-flexible coupled dynamics model of the mechanical arm, and real-time deviations of the measured frequency spectrum of the high-frequency IMU sensor and the model predicted frequency spectrum are combined to dynamically optimize the gain coefficient in the rigid-flexible coupled dynamics model through a Lyapunov stability algorithm, so as to real-time correct the output accuracy of the pre-compensation torque and output the corrected compensation torque; S3, vibration suppression instruction execution: the corrected compensation torque output in step S2 is injected into the feedforward channel of the joint motor control loop, a high-gain controller covering the three times frequency domain of the vibration base frequency is output through a response bandwidth, and a vibration suppression torque is output to the joint actuator to real-time offset the residual vibration at the end of the mechanical arm.

2. The method of claim 1, wherein: The double-IMU data fusion method of step S1 comprises the following steps: S11, based on the precise time synchronization protocol, the high-frequency IMU sensor and the auxiliary IMU sensor are time-aligned at the microsecond level to establish a time domain basis for subsequent frequency domain processing; S12, using the synchronized data in step S11, wavelet threshold denoising processing is performed on the original data collected by the high-frequency IMU sensor to eliminate high-frequency interference components above 200 Hz, and a principal component analysis algorithm is used to separate effective vibration signals and environmental temperature drift noise to output pure vibration signals; S13, based on the pure vibration signals output in step S12, real-time vibration energy values are calculated, and data weights are dynamically allocated according to energy thresholds, and the specific method is as follows: When the vibration energy is greater than 0.1 times the gravity acceleration, the high-frequency IMU sensor data is used as the main output; When the vibration energy is less than or equal to 0.1 times the gravity acceleration, the auxiliary IMU sensor data is switched to for compensation output.

3. The method of claim 1, wherein: The pre-compensation torque generation of step S2 is step S21, which comprises the following steps: S211, based on the vibration data fused in step S1, a harmonic reducer step loading experiment is performed in a constant temperature environment of 25 DEG C to establish a three-dimensional mapping relationship between joint stiffness and load and speed; when the joint speed is greater than 1 m / s, the following dynamic adjustments are synchronously performed based on the basic mapping relationship: According to the speed increment gradient, the stiffness compensation coefficient is improved to enhance the joint anti-deformation ability to suppress high-frequency vibration; According to the inverse function of the joint speed, the pre-tightening force output value is reduced to avoid the risk of resonance caused by over-constraint; The pre-tightening force reference value is refreshed every 100 ms, the stiffness compensation coefficient and the pre-tightening force output value are fused to generate real-time stiffness compensation parameters; S212, based on the real-time stiffness compensation parameter generated in step S21, identify the resonance frequency band of 80 to 150 Hz through finite element modal analysis, locate the mechanical arm vibration energy concentration area; based on the diagnosis result of this resonance frequency band, inject band-stop filter instructions into the pre-compensation torque to suppress the transmission of resonance energy to the end of the mechanical arm; S213, dynamically couple the real-time stiffness compensation parameter of step S21 and the band-stop filter instruction of step S22 to form a composite vibration suppression strategy of time-domain stiffness enhancement and frequency-domain energy blocking, verify the convergence of the composite vibration suppression strategy through Lyapunov stability condition, ensure that the torque instruction input by the gain correction module meets global asymptotic stability, and then optimize the composite vibration suppression strategy based on the verification result, the specific method is as follows: When high-frequency vibration dominates, the stiffness compensation weight is strengthened; When the resonance frequency band is concentrated, the band-stop filter weight is increased; S214, encapsulate the optimized composite vibration suppression strategy as a pre-compensation torque instruction package, and transmit it to the gain correction module through a standard industrial bus protocol.

4. The method of claim 1, wherein: The step S2 further includes a step S22 for suppressing trajectory discontinuous vibration, which includes the following steps: S221, re-plan the target motion trajectory of the mechanical arm using a quintic polynomial, limit the acceleration to within 300m / s³ through a time scaling algorithm, and eliminate the acceleration step at the trajectory corner; S222, insert a sinusoidal function transition section at the trajectory corner to make the acceleration curve smooth and gradually change, thereby directly reducing the impact vibration component in the high-frequency IMU sensor data input in step S1; S223, input the optimized trajectory instruction into the vibration data cooperative acquisition module of step S1 as the motion reference for high-frequency IMU sensor data acquisition, and simultaneously link the structure flexibility vibration suppression module of step S2 to dynamically generate a pre-compensation torque based on the trajectory characteristics, the specific method is as follows: Strengthen the stiffness compensation weight in the high-speed section to enhance the joint deformation resistance; Increase the strength of the band-stop filter instruction in the corner section to block the transmission of resonance energy.

5. The method of claim 1, wherein: The step S2 further includes a step S23 for suppressing base disturbance transmission, which includes the following steps: S231, measure the roll angle, pitch angle and yaw angle deviation of the base attitude in real time through the auxiliary IMU sensor of step S1; S232, establish a kinematics conversion model of the mechanical arm based on space vector chain multiplication, map the end pose offset caused by the base attitude angle deviation of step S231 into a joint space compensation torque through the transpose of the Jacobian matrix, and eliminate the kinematics transmission error of the mechanical arm end pose caused by the base disturbance; S233, superimpose the joint space compensation torque of step S232 to the corrected compensation torque output by step S2 in proportion, wherein the gain coefficient of step S2 is in positive proportion to the attitude angle deviation of step S231; then input the superimposed total compensation torque into the high-gain controller of step S3, and through the feedforward channel of step S3, real-time offset the end drift caused by the base disturbance to achieve base disturbance suppression.

6. The method of claim 1, wherein: The step S2 further includes a step S24 for suppressing load mutation vibration, which includes the following steps: S241, real-time detect the load change rate through the torque sensor installed on the output end of the joint harmonic reducer, when the load change rate is detected to exceed 50 N·m / ms, based on the three-dimensional mapping relationship between joint stiffness, load and speed established in step S211, the stiffness compensation coefficient of step S211 is proportionally increased according to the load mutation intensity, and the stiffness matrix value in the rigid-flexible coupling dynamics model of step S2 is updated synchronously, so that the rigid-flexible coupling dynamics model matches the sudden working condition in real time, and the low-frequency resonance caused by dynamics mismatch is eliminated; S242, link step S213 composite vibration suppression strategy generation module, based on the stiffness matrix value updated in step S241, the pre-compensation torque is regenerated, the high gain controller of step S3 is used to perform frequency domain spread spectrum processing on the pre-compensation torque, cover the low frequency shaking frequency band, then through the feedforward channel of step S3, the spread spectrum vibration suppression torque is injected into the joint actuator of step S3, real-time offset the low-frequency shaking caused by load mutation, suppress the residual amplitude.

7. The method of claim 1, wherein: The step S2 further comprises a step S25 kinematic model thermal deformation correction, which comprises the following steps: S251, deploy a temperature sensor network on the joints and links of the mechanical arm, real-time monitor the temperature change of-10℃ to 60℃, at the same time, based on the thermal expansion coefficient of the material, calculate the thermal deformation amount of the links of the mechanical arm, analyze the influence of thermal deformation on the vibration data acquisition accuracy of step S1, and quantify the end positioning error caused by thermal drift; S252, adjust the link size parameters in the rigid-flexlexible coupling dynamics model of step S2 according to the thermal deformation amount, and update the geometric parameters in the three-dimensional mapping relationship between joint stiffness, load and speed established in step S211, so as to eliminate the kinematic model mismatch caused by thermal deformation and suppress the model prediction spectrum shift; S253, automatically calibrate the core parameters of the rigid-flexible coupling kinematic model database of step S2 every 30 minutes, link the composite vibration suppression strategy generation module of step S213, regenerate the pre-compensation torque based on the updated rigid-flexible coupling kinematic model of step S2, and real-time offset the end low-frequency drift caused by thermal deformation through the high gain controller of step S3.

8. The method of claim 1, wherein: The step S3 is followed by step S4 vibration suppression effect closed loop verification, which comprises the following steps: S41, based on the high-frequency IMU sensor data of step S1 and the pure vibration signal output by step S12, extract the acceleration amplitude before and after vibration suppression to calculate the vibration suppression rate, when the vibration suppression rate is less than 70%, execute step S42, when the vibration suppression rate is greater than or equal to 90%, execute step S43; S42, improve the sampling frequency of the high-frequency IMU sensor and the auxiliary IMU sensor of step S1, and strengthen the microsecond-level time synchronization accuracy of the high-frequency IMU sensor and the auxiliary IMU sensor; increase the dynamic adjustment frequency of the stiffness compensation weight and the band rejection filter intensity of step S213 to improve the convergence speed of the composite vibration suppression strategy; activate the trajectory re-planning module of step S221 to reduce the joint speed of the mechanical arm to a safety threshold. S43, reduce the torque output of the joint motor through the standard industrial bus protocol of step S214, reduce the total power consumption of the system while maintaining the vibration suppression rate greater than 85%.

9. A mechanical arm end active vibration suppression system for implementing the active vibration suppression method of any one of claims 1 to 8, characterized by: Comprise: A multi-sensor fusion module comprising a high-frequency IMU sensor, an auxiliary IMU sensor and a joint strain network, the high-frequency IMU sensor is installed on the end flange of the mechanical arm for real-time acquisition of three-axis acceleration, angular velocity and vibration frequency spectrum data; the auxiliary IMU sensor is installed on the base of the mechanical arm for synchronous monitoring of roll angle, pitch angle and yaw angle; the joint strain network is used to form a dual-band monitoring network for covering the 0-500Hz vibration frequency domain; A double-loop control module comprising a prediction unit and a correction unit, the prediction unit is used to store the rigid-flexible coupling dynamics model of the mechanical arm and output the pre-compensation torque instruction; the correction unit dynamically optimizes the gain coefficient in the rigid-flexible coupling dynamics model through Lyapunov stability algorithm; An execution module comprising a high-gain controller, the high-gain controller is used to output vibration suppression torque to the joint actuator, and the response bandwidth covers the three times frequency domain of the vibration fundamental frequency.

Images