Method, device and storage medium for inhibiting robot joint module from shaking

Abstract

The application discloses a control method and device for inhibiting robot joint module jitter and a storage medium, relates to the technical field of robot motion control, and comprises the following steps: based on the running state of a robot, a plurality of source motion signals are associated and aligned according to a unified clock reference to obtain an original jitter signal set; the original jitter signal set is subjected to frequency spectrum decoupling processing according to a demarcation frequency to separate high-frequency resonance components, medium-frequency error components and low-frequency disturbance components; the three types of jitter components are compensated in layers according to a full-frequency-band compensation strategy to obtain a layered compensation signal; and the layered compensation signal is superimposed and fused with a servo three-loop control signal to generate a driving instruction for controlling the running of a joint module, so that the full-frequency-band jitter of the joint module is inhibited. Through the full-process jitter inhibition scheme, the application solves the problem of poor joint anti-jitter effect of the robot, improves the full-frequency-band jitter inhibition precision, and enhances the self-adaptive capability of the robot to complex working conditions and the joint running stability.

Description

Technical Field

[0001] This application relates to the field of robot motion control technology, and in particular to a control method, device and storage medium for suppressing the jitter of robot joint modules. Background Technology

[0002] In scenarios involving high-precision control and stable operation of industrial robot joint modules, the precise suppression of joint end-effector jitter, adaptive compensation for dynamic disturbances, and effective control of full-frequency vibration are directly related to the repeatability, work efficiency, operational stability, and service life of industrial robots.

[0003] In related technologies, a single control strategy is used to suppress vibration and control motion of robot joint modules by adopting traditional proportional-integral-derivative control, simple low-pass or band-stop filtering, and unified parameter tuning. This approach relies on traditional proportional-integral-derivative control algorithms, fixed filtering methods, and manual parameter tuning, which makes vibration control difficult to adapt to complex operating scenarios, resulting in poor anti-vibration performance of robot joints.

[0004] The above content is only used to help understand the technical solution of this application and does not represent an admission that the above content is prior art. Summary of the Invention

[0005] The main objective of this application is to provide a control method, device, and storage medium for suppressing the vibration of robot joint modules, aiming to solve the technical problem of poor anti-vibration performance of robot joints.

[0006] To achieve the above objectives, this application proposes a control method for suppressing the vibration of a robot joint module, the method comprising: Based on the robot's operating state, the multi-source motion signals are correlated and aligned according to the clock reference to obtain the original jitter signal set; The original jitter signal set is decoupled according to the boundary frequency spectrum to separate the high-frequency resonant component, the intermediate frequency error component and the low-frequency disturbance component. The high-frequency resonant component, the mid-frequency error component, and the low-frequency disturbance component are compensated in layers according to the full-band compensation strategy, and the layered compensation signal is obtained by sorting. The layered compensation signal is superimposed and fused with the servo three-loop control signal to generate drive commands for controlling the operation of the joint module, thereby suppressing the full-frequency jitter of the robot joint module.

[0007] In one embodiment, based on the robot's operating status data and the clock reference, a timing alignment interval is defined to determine the correction reference for the multi-source motion signal; The multi-source motion signal is calibrated for timing deviation based on the calibration benchmark, and the time-calibrated motion signal is obtained by correlation. The time-calibrated motion signal is fused and arranged with the robot's posture operation signal to form a global motion signal group; Effective fluctuation signals with body sway and offset characteristics are selected from the global motion signal set, and the effective fluctuation signals are aligned according to the clock reference to obtain the original jitter signal set.

[0008] In one embodiment, based on the original jitter signal, and combined with the vibration law of robot mechanical transmission and the characteristics of interference in the field environment, a spectrum decoupling partitioning system including the boundary frequency is constructed; The various vibration signals in the original jitter signal set are coupled and correlated and decomposed using the aforementioned spectrum decoupling partitioning system to obtain layered spectrum signals; Based on the defined boundary frequency, the signal level and signal category of the layered spectrum signal are identified, and individual signals with different vibration properties are obtained. Individual signals of the same type and frequency band are uniformly grouped together to obtain the high-frequency resonant component, the mid-frequency error component, and the low-frequency disturbance component.

[0009] In one embodiment, the global vibration distribution features in the original jitter signal are extracted to distinguish the range of interference features corresponding to the body's own vibration and external environmental disturbances; Based on the range of the interference characteristics, and combined with the robot's full-scene operation data, the inherent vibration evolution law of the mechanical rotation result of the body is sorted out, and the basic frequency range of the body vibration is determined. Using the aforementioned basic frequency range as a benchmark, and combining it with the actual working conditions on site, we analyze the vibration characteristics caused by various external disturbances and define the vibration frequency fluctuation range corresponding to various environmental disturbances. By integrating the basic frequency range and the vibration frequency floating range, and uniformly setting hierarchical frequency boundary points, a spectrum decoupling partitioning system including the boundary frequencies is obtained.

[0010] In one embodiment, the hierarchical compensation execution criteria and compensation control intensity corresponding to the high-frequency resonant component, the mid-frequency error component, and the low-frequency disturbance component are matched according to the full-band compensation strategy. Based on the compensation adjustment intensity, appropriate vibration cancellation adjustment is applied to the high-frequency resonant component, the mid-frequency error component, and the low-frequency disturbance component respectively to obtain the compensation component; According to the hierarchical compensation execution criteria and the compensation components, the high-frequency resonance component, the mid-frequency error component, and the low-frequency disturbance component are processed by deviation offsetting to obtain the hierarchical compensation signal.

[0011] In one embodiment, according to the hierarchical compensation execution criteria, the compensation component corresponding to the high-frequency resonant component is matched, and the high-frequency resonant component is subjected to structural adaptation deviation offsetting to obtain the high-frequency corrected vibration component. According to the hierarchical compensation execution criteria, the compensation component corresponding to the intermediate frequency error component is matched, and the intermediate frequency error component is subjected to operational state linkage deviation offsetting to obtain the intermediate frequency corrected vibration component. Based on the shaping state of the high-frequency corrected vibration component and the mid-frequency corrected vibration component, the compensation component corresponding to the low-frequency disturbance component is matched, and the low-frequency disturbance component is subjected to scene adaptive deviation offsetting to obtain the low-frequency corrected vibration component. The high-frequency corrected vibration component, the mid-frequency corrected vibration component, and the low-frequency corrected vibration component are systematically fused and aggregated according to the spectral arrangement rules to obtain the layered compensation signal.

[0012] In one embodiment, based on the servo three-loop control signal during the operation of the robot joint module, the hierarchical compensation signal is time-synchronized and calibrated for working condition dimension adaptation to obtain the calibrated hierarchical compensation signal. The calibrated hierarchical compensation signal is embedded into the servo three-loop control signal and superimposed to form a control signal; The control signal is logically normalized and transformed according to the robot's operating logic to obtain the drive command corresponding to the joint module; The drive commands control the joint module to complete corresponding actions, and the compensation signals within the drive commands suppress various types of body vibration.

[0013] In one embodiment, during the operation of the robot joint module, global vibration state feedback data generated during the operation of the joint module is collected. The full-range vibration state feedback data is compared with the preset stable operation standard data in all aspects to determine the actual implementation difference of the current full-band jitter suppression. Based on the actual implementation differences, there are control defects and adaptation loopholes due to the mismatch between the positioning-level compensation method and the on-site operating conditions. Based on the aforementioned control defects and adaptation loopholes, the full-band compensation strategy is adjusted, and the hierarchical compensation execution criteria and compensation control intensity corresponding to the vibration components are fine-tuned according to the strategy adjustment range, so as to optimize the vibration reduction effect of the next jitter suppression.

[0014] Furthermore, to achieve the above objectives, this application also proposes a control device for suppressing robot joint module jitter. The control device for suppressing robot joint module jitter includes: a memory, a processor, and a computer program stored in the memory and executable on the processor. The computer program is configured to implement the steps of the control method for suppressing robot joint module jitter as described above.

[0015] In addition, to achieve the above objectives, this application also proposes a storage medium, which is a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the steps of the control method for suppressing robot joint module jitter as described above.

[0016] This application provides a control method for suppressing the jitter of robot joint modules. The method includes: based on the robot's operating state, aligning and associating multi-source motion signals (joint motor current, encoder position, joint output acceleration, and joint torque) according to a unified clock reference to obtain an original jitter signal set; then, performing spectral decoupling processing on the original jitter signal set according to a preset boundary frequency to separate high-frequency resonant components, mid-frequency error components, and low-frequency disturbance components; next, performing layered compensation on each frequency band jitter component according to a full-band compensation strategy and organizing it to obtain a layered compensation signal; finally, superimposing and fusing the layered compensation signal with the servo three-loop control signal to generate drive commands for controlling the joint module's operation. This technical solution solves the problems of traditional proportional-integral-microcontroller (PIMS) methods in existing technologies. The proposed solution addresses several technical issues: poor adaptability to nonlinear disturbances and load changes in PID (Proportional-Integral-Derivative) control; phase lag introduced by simple low-pass or band-stop filtering, reducing system response speed; inability to differentiate jitter sources and address jitter in different frequency bands; reliance on manual parameter tuning leading to readjustment after changes in operating conditions; and insufficient robustness and versatility. This solution improves the full-band jitter suppression effect of robot joint modules, ensuring system dynamic response performance, enhancing adaptability under different operating conditions, significantly improving joint motion smoothness and repeatability accuracy without modifying the mechanical structure, and enhancing the versatility and deployment convenience of the solution for various reducer joint modules.

[0017] In summary, this application solves the technical problem of poor anti-shake performance of robot joints by adopting a full-process jitter suppression scheme based on multi-source motion signal correlation alignment, spectrum decoupling and hierarchical compensation, and servo three-ring signal fusion. It improves the full-band jitter suppression accuracy and enhances the system's adaptability and operational stability. Attached Figure Description

[0018] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0019] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a flowchart illustrating the first embodiment of the control method for suppressing robot joint module jitter according to this application; Figure 2 This is a schematic diagram of the modules in this application; Figure 3 This is a flowchart illustrating the fourth embodiment of the control method for suppressing robot joint module jitter in this application; Figure 4 This is a flowchart illustrating the sixth embodiment of the control method for suppressing robot joint module jitter in this application; Figure 5 This is a flowchart illustrating the eighth embodiment of the control method for suppressing robot joint module jitter in this application; Figure 6 This is a schematic diagram of the control device for suppressing the vibration of the robot joint module in this application.

[0021] The purpose, features, and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0022] It should be understood that the specific embodiments described herein are merely illustrative of the technical solutions of this application and are not intended to limit this application.

[0023] In related technologies, a single control strategy is used to suppress vibration and control motion of robot joint modules by adopting traditional proportional-integral-derivative control, simple low-pass or band-stop filtering, and unified parameter tuning. This approach relies on traditional proportional-integral-derivative control algorithms, fixed filtering methods, and manual parameter tuning, which makes vibration control difficult to adapt to complex operating scenarios, resulting in poor anti-vibration performance of robot joints.

[0024] This application provides a solution: First, based on the robot's operating state, multi-source motion signals are correlated and aligned according to a clock reference to obtain an original jitter signal set. Then, the original jitter signal set is decoupled according to the boundary frequency spectrum to separate high-frequency resonant components, mid-frequency error components, and low-frequency disturbance components. Next, the high-frequency resonant components, mid-frequency error components, and low-frequency disturbance components are compensated in layers according to a full-band compensation strategy to obtain a layered compensation signal. Finally, the layered compensation signal is superimposed and fused with the servo three-loop control signal to generate drive commands for controlling the operation of the joint module, thereby suppressing the full-band jitter of the robot joint module.

[0025] It should be noted that the executing entity in this embodiment can be a computing service device with data processing, network communication, and program execution functions, such as a tablet computer, personal computer, or mobile phone, or an electronic device capable of performing the above functions, a control device for suppressing robot joint module vibration, etc. The following description uses a control device for suppressing robot joint module vibration as an example to illustrate this embodiment and the subsequent embodiments.

[0026] To better understand the technical solution of this application, a detailed description will be provided below in conjunction with the accompanying drawings and specific implementation methods.

[0027] This application provides a control method for suppressing the vibration of a robot joint module, referring to... Figure 1 , Figure 1 This is a flowchart illustrating the first embodiment of the control method for suppressing robot joint module vibration according to this application.

[0028] In this embodiment, the control method for suppressing robot joint module jitter includes steps S10~S40: Step S10: Based on the robot's operating state, the multi-source motion signals are correlated and aligned according to the clock reference to obtain the original jitter signal set.

[0029] The robot's operating state refers to the complex dynamic conditions under which joint modules switch from static to dynamic states, high-speed start-stop, variable load, trajectory switching, and reducer return backlash and flexibility. The clock reference is a high-precision synchronous time base signal used to unify the sampling timing of multiple sensors and eliminate time deviations. Multi-source motion signals refer to four types of real-time feedback signals: joint motor current, encoder position, joint output acceleration, and joint torque. The raw jitter signal set refers to a set of digital signals that are time-consistent, dimensionally complete, and free of phase deviation after timing alignment, timestamping, and analog-to-digital conversion.

[0030] In this embodiment, the multi-source motion signal acquisition and correlation alignment can be triggered in three ways. First, state transition triggering: when the robot joint module switches from a stationary state to motion start-up, high-speed start-stop, or trajectory switching, the system detects the state transition event, immediately synchronizes multi-sensor acquisition, and performs clock reference alignment to form an initial raw jitter signal set. Second, load mutation triggering: when the joint torque and motor current change more than the load mutation threshold within a preset period, it is determined to be a variable load condition, triggering a new round of multi-source signal re-acquisition and timing re-alignment to update the raw jitter signal set. Third, periodic refresh triggering: the system continuously performs clock reference synchronization and multi-source signal correlation alignment according to a preset sampling period, updating the raw jitter signal set frame by frame, covering the dynamic disturbance process under the reducer's return clearance and flexible action.

[0031] Once the jitter suppression control device receives the operating status trigger signal, it begins to perform synchronous acquisition, timing calibration, and correlation alignment of multi-source motion signals.

[0032] For example, there are three methods for synchronous acquisition and timing alignment of multi-source motion signals. The first method is hardware-based hard synchronization phase-locked loop acquisition, which sends synchronous sampling pulses to current sensors, encoders, accelerometers, and torque sensors through a unified clock reference bus. Each sensor strictly acquires the original analog signals in parallel according to the pulse edges. After amplification, filtering, and analog-to-digital conversion, the digital signals are output and then uniformly stamped with high-precision timestamps to directly generate a set of original jitter signals with strict timing alignment. This method has extremely small timing errors and high synchronization accuracy, and can completely eliminate sampling delay deviations between multiple sensors, making it suitable for high-speed and high-frequency jitter suppression scenarios.

[0033] The second method, event-driven soft-synchronous fusion acquisition, uses encoder position zero-crossing, acceleration peak, and torque mutation point as synchronization event markers. After acquiring signals from each sensor separately, cross-sensor timing calibration and signal correlation alignment are performed using these event markers. This dynamically compensates for the inherent delay differences among the sensors, generating a timing-aligned raw jitter signal set. This method does not rely on a hardware synchronization bus, has strong adaptability, and can flexibly adapt to sensor combinations of different brands and interface types, making it suitable for scenarios with compatibility across multiple joint module models.

[0034] The third type of timing deviation adaptive compensation acquisition first asynchronously acquires multi-source motion signals at a fixed sampling frequency, calculates the dynamic timing deviation between each sensor in real time, establishes a timing deviation compensation model, corrects the timestamps of each signal online and completes correlation alignment, and outputs a set of original jitter signals with consistent timing. This method requires no hardware modification, has strong adaptive capabilities, and can automatically offset timing drift caused by changes in operating conditions, making it suitable for non-stationary operating conditions such as variable loads and flexible joints.

[0035] After the multi-source motion signal association and alignment are completed, a set of original jitter signals with consistent timing and complete dimensions is output for subsequent spectrum decoupling processing.

[0036] It should be noted that when a transition in the joint module's operating state or a sudden change in load is detected, the system automatically increases the sampling rate from the base 5kHz to 10kHz and maintains it for 3 control cycles. Simultaneously, the sampling rate is dynamically adjusted based on the real-time vibration amplitude: when the vibration amplitude exceeds 120% of the preset threshold, the sampling rate gradually increases to a maximum of 20kHz. When the vibration amplitude falls below 50% of the threshold, the sampling rate gradually decreases back to the base 5kHz. All sampling processes maintain strict synchronization across multiple sensors via a hardware clock bus, thereby ensuring the accuracy of high-frequency jitter signal acquisition while reducing system computing power and bandwidth consumption.

[0037] Step S20: Decouple the original jitter signal set according to the boundary frequency spectrum to separate the high-frequency resonant component, the intermediate frequency error component and the low-frequency disturbance component.

[0038] The boundary frequency refers to the preset frequency threshold used to divide the high-frequency, mid-frequency, and low-frequency jitter bands, including the high-frequency boundary frequency and the low-frequency boundary frequency. Spectrum decoupling processing refers to the process of independently splitting the original jitter signal set according to preset frequency bands and eliminating inter-band coupling interference based on frequency domain filtering and signal decomposition algorithms. The high-frequency resonant component refers to the high-frequency vibration signal caused by the mechanical resonance of the reducer and connecting rod, which is higher than the high-frequency boundary frequency. The mid-frequency error component refers to the mid-frequency fluctuation signal caused by servo tracking error and positioning overshoot, which is between the low-frequency and high-frequency boundary frequencies. The low-frequency disturbance component refers to the low-frequency drift signal caused by load changes, friction, and external force disturbances, which is lower than the low-frequency boundary frequency.

[0039] In this embodiment, frequency decoupling processing can be achieved in three ways. First, fixed-frequency hierarchical filtering decoupling is used. A fixed high-frequency boundary frequency of 200 Hz and a low-frequency boundary frequency of 20 Hz are preset. The original jitter signal set is filtered step-by-step through a high-pass filter, a band-pass filter, and a low-pass filter, directly separating the high-frequency resonant component, the intermediate-frequency error component, and the low-frequency disturbance component. This method has a simple algorithm, low computational load, and high real-time performance, making it suitable for jitter suppression scenarios in conventional joint modules with stable operating conditions and fixed resonant frequencies.

[0040] Secondly, the system employs adaptive frequency decoupling based on operating conditions. It analyzes the power spectral density of the original jitter signal set in real time, identifies the main peak of mechanical resonance and the characteristic peak of disturbance under the current operating condition, and dynamically and adaptively adjusts the high-frequency and low-frequency boundary frequencies. Then, it uses an adaptive filter bank to perform frequency band decoupling, separating three types of jitter components that precisely match the current operating conditions. This method can dynamically adapt to frequency band divisions as operating conditions change, offers high decoupling accuracy, and is suitable for complex dynamic operating conditions such as high-speed start-stop, variable load, and trajectory switching.

[0041] Third, mode decomposition combined with spectral decoupling involves performing empirical mode decomposition on the original jitter signal set to extract different intrinsic mode function components. Then, combined with a preset boundary frequency, mode and spectrum joint screening is performed to eliminate mode coupling interference, accurately separating the high-frequency resonant components, mid-frequency error components, and low-frequency disturbance components without cross-interference. This method can completely eliminate inter-band mode coupling and achieves high decoupling purity, making it suitable for flexible joint modules where reducer backlash and flexibility are significant and multi-mode coupling is severe.

[0042] After the spectrum decoupling process is completed, the high-frequency resonant component, mid-frequency error component, and low-frequency disturbance component, which are independent and free from coupling interference, are output for subsequent hierarchical compensation processing.

[0043] Step S30: Based on the full-band compensation strategy, the high-frequency resonant component, the mid-frequency error component, and the low-frequency disturbance component are compensated in layers to obtain the layered compensation signal.

[0044] The full-band compensation strategy refers to a hierarchical control strategy that matches dedicated compensation algorithms to the three types of jitter components—high frequency, mid frequency, and low frequency—to achieve blind-spot-free suppression across the entire frequency band. Hierarchical compensation refers to the process of using differentiated compensation mechanisms for different frequency band jitter components and independently generating compensation amounts for each band. The hierarchical compensation signal is the synthesized compensation signal obtained by organizing and normalizing the high-frequency, mid-frequency, and low-frequency compensation amounts according to the control timing sequence.

[0045] In this embodiment, full-band hierarchical compensation can be achieved through three methods. First, hierarchical compensation using modal active damping, model reference adaptive feedforward, and extended state observer applies modal active damping compensation to the high-frequency resonant components, directly suppressing mechanical resonance without phase lag. Model reference adaptive feedforward compensation is performed on the mid-frequency error components, dynamically correcting the feedforward gain and eliminating tracking overshoot. Extended state observer perturbation compensation is applied to the low-frequency perturbation components, estimating and canceling the total perturbation in real time. The three types of compensation quantities are then normalized to obtain the hierarchical compensation signal. This method is highly targeted, has no phase lag, and possesses strong adaptive capabilities, making it the core compensation scheme for full-band jitter suppression.

[0046] Secondly, a layered compensation approach is employed, combining notch resonant suppression, sliding mode error compensation, and disturbance observation feedforward. For high-frequency resonant components, adaptive notch filtering suppresses resonance peaks. For mid-frequency error components, sliding mode control facilitates rapid convergence and error tracking. For low-frequency disturbance components, a disturbance observer provides real-time estimation and feedforward compensation. These three types of compensation are then fused to generate a layered compensation signal. This approach exhibits strong resistance to parameter perturbations and high robustness, making it suitable for industrial robotic arm scenarios with time-varying parameters and drastic load fluctuations.

[0047] Thirdly, a layered compensation approach is employed, combining frequency-domain damping shaping, adaptive robust compensation, and low-frequency friction compensation. This approach reshapes the system's frequency response by applying frequency-domain damping to the high-frequency resonant components. Adaptive robust compensation is performed on the mid-frequency error components to suppress uncertainty errors. Nonlinear friction compensation is applied to the low-frequency disturbance components to counteract low-speed jitter caused by friction. The resulting layered compensation signal is then integrated. This method offers excellent dynamic performance and good low-speed stability, making it suitable for high-precision, stable control scenarios such as collaborative robots and flexible joint actuators.

[0048] After completing the full-band hierarchical compensation, the output hierarchical compensation signal with timing matching, amplitude normalization, and conflict-free superposition is used for subsequent servo control fusion.

[0049] Step S40: The layered compensation signal is superimposed and fused with the servo three-ring control signal to generate drive commands for controlling the operation of the joint module, so as to suppress the full-band jitter of the robot joint module.

[0050] Servo three-loop control signals refer to the basic control signals output from the position loop, speed loop, and current loop of the servo driver. Superposition and fusion refers to the synthesis process of aligning the layered compensation signals with the servo three-loop control signals in terms of timing, amplitude matching, weighted superposition, and amplitude limiting protection. Drive commands refer to the final pulse width modulation (PWM) control commands output to the servo driver after fusion, used to control the operation of the joint motor. Full-frequency jitter refers to vibration and jitter across the entire frequency range, covering high-frequency mechanical resonance, mid-frequency tracking error, and low-frequency load disturbances.

[0051] In this embodiment, the superposition and fusion of the hierarchical compensation signal and the servo three-loop control signal can be achieved in three ways. First, direct weighted linear superposition and fusion: the hierarchical compensation signal is directly superimposed onto the speed loop and current loop outputs of the servo three-loop control signal with fixed weights, and then a drive command is generated after amplitude limiting protection. This method has simple fusion logic, low computational latency, and strong real-time performance, making it suitable for conventional jitter suppression scenarios with high response speed requirements and small compensation amplitude.

[0052] Secondly, error closed-loop feedback correction fusion introduces the hierarchical compensation signal as a feedforward compensation quantity into the servo three-loop control closed loop, providing real-time feedback on the fused control error, dynamically adjusting the compensation weights and correcting the fusion result in the closed loop, and generating high-precision drive commands. This method offers high fusion accuracy and can automatically cancel compensation residuals, making it suitable for collaborative robot scenarios requiring high-precision positioning and low residual vibration.

[0053] Third, model prediction and dynamic fusion are employed. Based on the joint module dynamics model, the jitter trend at the next moment is predicted, and the fusion weights of the hierarchical compensation signal and the servo three-loop control signal are dynamically adjusted to compensate for the jitter trend in advance and generate forward-looking drive commands. This method has predictive suppression capabilities and can suppress jitter recurrence caused by time delay, making it suitable for industrial robotic arm scenarios with high-speed, high-dynamic trajectory motion.

[0054] After the superposition and fusion are completed, the output drive commands are accurate in timing, reasonable in amplitude, without overshoot, and without phase lag, controlling the joint module to operate smoothly and achieving full-band jitter suppression.

[0055] Further, please refer to Figure 2 , Figure 2This is a schematic diagram of the module in this application. The motor q-axis current signal, encoder position / speed signal, joint output acceleration signal, and joint torque sensor signal are acquired in real time at a sampling rate of 10 kHz (10 Kilohertz). After signal acquisition by the multi-source signal acquisition module (including current sensor, encoder, accelerometer, and torque sensor), improved complementary filtering is performed by the jitter separation processing module (embedded processor). The jitter components are separated into layers according to high-frequency components (greater than 200 Hz, corresponding to reducer / connecting rod mechanical resonance), mid-frequency components (20 Hz to 200 Hz, corresponding to servo tracking error and overshoot), and low-frequency components (less than 20 Hz, corresponding to load changes, friction, and external force disturbances). The separation formula is: the total jitter signal (S_total, Signal_total) equals the high-frequency resonance component (S_high, Signal_high) plus the mid-frequency error component (S_mid, Signal_mid) plus the low-frequency disturbance component (S_low, Signal_low). Subsequently, through a hierarchical adaptive compensation module, the high-frequency components are directly suppressed by modal active damping without phase lag; the mid-frequency components are dynamically corrected by model reference adaptive feedforward to eliminate tracking overshoot; and the low-frequency components are estimated and compensated online in real time by an extended state observer (ESO). Then, the control quantity is fused according to the formula that the total output control quantity (U_out, Output) equals the position loop output (U_pos, Position) plus adaptive feedforward compensation (U_feed, Feedforward) plus modal damping compensation (U_damp, Damp) plus extended state observer disturbance compensation (U_eso, ESO). The fused total control quantity is output to the servo drive control module, which outputs a pulse width modulation (PWM) signal to drive the joint motor to run smoothly, ultimately achieving full-band jitter suppression of the robot joint module.

[0056] Second Embodiment This embodiment provides an exemplary scheme for timing alignment of multi-source motion signals and generation of the original jitter signal set. In this example, the timing alignment interval is first defined based on robot running status data and clock reference, and a correction reference is determined. Then, the timing deviation of the multi-source motion signals is calibrated. Next, the robot posture operation completion signal is fused to form a global motion signal group. Finally, effective fluctuation signals are selected and aligned to obtain the original jitter signal set. Step S10 includes steps A11 to A14: Step A11: Based on the robot's operating status data and the clock reference, define the timing alignment interval and determine the correction reference for the multi-source motion signal.

[0057] Step A12: Perform timing deviation calibration on the multi-source motion signal according to the calibration benchmark, and obtain the timing-calibrated motion signal.

[0058] Step A13: The timing-calibrated motion signal is fused and arranged with the robot's posture operation signal to form a global motion signal group.

[0059] Step A14: Select effective fluctuation signals with body swaying and offset characteristics from the global motion signal group, and align the effective fluctuation signals according to the clock reference to obtain the original jitter signal set.

[0060] Operational status data refers to the set of state parameters characterizing the real-time operating condition of a robot joint module, and is the core basis for defining timing alignment intervals. Examples include joint start / stop status, load change amplitude, trajectory switching time, reducer return backlash trigger signal, and flexible deformation. Clock reference refers to a high-precision synchronous time base signal used to unify the sampling timing of multiple sensors and eliminate time deviations; it is the unified time scale for all signal calibration and alignment. Examples include the internal crystal oscillator clock of the servo driver, the external synchronous bus clock, and the global system clock. Timing alignment intervals refer to the time windows defined based on operational status data that require multi-source signal synchronous calibration, covering the complete process of jitter generation and persistence. Examples include start / stop transition intervals, load change intervals, trajectory switching intervals, and flexible vibration intervals. Correction references refer to the reference time points or reference signal sequences selected within the timing alignment interval as the reference for time calibration of all multi-source motion signals. Examples include encoder position zero-crossing points, motor current change points, acceleration peak points, and torque change points.

[0061] Timing-calibrated motion signals refer to the joint motor current, encoder position, joint output acceleration, and joint torque signals after timing deviation calibration to eliminate sampling delays and time drift between multiple sensors. Robot posture completion signals are feedback signals characterizing the robot joint's completion of a preset posture action and entry into a stable or stopped state; they are used to delineate the complete motion cycle and jitter-related intervals. Global motion signal sets refer to a multi-dimensional signal set covering the complete motion cycle, formed by fusing and arranging the timing-calibrated motion signals and posture completion signals in chronological order. Body sway offset characteristics are core fluctuation characteristics that distinguish normal motion signals from jitter signals and are the basis for selecting effective jitter signals. Examples include signal amplitude abrupt change characteristics, frequency abrupt change characteristics, phase offset characteristics, and torque fluctuation characteristics. Effective fluctuation signals refer to signal components with body sway offset characteristics directly related to joint jitter, selected from the global motion signal set, after eliminating steady-state signals and noise interference from normal motion. The original jitter signal set refers to a set of digital signals that are time-consistent, dimensionally complete, and contain only jitter-related information, formed by realigning the effective fluctuation signals according to a clock reference.

[0062] In this example, when defining the timing alignment interval and determining the correction reference based on the robot's operating status data and clock reference, it can be done by triggering state transitions. Alternatively, by monitoring the changing trend of the operating status data in real time, the timing of jitter can be predicted in advance, and the timing alignment interval can be dynamically defined and the correction reference selected, thus completing the preliminary preparation for timing calibration.

[0063] After determining the calibration benchmark, the timing deviation calibration process for multi-source motion signals is initiated. For each multi-source motion signal, the timing deviation between it and the calibration benchmark is calculated sequentially. Based on this deviation, the timestamp of the signal is corrected, and the timing-calibrated motion signal is obtained. Then, the timing-calibrated motion signal and the robot posture operation completion signal are fused and arranged in chronological order to form a global motion signal group covering the complete motion cycle. Finally, effective fluctuation signals with body sway offset characteristics are selected from the global motion signal group, and the effective fluctuation signals are realigned according to the clock benchmark to obtain the original jitter signal set. In this way, through hierarchical timing calibration and signal filtering, the purity and timing consistency of the original jitter signal set are improved, avoiding subsequent spectral decoupling errors caused by signal mixing and timing deviation.

[0064] For example, there are two methods for filtering valid fluctuation signals and generating the original jitter signal set through a global motion signal group. The first method is a frame-by-frame serial filtering based on feature thresholds. According to a preset body sway offset feature threshold, starting from the beginning of the global motion signal group, feature parameters of each dimension of the signal are extracted frame by frame, and the feature parameters of each signal frame are compared with the preset threshold. After each frame's feature comparison is completed, it is simultaneously determined whether the frame belongs to a valid fluctuation signal. If it meets the feature threshold requirement, the frame signal is included in the valid fluctuation signal set, and the time span of valid signals is continuously accumulated. If it does not meet the requirement, the frame signal is skipped, and the judgment and filtering of the next frame continues. After all frames of the global motion signal group have completed frame-by-frame comparison and judgment, all valid fluctuation signals are summarized and re-aligned according to the clock reference to obtain the original jitter signal set. This method uses frame-by-frame feature comparison and threshold judgment calculation logic, and through strict feature filtering, normal motion signals and noise interference are eliminated, ensuring the purity of the original jitter signal set. It is suitable for normal operating conditions with obvious jitter characteristics and little interference.

[0065] The second method is parallel filtering based on feature clustering. This involves reading and analyzing the global feature distribution of the entire motion signal group, extracting the body sway offset feature parameters of all signal frames, and performing clustering based on the distribution characteristics of these parameters. Signal frames with feature parameters within the same preset fluctuation range and sequentially continuous are grouped into the same fluctuation feature range, and non-fluctuation boundary nodes between ranges are marked, generating multiple fluctuation feature ranges that are computationally independent of each other. Then, parallel verification processing is simultaneously initiated for all the divided fluctuation feature ranges. Within each fluctuation feature range, the feature parameters of all signal frames within the range are verified to ensure they all meet the body sway offset feature requirements, and the total duration of signal frames within ranges that meet the full range fluctuation requirements is calculated. After all parallel verification and statistics for all fluctuation feature ranges are completed, all valid fluctuation signal frames that meet the requirements are summarized and realigned according to a clock reference to obtain the original jitter signal set. This method employs a computational logic of global feature clustering and parallel verification across the entire range. By performing global feature distribution clustering in advance, it can lock in potential jitter signal ranges in advance, reducing the number of invalid frame-by-frame comparisons and improving the efficiency of signal filtering. It is suitable for complex operating conditions involving high-speed motion and long-cycle operation.

[0066] Third Embodiment This embodiment provides an exemplary scheme for spectral decoupling and hierarchical component extraction of the original jitter signal set. In this example, a spectral decoupling and partitioning system containing boundary frequencies is first constructed based on the original jitter signal set, combined with the vibration law of robot mechanical transmission and the characteristics of interference in the field environment. Then, the various vibration signals are coupled and decomposed through this system to obtain hierarchical spectral signals. Next, the signal levels and categories are identified according to the boundary frequencies to filter out individual signals with different vibration attributes. Finally, signal individuals of the same type and frequency band are uniformly grouped to obtain three types of jitter components. Step S20 includes steps B11 to B14: Step B11: Based on the original jitter signal, and combined with the vibration law of robot mechanical transmission and the characteristics of interference in the field environment, construct a spectrum decoupling partitioning system including the boundary frequency.

[0067] Step B12: The various vibration signals in the original jitter signal set are coupled and decomposed using the spectrum decoupling partitioning system to obtain layered spectrum signals.

[0068] Step B13: Identify the signal level and signal category of the layered spectrum signal according to the defined boundary frequency, and filter out individual signals with different vibration attributes.

[0069] Step B14: Collect all individual signals of the same type and frequency band to obtain the high-frequency resonant component, the mid-frequency error component, and the low-frequency disturbance component.

[0070] The vibration law of robot mechanical transmission refers to the inherent frequency characteristics and propagation law of the vibration generated by mechanical components such as reducers, links, and motors in robot joint modules during movement. It is the core basis for determining the boundary frequency. For example, the meshing vibration frequency of a harmonic reducer, the cycloidal vibration frequency of a rotary vector (RV) reducer, and the first-order natural resonant frequency of a link. The characteristics of on-site environmental interference refer to the frequency distribution and amplitude characteristics of external interference such as electromagnetic interference, mechanical shock, and ground vibration in the robot's working environment. It is an important reference for optimizing the spectrum decoupling system. For example, power frequency electromagnetic interference frequency, vibration frequency of surrounding equipment, and ground-conducted vibration frequency. The spectrum decoupling system is a signal decomposition system based on the mechanical vibration law and environmental interference characteristics, containing multi-level boundary frequencies and frequency band division rules. It is used to decompose mixed vibration signals into independent frequency band components. For example, a complete system including high-frequency boundary frequencies, low-frequency boundary frequencies, frequency band assignment rules, and coupling interference elimination rules.

[0071] Coupling-correlation decomposition refers to the process of separating and eliminating inter-band cross-interference of vibration signals from different frequency bands that are coupled together in the original jitter signal. It is a key step in obtaining a clean, layered spectral signal. For example, it eliminates the coupling interference of high-frequency resonance on the mid-frequency error signal and eliminates the baseline drift interference of low-frequency disturbances on the high-frequency signal. A layered spectral signal is a collection of spectral signals arranged in layers according to frequency, obtained after coupling-correlation decomposition, containing mixed signal components of different frequency bands. Examples include high-frequency mixed spectral signals, mid-frequency mixed spectral signals, and low-frequency mixed spectral signals.

[0072] Signal levels refer to the three signal levels—high frequency, mid frequency, and low frequency—defined by boundary frequencies, used to distinguish the frequency range of vibration signals. Examples include high-frequency level, mid-frequency level, and low-frequency level. Signal categories refer to the three signal categories—mechanical resonance, tracking error, and load disturbance—defined by the source of vibration, used to distinguish the cause of vibration signals. Examples include mechanical resonance signals, tracking error signals, and load disturbance signals. A signal unit refers to an independent signal unit with a single frequency range and a single vibration source, obtained after identification; it is the basic unit for aggregating the final jitter components. Examples include a 250 Hz reducer meshing resonance signal, a 50 Hz servo tracking error signal, and a 5 Hz load variation disturbance signal.

[0073] In this example, when constructing a spectrum decoupling system based on the original jitter signal set, the vibration characteristics of the robot's mechanical transmission, and the characteristics of environmental interference, a layered filtering method with fixed boundary frequencies can be used. Alternatively, by analyzing the power spectral density of the original jitter signal set in real time, and combining it with a pre-stored database of mechanical transmission vibration characteristics and a database of environmental interference, the boundary frequencies and frequency band division rules can be dynamically adjusted to construct an adaptive spectrum decoupling system, thus completing the preliminary preparations for spectrum decoupling.

[0074] After completing the construction of the spectrum decoupling and partitioning system, the coupling correlation decomposition process is initiated. Through this system, various vibration signals in the original jitter signal set are separated by frequency band and coupling interference is eliminated to obtain layered spectrum signals. Then, the signal level and signal category of the layered spectrum signals are identified according to the boundary frequency, and individual signals with single vibration attributes are selected. Finally, signal individuals of the same type and frequency band are uniformly collected to obtain high-frequency resonant components, mid-frequency error components, and low-frequency disturbance components. In this way, through layered coupling decomposition and precise identification, the purity of spectrum decoupling and the accuracy of component extraction are improved, avoiding poor subsequent compensation effects caused by frequency band cross-interference and signal mixing.

[0075] For example, there are two methods for obtaining the three types of jitter components through layered spectrum signal identification and aggregation. The first method is a frame-by-frame serial identification based on both frequency band and category dimensions. According to the boundary frequency and signal category determination rules in the spectrum decoupling partitioning system, starting from the beginning position of the layered spectrum signal, the frequency features and source features of each signal unit are extracted frame by frame, and the features of each signal unit are compared with the preset rules in turn. After the feature comparison of each frame of signal is completed, the signal level and signal category to which the signal unit belongs are determined simultaneously. If it meets the determination rules of a certain level and category, the signal unit is included in the signal individual set of the corresponding category, and the signal duration of that category is continuously accumulated. If it does not meet any rules, it is determined to be an interference signal and is removed, and the determination and filtering of the next frame continues. After all layered spectrum signals have been compared and determined frame by frame, all signal individuals of the same type and frequency band are summarized to obtain the high-frequency resonant component, the mid-frequency error component, and the low-frequency disturbance component, respectively. This method adopts a two-dimensional frame-by-frame determination and interference removal calculation logic, and ensures the accuracy of component extraction through strict feature matching. It is suitable for conventional working conditions with single vibration sources and few interferences.

[0076] The second method is a full-band parallel discrimination combining modal and spectral analysis. This involves reading and performing global modal analysis on the layered spectral signals, extracting the inherent modal and frequency characteristics of all signal units. Based on the joint distribution of modal and frequency characteristics, clustering is performed, grouping signal units with the same modal characteristics and frequency range into the same signal category cluster. Category boundary nodes between clusters are marked, generating multiple signal category clusters that are computationally independent of each other. Parallel verification is then initiated simultaneously for all the divided signal category clusters. Within each cluster, the characteristics of all signal units are verified to ensure they all meet the criteria for the corresponding category, and the total duration of signal units meeting the overall cluster category requirements is calculated. After all parallel verification and statistics for all signal category clusters are completed, all compliant signals of the same type and frequency band are aggregated, yielding high-frequency resonant components, mid-frequency error components, and low-frequency disturbance components. This method employs a computational logic of joint modal and spectral clustering and parallel verification of the entire cluster. By performing global modal analysis in advance, it can lock signal clusters from different vibration sources, reduce the number of invalid frame-by-frame comparisons, and significantly improve the efficiency of spectral decoupling. It is suitable for working conditions with multimodal coupling and complex interference.

[0077] Fourth embodiment This embodiment provides an exemplary scheme for adaptively constructing a spectrum decoupling partitioning system. In this example, the global vibration distribution characteristics in the original jitter signal are first extracted to distinguish the interference characteristic ranges corresponding to the robot's inherent vibration and external environmental disturbances. Then, combined with the robot's full-scene operation data, the inherent vibration evolution law of the robot's mechanical rotation is analyzed to determine the fundamental frequency range of the body vibration. Next, using the fundamental frequency range as a benchmark, and combined with the actual on-site operating conditions, vibration frequency fluctuation ranges corresponding to various environmental disturbances are delineated. Finally, the two types of ranges are integrated to uniformly set hierarchical frequency boundary points, resulting in a spectrum decoupling partitioning system including the boundary frequencies. Please refer to... Figure 3 , Figure 3 This is a flowchart illustrating the fourth embodiment of the control method for suppressing robot joint module vibration according to this application. Step B11 includes steps C11 to C14: Step C11: Extract the global vibration distribution features from the original jitter signal and distinguish the range of interference features corresponding to the body's own vibration and external environmental disturbances.

[0078] Step C12: Based on the range of the interference characteristics, and combined with the robot's full-scene operation data, sort out the inherent vibration evolution law of the mechanical rotation result of the body, and determine the basic frequency range of the body vibration.

[0079] Step C13: Using the aforementioned base frequency range as a benchmark, and combining it with the actual working conditions on site, analyze the vibration characteristics caused by various external disturbances, and delineate the vibration frequency fluctuation range corresponding to various environmental disturbances.

[0080] Step C14: Integrate the basic frequency range and the vibration frequency floating range, and uniformly set the hierarchical frequency boundary points to obtain a spectrum decoupling partitioning system including the boundary frequencies.

[0081] Global vibration distribution characteristics refer to the set of power spectral density distribution, amplitude distribution, and time-domain fluctuation characteristics of the original jitter signal across the entire frequency range, and are the core basis for distinguishing different vibration sources. Examples include the frequency peak location, peak amplitude, bandwidth, and time-domain fluctuation period of the vibration signal. Inherent vibration of the robot body refers to the inherent vibration generated by the mechanical structure of the robot's joint module during movement, and is the core object of the spectrum decoupling partitioning system. Examples include the cycloidal vibration of a rotary vector reducer (RV), the meshing vibration of a harmonic reducer, and the first-order natural vibration of a connecting rod.

[0082] External environmental disturbances refer to non-organic vibration interference from the external environment in the robot's working environment. These are interference signals that need to be eliminated or processed separately during spectrum decoupling. Examples include vibrations from surrounding equipment, ground-transmitted vibrations, and signal fluctuations caused by electromagnetic interference. The interference characteristic range refers to the distribution range of the external environmental disturbance signal in the frequency and amplitude domains, serving as the boundary between machine vibration and environmental interference. Examples include the 50 Hz range for power frequency electromagnetic interference and the 100 Hz to 150 Hz range for vibrations from surrounding motors. Robot full-scenario operation data refers to the historical vibration data collected under all typical working conditions of the robot, including no-load, full-load, high-speed, low-speed, start-stop, and trajectory switching. This data forms the basis for understanding the inherent vibration evolution patterns. Examples include vibration data from no-load operation, vibration data from full-load handling, and vibration data from high-speed trajectory tracking.

[0083] The inherent vibration evolution law refers to the intrinsic law governing the changes in the frequency and amplitude of robot body vibration with operating speed, load size, and motion posture. It is the core basis for determining the fundamental frequency range of robot body vibration. For example, the meshing frequency of a reducer changes linearly with motor speed, and the resonant frequency of a link decreases slightly with increasing load. The fundamental frequency range of robot body vibration refers to the frequency coverage of the robot body's inherent vibration under all operating conditions, and is the core reference range of the spectrum decoupling system. For example, the fundamental resonant range of 150 Hz to 300 Hz for rotary vector reducer joints, and the fundamental resonant range of 200 Hz to 400 Hz for harmonic reducer joints.

[0084] Actual on-site operating conditions refer to the specific working environment and operating state of the robot, serving as the basis for defining the frequency fluctuation range of environmental disturbances. Examples include assembly line operation conditions, warehouse handling operation conditions, and precision assembly operation conditions. Vibration characteristics refer to the frequency stability, amplitude fluctuation, and duration of different types of external environmental disturbances, serving as the basis for distinguishing different types of environmental disturbances. Examples include the low-frequency, wide-amplitude characteristics of ground vibration, the high-frequency, narrow-band characteristics of electromagnetic interference, and the transient pulse characteristics of equipment impacts.

[0085] The vibration frequency fluctuation range refers to the frequency fluctuation range of various external environmental disturbances under current field conditions. It is the range used to avoid interference in the spectrum decoupling system. For example, the 0 Hz to 15 Hz fluctuation range of ground vibration at the current site, and the 80 Hz to 120 Hz fluctuation range of surrounding equipment vibration. The hierarchical frequency boundary points refer to the specific frequency values ​​used to divide the high-frequency, mid-frequency, and low-frequency jitter bands, and are the core parameters of the spectrum decoupling system. For example, the 20 Hz low-frequency boundary point and the 200 Hz high-frequency boundary point.

[0086] In this example, when constructing the spectrum decoupling partitioning system, it can be done by using a layered filtering method with fixed boundary frequencies. Alternatively, by extracting the global vibration distribution characteristics of the original jitter signal in real time, and combining it with a pre-stored database of robot operation in all scenarios and a database of on-site environmental interference, the layered frequency boundary points can be dynamically adjusted to construct an adaptive spectrum decoupling partitioning system, thus completing the preliminary preparation for spectrum decoupling.

[0087] After extracting the full-domain vibration distribution characteristics and distinguishing the range of interference characteristics, the process of sorting out the inherent vibration evolution law is initiated. Combining the robot's full-scene operation data, the law of the body vibration frequency changing with the working conditions is analyzed, and the basic frequency range of the body vibration is determined. Then, using the basic frequency range as a benchmark, the vibration characteristics of various external interferences are sorted out in combination with the actual on-site working conditions, and the corresponding vibration frequency fluctuation range is delineated. Finally, the basic frequency range and the vibration frequency fluctuation range are integrated, and the hierarchical frequency boundary points are uniformly set to obtain a spectrum decoupling division system including the boundary frequency. In this way, by separating and analyzing the body vibration and environmental interference, the working condition adaptability and anti-interference capability of the spectrum decoupling division system are improved, and the decoupling error of fixed boundary frequency under complex working conditions is avoided.

[0088] For example, there are two methods for setting hierarchical frequency boundary points by integrating the fundamental frequency range and the vibration frequency fluctuation range. The first method is a point-by-point serial setting based on interval boundary avoidance. Starting from the upper and lower boundaries of the fundamental vibration frequency range, the vibration frequency fluctuation range affected by environmental disturbances is checked outwards sequentially. For each environmental disturbance range checked, the corresponding boundary point is adjusted to the gap between the two ranges, ensuring that the boundary point avoids the frequency range of all environmental disturbances. After all environmental disturbance ranges have been checked, the low-frequency boundary point and the high-frequency boundary point are finally determined, forming a spectrum decoupling partitioning system. This method uses the calculation logic of boundary avoidance and point-by-point adjustment. By strictly avoiding environmental interference frequencies, it ensures the purity of spectrum decoupling and is suitable for conventional operating conditions with few types of environmental interference and clear distribution.

[0089] The second approach involves parallel settings for global interval optimization. First, global interval data is read, loading the upper and lower limits of the body vibration fundamental frequency interval, the upper and lower limits of all identified environmental disturbance vibration frequency fluctuation intervals, and the corresponding interference intensity coefficients. The interference intensity coefficients are calculated by weighting the frequency and average amplitude of the environmental disturbance in historical operational data. Then, interval overlap analysis is initiated, traversing all environmental disturbance fluctuation intervals and the body vibration fundamental frequency interval. The overlap length and degree of overlap between each pair of intervals are calculated. The overlap degree is defined as the ratio of the overlap length to the total length of the body vibration fundamental frequency interval. Simultaneously, all high-priority interference intervals that overlap with the body vibration intervals are marked.

[0090] Based on this, a multi-objective optimization objective function is constructed. The objective function F is composed of three weighted components: body vibration coverage F1, environmental interference avoidance F2, and frequency band balance F3. The formula is: F = w1×F1 + w2×F2 + w3×F3, where w1, w2, and w3 are preset weight coefficients with default values ​​of 0.5, 0.3, and 0.2, respectively. Body vibration coverage F1 is defined as the intersection length of the mid-frequency band interval divided by the current boundary point combination and the body vibration fundamental frequency interval, divided by the total length of the body vibration fundamental frequency interval. The value ranges from 0 to 1, with a larger value indicating more complete body vibration coverage. Environmental interference avoidance F2 is defined as 1 minus the weighted sum of interference intensity at the intersection of all environmental disturbance fluctuation intervals and the three divided frequency bands. A larger value indicates better interference avoidance. Frequency band balance F3 is defined as 1 minus the ratio of the standard deviation of the bandwidth of the three frequency bands to the average bandwidth. A larger value indicates more balanced frequency band division.

[0091] A parallel computing architecture is employed, dividing all possible low-frequency and high-frequency boundary point combinations into multiple computational subtasks. These subtasks are then distributed across different computing cores for simultaneous decoupling effect scoring. Each subtask substitutes its corresponding boundary point combination into the objective function to calculate its comprehensive score F. After all subtasks have completed their calculations, the system automatically selects the boundary point combination with the highest comprehensive score. Simultaneously, it verifies whether the bandwidth of the three frequency bands under this combination meets the minimum bandwidth requirement. If it does, it is used as the final hierarchical frequency boundary point, generating the corresponding spectrum decoupling partitioning system. If it does not meet the requirement, the next highest-scoring combination is selected for verification until the optimal combination that meets the requirements is found.

[0092] This method adopts the logic of global multi-objective optimization and full combinatorial parallel computing. By comprehensively considering three core indicators—complete coverage of body vibration, effective avoidance of environmental interference, and balanced frequency band division—it achieves the global optimal configuration of boundary points. It completely solves the problem of reduced decoupling accuracy of traditional fixed boundary frequency or serial avoidance methods when multiple interference intervals overlap and the boundaries between body vibration and environmental interference are blurred. It is especially suitable for harsh working conditions in industrial sites where multiple devices operate densely and ground vibration and electromagnetic interference coexist.

[0093] Fifth embodiment This embodiment provides an exemplary scheme for generating a full-band hierarchical compensation signal. In this example, the hierarchical compensation execution criteria and compensation adjustment intensity corresponding to the three types of jitter components are first matched with the full-band compensation strategy. Then, according to the compensation adjustment intensity, appropriate vibration cancellation calibration is applied to each frequency band component to obtain the compensation component. Finally, deviation offset processing is performed on the three types of jitter components according to the hierarchical compensation execution criteria, and the signal is then processed to obtain the hierarchical compensation signal. Step S30 includes steps D11 to D13: Step D11: In accordance with the full-band compensation strategy, match the hierarchical compensation execution criteria and compensation control intensity corresponding to the high-frequency resonant component, the mid-frequency error component, and the low-frequency disturbance component.

[0094] Step D12: Based on the compensation adjustment strength, apply appropriate vibration cancellation adjustment to the high-frequency resonant component, the mid-frequency error component, and the low-frequency disturbance component respectively to obtain the compensation component.

[0095] Step D13: According to the hierarchical compensation execution criteria and the compensation components, the high-frequency resonance component, the mid-frequency error component, and the low-frequency disturbance component are processed by deviation offsetting to obtain the hierarchical compensation signal.

[0096] Hierarchical compensation execution criteria refer to the specific rules and procedures for executing compensation algorithms tailored to jitter components in different frequency bands. These are the operational specifications for vibration cancellation calibration. Examples include modal active damping execution criteria, model reference adaptive feedforward execution criteria, and extended state observer disturbance compensation execution criteria. Compensation control strength refers to the set of adjustable parameters used to control the intensity of the compensation effect, response speed, and phase characteristics. This is the core control method for achieving precise compensation. Examples include compensation gain coefficient, damping coefficient, feedforward coefficient, observer bandwidth, and response time constant.

[0097] Vibration cancellation calibration refers to the process of generating a cancellation signal with the same amplitude but opposite phase as the original jitter signal, based on compensation execution criteria and control intensity. It is the core action for jitter suppression. Examples include modal damping force generation, adaptive feedforward calculation, and disturbance estimation. The compensation component refers to the independent compensation signal generated after vibration cancellation calibration, used to cancel jitter in the corresponding frequency band. It is the input object for deviation offsetting processing. Examples include high-frequency modal damping compensation components, mid-frequency adaptive feedforward compensation components, and low-frequency extended state observer disturbance compensation components. Deviation offsetting processing refers to the superposition and cancellation process after phase alignment and amplitude matching of the compensation component and the original jitter component in the corresponding frequency band. It is a crucial step in eliminating jitter.

[0098] In this example, when matching the hierarchical compensation execution criteria and the compensation control intensity, the modal active damping, model reference adaptive feedforward, and extended state observer hierarchical compensation methods provided in the first embodiment can be used as a reference. Alternatively, by analyzing the amplitude, frequency, and variation trend of jitter components in each frequency band in real time, the compensation control intensity parameters can be dynamically adjusted to generate adaptive hierarchical compensation execution criteria, thereby achieving dynamic optimization of the compensation effect.

[0099] After matching the hierarchical compensation execution criteria with the compensation control intensity, the vibration cancellation calibration process is initiated. Based on the compensation control intensity, appropriate vibration cancellation calibration is applied to the high-frequency resonant component, the mid-frequency error component, and the low-frequency disturbance component to generate corresponding compensation components. Then, according to the hierarchical compensation execution criteria, the original jitter components of each compensation component and the corresponding frequency band are offset to eliminate the jitter components of each frequency band. Finally, the processed signals of each frequency band are sorted according to the control timing and the amplitude is normalized to obtain the hierarchical compensation signal. In this way, through hierarchical precise matching and dynamic control, the accuracy and adaptability of full-band jitter compensation are improved, avoiding the problem of insufficient or over-compensation when fixed compensation parameters change with operating conditions.

[0100] For example, there are two methods for obtaining a layered compensation signal through deviation offsetting processing. The first is a hierarchical serial offsetting based on frequency band priority. Deviation offsetting processing for each frequency band is performed sequentially according to the priority order: high frequency first, then mid-frequency, and finally low frequency. First, phase alignment and amplitude matching are performed on the high-frequency resonant component and the high-frequency compensation component to offset the high-frequency jitter. Then, synchronous offsetting is performed on the mid-frequency error component and the mid-frequency compensation component to eliminate mid-frequency tracking errors. Finally, offsetting is performed on the low-frequency disturbance component and the low-frequency compensation component to cancel low-frequency load disturbances. After all frequency band offsetting processing is completed, the processed signals of each frequency band are spliced ​​together in time sequence and their amplitudes are normalized to obtain the layered compensation signal. This method uses hierarchical serial processing and priority scheduling logic. By prioritizing the suppression of high-frequency resonance to avoid its interference with mid- and low-frequency compensation, the stability of the compensation is ensured, making it suitable for high-speed motion and conditions with significant high-frequency resonance.

[0101] The second method is multi-dimensional coupled parallel offsetting. It performs global reading and coupling relationship analysis on the high-frequency resonant component, mid-frequency error component, low-frequency disturbance component, and their corresponding compensation components. It identifies cross-coupling interference between different frequency bands and simultaneously initiates parallel deviation offsetting processing for the three frequency bands. During the offsetting process, it compensates for the cross-coupling effects between frequency bands in real time. After all offsetting and coupling compensation for all frequency bands are completed, the processed signals of each frequency band are time-synchronized, weighted, fused, and amplitude normalized to obtain a layered compensation signal. This method employs parallel processing and cross-coupling compensation logic. By processing all frequency bands simultaneously and eliminating mutual interference between them, it improves the overall accuracy and response speed of the compensation, making it suitable for high-precision operation scenarios with multi-modal coupling and complex operating conditions.

[0102] Sixth Embodiment This embodiment provides an exemplary scheme for fusing and aggregating layered deviation hedging and corrective vibration components. In this example, firstly, the compensation component corresponding to the high-frequency resonance component is matched according to the hierarchical compensation execution criterion, and structural adaptation deviation hedging is performed to obtain the high-frequency corrective vibration component. Then, the compensation component corresponding to the mid-frequency error component is matched, and operational state linkage deviation hedging is performed to obtain the mid-frequency corrective vibration component. Next, based on the shaping state of the mid- and high-frequency corrective vibration components, the compensation component corresponding to the low-frequency disturbance component is matched, and scene adaptive deviation hedging is performed to obtain the low-frequency corrective vibration component. Finally, the three types of corrective vibration components are fused and aggregated in an orderly manner according to the spectral arrangement rules to obtain the layered compensation signal. Please refer to... Figure 4 , Figure 4 This is a flowchart illustrating the sixth embodiment of the control method for suppressing robot joint module vibration according to this application. Step D13 includes steps E11-E14: Step E11: According to the hierarchical compensation execution criteria, match the compensation component corresponding to the high-frequency resonant component, and perform structural adaptation deviation offset on the high-frequency resonant component to obtain the high-frequency corrected vibration component.

[0103] Step E12: According to the hierarchical compensation execution criteria, match the compensation component corresponding to the intermediate frequency error component, and perform operational state linkage deviation offsetting on the intermediate frequency error component to obtain the intermediate frequency corrected vibration component.

[0104] Step E13: Based on the shaping state of the high-frequency corrected vibration component and the mid-frequency corrected vibration component, match the compensation component corresponding to the low-frequency disturbance component, and perform scene adaptive deviation offsetting on the low-frequency disturbance component to obtain the low-frequency corrected vibration component.

[0105] Step E14: The high-frequency corrected vibration component, the mid-frequency corrected vibration component, and the low-frequency corrected vibration component are fused and aggregated in an orderly manner according to the spectral arrangement rules to obtain the layered compensation signal.

[0106] Structural adaptation offsetting refers to the method of adjusting the offsetting phase and amplitude to cancel out the deviations by combining the inherent characteristics of the joint module's mechanical structure (such as reducer type and connecting rod stiffness), which can accurately match the resonance characteristics of different structures.

[0107] High-frequency corrected vibration components refer to vibration signals that have had high-frequency mechanical resonance eliminated after structurally adapted deviation offsetting, serving as the foundational input for subsequent mid-frequency and low-frequency compensation. Mid-frequency error components refer to mid-frequency fluctuation signals in the 20Hz to 200Hz range caused by servo tracking errors and positioning overshoot, and are the target of mid-frequency layer deviation offsetting. Operational state-linked deviation offsetting refers to a deviation cancellation method that dynamically adjusts the offsetting gain based on the joint's current operating speed, acceleration, and load, adapting to tracking error characteristics under different motion states.

[0108] The intermediate frequency (IF) corrected vibration component refers to the vibration signal after being offset by a linkage-based deviation adjustment during operation, eliminating IF tracking error and overshoot. The forming state refers to the amplitude attenuation, phase deviation, and stability indicators of the IF corrected vibration components, used to guide the dynamic adjustment of low-frequency compensation.

[0109] Low-frequency disturbance components refer to low-frequency drift signals below 20 Hz caused by load changes, friction, and external force disturbances; these are the targets for low-frequency layer deviation offsetting. Scene-adaptive deviation offsetting refers to a deviation cancellation method that adjusts the observer bandwidth and compensation strength according to the on-site operation scenario (such as precision assembly, heavy-duty handling, and high-speed sorting), adapting to the disturbance characteristics of different scenarios. Low-frequency corrected vibration components refer to vibration signals that have eliminated low-frequency load and friction disturbances after scene-adaptive deviation offsetting.

[0110] The spectral arrangement rule refers to the rule of arranging the corrected vibration components in order of frequency from high to low, which can ensure the timing consistency and phase accuracy of the fused signal.

[0111] In this example, when performing layered deviation offsetting and fusion aggregation, the method of full-band layered compensation can be referenced. Alternatively, by monitoring the shaping state of the corrected vibration components in each frequency band in real time, the parameters and timing of low-frequency deviation offsetting can be dynamically adjusted to achieve coordinated optimization of high, medium, and low-frequency compensation, thereby improving the overall compensation effect.

[0112] After matching the high-frequency resonant component and the high-frequency compensation component, a structurally adaptive deviation offsetting process is initiated. Based on the reducer type and linkage stiffness parameters of the joint module, the phase and amplitude of the high-frequency compensation component are adjusted to precisely offset the high-frequency resonant component, resulting in a high-frequency corrected vibration component. Next, the intermediate-frequency compensation component corresponding to the intermediate-frequency error component is matched. Considering the current operating speed and load of the joint, a state-linked deviation offsetting process is executed to obtain the intermediate-frequency corrected vibration component. Subsequently, based on the amplitude attenuation rate and phase deviation of the intermediate-frequency corrected vibration component, the observer bandwidth and compensation gain of the low-frequency compensation component are dynamically adjusted. Scene-adaptive deviation offsetting is then performed on the low-frequency disturbance component to obtain the low-frequency corrected vibration component. Finally, the three types of corrected vibration components are time-synchronized and fused according to the spectral arrangement from high to low frequency. After amplitude normalization, a layered compensation signal is obtained. Through layered differentiated offsetting and collaborative optimization, the accuracy and adaptability of full-band jitter compensation are improved, and mutual interference between compensations in different frequency bands is avoided.

[0113] For example, there are two methods for obtaining layered compensation signals through layered deviation offsetting and fusion aggregation. The first is a time-progressive offsetting fusion based on spectral priority. Deviation offsetting is performed sequentially for each frequency band, prioritizing high frequencies, followed by mid-frequency, and finally low frequencies. First, high-frequency structural adaptation deviation offsetting is completed, and the stability of the high-frequency corrected vibration component is verified. After the high-frequency resonance is completely suppressed, mid-frequency operating state linkage deviation offsetting is initiated to eliminate mid-frequency tracking errors. Finally, based on stable mid-to-high frequency corrected vibration components, low-frequency scene adaptive deviation offsetting is performed to counteract low-frequency disturbances. After all frequency band offsetting is completed, the three types of corrected vibration components are sequentially spliced ​​and their amplitudes normalized to obtain the layered compensation signal. This method employs a time-progressive and step-by-step verification logic. By prioritizing the suppression of high-frequency resonances to avoid their interference with mid-to-low frequency compensation, it ensures the stability of the compensation system and is suitable for high-speed motion and conditions with severe high-frequency resonance.

[0114] The second method involves parallel processing of high and medium frequencies with low-frequency feedback linkage for offsetting. This simultaneously initiates high-frequency structural adaptation offsetting and mid-frequency operating state linkage offsetting, real-time acquisition of the forming state data of the high and medium frequency corrected vibration components, and feedback to the low-frequency compensation module. The low-frequency compensation module dynamically adjusts its compensation parameters and offsetting timing based on the compensation effect of the high and medium frequencies, performing scenario-adaptive offsetting. After all offsetting processing in the three frequency bands is completed, the three types of corrected vibration components undergo global temporal synchronization, cross-coupling compensation, and weighted fusion. After amplitude normalization, a layered compensation signal is obtained. This method employs parallel processing and feedback linkage logic. Parallel processing of high and medium frequencies improves response speed, while low-frequency feedback linkage achieves global optimization of the compensation effect, improving compensation accuracy and dynamic performance under complex operating conditions. It is suitable for high-precision operation scenarios with multi-modal coupling and frequent load changes.

[0115] For example, the system synchronously initiates high-frequency structural adaptation deviation offsetting and mid-frequency operating state linkage deviation offsetting with a 10kHz control cycle. The high-frequency module adjusts the phase offset and damping coefficient of the high-frequency modal damping compensation component in real time according to the pre-stored cycloidal wheel meshing stiffness of the rotary vector reducer (RV) and the first-order natural frequency parameters of the connecting rod, and performs point-by-point phase alignment deviation offsetting with the high-frequency resonant component. The mid-frequency module synchronously collects the current running speed, acceleration and load torque data of the joint, and dynamically adjusts the gain coefficient and update step size of the model reference adaptive feedforward compensation according to the preset speed and gain mapping table, and performs operating state linkage deviation offsetting with the mid-frequency error component. Simultaneously, the system collects the amplitude attenuation rate, phase deviation, and fluctuation standard deviation of the high-frequency correction vibration components in real time at a sampling frequency of 20kHz as forming state data. This data is fed back to the low-frequency compensation module via an internal high-speed bus. The low-frequency compensation module first calculates the high-frequency comprehensive stability index (amplitude attenuation rate × 0.6 + reciprocal of phase deviation × 0.3 + reciprocal of fluctuation standard deviation × 0.1). When the comprehensive stability reaches 95% or higher, the system automatically initiates scene adaptive deviation offsetting and dynamically adjusts the bandwidth (adjustment range 5Hz~20Hz) and disturbance compensation gain (adjustment coefficient 0.8~1.2) of the Extended State Observer (ESO) based on the high-frequency stability value. When the high-frequency stability is below 80%, the system automatically reduces the low-frequency compensation gain and extends the offsetting response time to avoid system oscillation caused by compensation superposition. After the deviation offsetting processing of the three frequency bands is completed, the system first performs nanosecond-level global timing synchronization on the three types of corrected vibration components based on a unified clock reference. Then, it generates a cross-coupling compensation amount based on the real-time calculated cross-coupling coefficients of the high, medium and low frequency bands (calculated from the cross-correlation function of the vibration energy of adjacent frequency bands), which is superimposed on the corrected vibration components of the corresponding frequency bands to eliminate mutual interference between frequency bands. Finally, it performs weighted fusion according to the dynamic weights of 0.2 for high frequency, 0.5 for medium frequency and 0.3 for low frequency, and normalizes the fused signal to the standard control voltage range of -10V to +10V to obtain the final hierarchical compensation signal. This method reduces the compensation response delay to less than 100 microseconds through parallel processing of high and medium frequencies and achieves global collaborative optimization of the compensation effect across the entire frequency band through low-frequency feedback linkage. It completely solves the problem of mutual interference between compensations of different frequency bands under multimodal coupling conditions, and is especially suitable for high-precision operation scenarios such as precision assembly and electronic manufacturing with frequent load fluctuations and high-speed trajectory switching.

[0116] Seventh Embodiment This embodiment provides an exemplary scheme for fusing layered compensation signals and servo three-loop control signals and generating drive commands. In this example, the layered compensation signals are first synchronized with the timing and adapted to the working conditions based on the servo three-loop control signals during the operation of the robot joint module to obtain calibrated layered compensation signals. Then, the calibrated layered compensation signals are embedded into the servo three-loop control signals and superimposed to form control signals. Next, the control signals are logically normalized and transformed according to the robot's operating logic to obtain the drive commands corresponding to the joint modules. Finally, the drive commands are used to control the joint modules to complete the corresponding actions and suppress various vibration phenomena of the robot body. Step S40 includes steps F11 to F14: Step F11: Based on the servo three-loop control signal during the operation of the robot joint module, perform time synchronization and working condition dimension adaptation calibration on the hierarchical compensation signal to obtain the calibrated hierarchical compensation signal.

[0117] Step F12: Embed the calibrated layered compensation signal into the servo three-loop control signal and superimpose them to form a control signal.

[0118] Step F13: Perform logic regularization and transformation processing on the control signal according to the robot's operating logic to obtain the drive command corresponding to the joint module.

[0119] Step F14: Control the joint module to complete the corresponding action through the drive command, and suppress various shaking phenomena of the machine body through the compensation signal in the drive command.

[0120] Servo three-loop control signals refer to the basic control signals output by the position loop, speed loop, and current loop of the servo driver. They are the reference signals for the calibration and fusion of layered compensation signals. For example, the position control signal output by the position loop, the speed control signal output by the speed loop, and the current control signal output by the current loop. Operating condition dimension adaptation calibration refers to the calibration process of dynamically adjusting the amplitude, phase, and bandwidth of the layered compensation signals based on operating condition parameters such as the current operating speed, load, and motion posture of the joint module.

[0121] The calibrated layered compensation signal refers to the compensation signal that, after timing synchronization and working condition adaptation calibration, is completely matched with the current servo control state. The control signal refers to the comprehensive control signal formed by embedding the calibrated layered compensation signal into the servo three-loop control signal and superimposing it, encompassing basic control and jitter compensation. The robot's operating logic refers to the preset joint motion trajectory, speed limits, acceleration limits, safety protection rules, and other control logic, which serves as the basis for control signal conversion. Logic normalization and conversion processing refers to the process of performing amplitude limiting protection, dead zone compensation, nonlinear correction, and format conversion on the control signal to make it conform to the servo driver input requirements. Various body jitter phenomena refer to vibrations and jitters across the entire frequency range, covering high-frequency mechanical resonance, mid-frequency tracking error, and low-frequency load disturbances, and are the ultimate targets of this solution for suppression.

[0122] In this example, when fusing the hierarchical compensation signal and the servo three-loop control signal and generating drive commands, a direct weighted linear superposition fusion method can be used. Alternatively, by monitoring the status of the servo three-loop control signal and the joint operating conditions in real time, the embedding position and fusion weight of the hierarchical compensation signal can be dynamically adjusted to achieve optimal fusion of the compensation signal and the basic control signal, thereby improving the control accuracy and jitter suppression effect of the drive commands.

[0123] After generating the layered compensation signal, the timing synchronization and working condition adaptation calibration process is initiated. Based on the time reference of the servo three-loop control signals, the layered compensation signal is precisely time-aligned. Simultaneously, considering the joint's current running speed, load, and motion posture, the amplitude, phase, and bandwidth of the layered compensation signal are dynamically adjusted to obtain the calibrated layered compensation signal. Next, the calibrated layered compensation signal is embedded into the servo position loop, velocity loop, and current loop according to preset embedding rules, and superimposed with the basic control signals of each loop to form the control signal. Subsequently, according to the robot's preset operating logic, the control signal undergoes logic regularization processing such as amplitude limiting protection, dead zone compensation, nonlinear correction, and format conversion to obtain drive commands that meet the input requirements of the servo driver. Finally, the drive commands are output to the servo driver to control the joint motors to run smoothly. The layered compensation signal embedded in the drive command cancels out various jitter phenomena generated by the machine in real time. Through precise timing synchronization and embedded fusion, the coordination between the compensation signal and the basic control signal is improved, avoiding compensation lag and overcompensation problems, and achieving precise suppression of jitter across the entire frequency band.

[0124] For example, there are two methods for generating control signals by embedding hierarchical compensation signals into the servo three-loop control signals. The first is a serial hierarchical embedding fusion based on the control level. Following the control level order of position loop, speed loop, and current loop, the hierarchical compensation signals for the corresponding frequency bands are embedded into the control signals of each loop sequentially. First, the low-frequency disturbance compensation component is embedded into the output of the position loop to compensate for position deviations caused by low-frequency load changes. Then, the intermediate-frequency error compensation component is embedded into the output of the speed loop to eliminate intermediate-frequency tracking errors and overshoot. Finally, the high-frequency resonance compensation component is embedded into the output of the current loop to suppress high-frequency mechanical resonance. After all the compensation signals for all frequency bands are embedded, they are superimposed to form the final control signal. This method uses serial hierarchical embedding with precise matching logic of frequency bands and loop levels. By embedding compensation signals for different frequency bands into different control loops, it fully utilizes the adjustment characteristics of each control loop, ensuring the stability of the compensation system and making it suitable for jitter suppression under normal operating conditions.

[0125] The second method involves multi-loop parallel collaborative embedding and fusion. This method decomposes the calibrated hierarchical compensation signal into position loop compensation components, velocity loop compensation components, and current loop compensation components, embedding them in parallel within the three-loop servo control signal. Simultaneously, during the fusion process, the cross-coupling effects between loops are calculated in real time, and the weights of the compensation components in each loop are dynamically adjusted to eliminate inter-loop coupling interference. After the compensation embedding and coupling compensation of all three loops are completed, they are superimposed to form the final control signal. This method employs parallel embedding and cross-coupling compensation logic. By simultaneously applying compensation to the three control loops and eliminating inter-loop interference, it improves the compensation response speed and overall accuracy, making it suitable for complex operating scenarios requiring high-speed motion and high precision.

[0126] Eighth embodiment This embodiment provides an exemplary scheme for closed-loop feedback optimization of full-band jitter suppression effect. In this example, full-domain vibration state feedback data during the operation of the joint module is first collected. Then, this data is compared with preset stable operation standard data to determine the actual implementation difference of full-band jitter suppression. Next, based on the difference, the control defects and adaptation loopholes that are mismatched between the hierarchical compensation method and the on-site working conditions are identified. Finally, based on the defects and loopholes, the full-band compensation strategy is adjusted and the hierarchical compensation execution criteria and compensation control intensity are finely adjusted to optimize the jitter suppression effect in the next iteration. Please refer to... Figure 5 , Figure 5 This is a flowchart illustrating the eighth embodiment of the control method for suppressing robot joint module vibration according to this application. Following step S40, steps G11-G14 are also included: Step G11: During the operation of the robot joint module, collect the global vibration state feedback data generated during the operation of the joint module.

[0127] Step G12 involves comparing the full-range vibration state feedback data with the preset stable operation standard data in all aspects to determine the actual implementation differences of the current full-band jitter suppression.

[0128] Step G13: Based on the actual implementation differences, identify the control defects and adaptation loopholes where the location-level compensation method does not match the on-site operating conditions.

[0129] Step G14: Based on the aforementioned control defects and adaptation loopholes, adjust the full-band compensation strategy, and fine-tune the hierarchical compensation execution criteria and compensation control intensity corresponding to the vibration components according to the strategy adjustment range, so as to optimize the vibration reduction effect of the next jitter suppression.

[0130] Global vibration status feedback data refers to the set of multi-dimensional vibration and control feedback signals collected in real time during the operation of the joint module, which comprehensively reflects the vibration suppression effect and is the core basis for evaluating the suppression effect. Examples include joint motor current signals, encoder position / velocity signals, joint output acceleration signals, joint torque signals, servo three-loop control output signals, and trajectory tracking error signals. Stable operation standard data refers to the preset set of benchmark signals and performance indicators for the joint module to operate smoothly without vibration under different working conditions, serving as a reference standard for judging differences in suppression effect. Examples include vibration amplitude thresholds under no-load / full-load conditions, vibration energy proportion standards for each frequency band, allowable range of trajectory tracking error, positioning accuracy indicators, and response time indicators.

[0131] Actual implementation discrepancies refer to the quantitative differences in amplitude, frequency, phase, and error between the full-domain vibration state feedback data and the standard data for stable operation, which form the basis for positioning and control defects. Examples include differences in high-frequency resonant residual amplitude, mid-frequency tracking overshoot, low-frequency disturbance compensation error, and phase lag. Control defects refer to problems such as improper parameter settings and illogical logic in the execution of compensation algorithms at various levels within the full-band compensation strategy, which are internal causes of poor suppression effects. For example, excessively large modal active damping coefficients lead to system overdamping, excessively slow adaptive feedforward gain update rates cause tracking lag, and insufficient bandwidth of the Extended State Observer (ESO) leads to inaccurate disturbance estimation.

[0132] Adaptation flaws refer to the mismatch between the full-band compensation strategy and the actual operating conditions on site, which are external causes of poor suppression effects. For example, the compensation parameters may not be adapted to the cycloidal vibration characteristics of the rotary vector reducer (RV), the impact of on-site electromagnetic interference on high-frequency signals may not be considered, and the load change characteristics under heavy load conditions may not be adapted.

[0133] In this example, when collecting global vibration state feedback data, a multi-source signal acquisition method can be used. Alternatively, real-time streaming acquisition can be employed, simultaneously controlling the joint module's operation while synchronously collecting feedback data from various dimensions to accumulate complete global vibration state feedback data, thereby completing the data acquisition for evaluating the suppression effect.

[0134] After collecting the full-range vibration status feedback data, the suppression effect evaluation process is initiated. The feedback data is compared with the preset stable operation standard data in a comprehensive and multi-dimensional manner to calculate the actual implementation differences of each frequency band. Then, based on the distribution characteristics of the actual implementation differences, the control defects and field condition adaptation loopholes of the corresponding hierarchical compensation methods are located. Finally, according to the type and severity of the defects and loopholes, the full-band compensation strategy is dynamically adjusted, the hierarchical compensation execution criteria and compensation control intensity of the corresponding vibration components are fine-tuned, and the compensation parameter library is updated. In this way, through closed-loop feedback optimization, the full-band jitter suppression effect is continuously improved, solving the problem of the suppression effect of fixed compensation parameters decreasing after long-term operation or changes in operating conditions.

[0135] For example, there are two ways to adjust the full-band compensation strategy based on actual implementation differences. The first is a successive closed-loop optimization based on the division of responsibilities among frequency bands. Following the order of high frequency, mid frequency, and low frequency, the actual implementation differences of each frequency band are compared sequentially to locate the control defects and adaptation loopholes in the corresponding compensation level. If there is resonance residue in the high-frequency band, the active modal damping coefficient and resonant frequency tracking parameters are adjusted. If there is tracking overshoot in the mid-frequency band, the adaptive feedforward gain update rate and feedforward coefficient are adjusted. If there is disturbance compensation error in the low-frequency band, the bandwidth of the extended state observer and the disturbance estimation gain are adjusted. After all defects and loopholes in all frequency bands are located, the compensation execution criteria and control intensity at the corresponding level are fine-tuned sequentially to complete the optimization of the full-band compensation strategy. This method uses the logic of dividing responsibilities among frequency bands and successive adjustments. By accurately locating problems in a single frequency band and optimizing them specifically, it ensures the stability and traceability of the optimization process, making it suitable for conventional operating conditions where the suppression effect of a single frequency band is poor.

[0136] The second approach is global multi-dimensional collaborative closed-loop optimization. This method performs global reading and multi-dimensional feature analysis on the vibration state feedback data across the entire domain. It simultaneously extracts the actual implementation differences across the high, mid, and low frequency bands, as well as the cross-coupling effects between different frequency bands, constructing a multi-dimensional evaluation system encompassing suppression effect, system stability, and response speed. Based on this evaluation system, the optimal adjustment amounts for compensation parameters at all levels are calculated synchronously. Considering the coupling effects between frequency bands, the entire frequency band compensation strategy is globally adjusted, and the compensation execution criteria and control intensity at each level are uniformly fine-tuned. After all parameters are adjusted, the comprehensive performance of the optimized compensation strategy is verified, ensuring that the suppression effect is improved without reducing the system's dynamic response and stability. This method employs global multi-dimensional analysis and collaborative optimization logic. By simultaneously considering the problems and cross-coupling effects of all frequency bands, it achieves the globally optimal configuration of the compensation strategy, making it suitable for high-precision operation scenarios with severe multi-frequency coupling and complex, variable operating conditions.

[0137] This application provides a control device for suppressing robot joint module jitter. The control device for suppressing robot joint module jitter includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the control method for suppressing robot joint module jitter in the above embodiment 1.

[0138] The following is for reference. Figure 6 The diagram illustrates a structural schematic of a control device suitable for implementing the embodiments of this application to suppress robot joint module jitter. The control device for suppressing robot joint module jitter in the embodiments of this application may include, but is not limited to, mobile terminals such as joint active dampers, magnetorheological joint dampers, electromagnetic joint dampers, high-torque rotary dampers, etc., as well as fixed terminals such as joint module dedicated servo drivers, integrated joint drive and control units, high-performance motion control platforms, etc. Figure 6 The control device shown for suppressing robot joint module vibration is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments of this application.

[0139] like Figure 6As shown, the control device for suppressing robot joint module jitter may include a processing unit 1001 (e.g., a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 1002 or a program loaded from a storage device 1003 into a random access memory (RAM) 1004. The RAM 1004 also stores various programs and data required for the operation of the control device for suppressing robot joint module jitter. The processing unit 1001, the read-only memory 1002, and the RAM 1004 are interconnected via a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to I / O interface 1006: input devices 1007 including, for example, touchscreens, touchpads, keyboards, mice, image sensors, microphones, accelerometers, gyroscopes, etc.; output devices 1008 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 1003 including, for example, magnetic tapes, hard disks, etc.; and communication devices 1009. Communication device 1009 allows the control device for suppressing robot joint module jitter to communicate wirelessly or wiredly with other devices to exchange data. Although control devices for suppressing robot joint module jitter with various systems are shown in the figures, it should be understood that it is not required to implement or possess all of the systems shown. More or fewer systems may be implemented alternatively.

[0140] Specifically, according to the embodiments disclosed in this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments disclosed in this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device, or installed from storage device 1003, or installed from read-only memory 1002. When the computer program is executed by processing device 1001, it performs the functions defined in the methods of the embodiments disclosed in this application.

[0141] The control device for suppressing robot joint module vibration provided in this application employs the control method for suppressing robot joint module vibration in the above embodiments, which can solve the technical problem of poor robot joint anti-vibration effect. Compared with the prior art, the beneficial effects of the control device for suppressing robot joint module vibration provided in this application are the same as the beneficial effects of the control method for suppressing robot joint module vibration provided in the above embodiments, and other technical features in the control device for suppressing robot joint module vibration are the same as those disclosed in the previous embodiment method, and will not be repeated here.

[0142] It should be understood that the various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0143] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0144] This application provides a computer-readable storage medium having computer-readable program instructions (i.e., a computer program) stored thereon, the computer-readable program instructions being used to execute the control method for suppressing robot joint module jitter in the above embodiments.

[0145] The computer-readable storage medium provided in this application may be, for example, a USB flash drive, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, system, or device. The program code contained on the computer-readable storage medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, radio frequency (RF), etc., or any suitable combination thereof.

[0146] The aforementioned computer-readable storage medium may be included in a control device for suppressing robot joint module jitter; or it may exist independently and not assembled into a control device for suppressing robot joint module jitter.

[0147] The aforementioned computer-readable storage medium carries one or more programs. When these programs are executed by a control device for suppressing robot joint module jitter, the control device for suppressing robot joint module jitter performs the following actions: based on the robot's operating state, it correlates and aligns multi-source motion signals according to a clock reference to obtain an original jitter signal set; it decouples the original jitter signal set according to the boundary frequency spectrum to separate high-frequency resonant components, intermediate-frequency error components, and low-frequency disturbance components; it performs layered compensation of the high-frequency resonant components, intermediate-frequency error components, and low-frequency disturbance components according to a full-band compensation strategy, and organizes them to obtain a layered compensation signal; it superimposes and fuses the layered compensation signal with the servo three-loop control signal to generate a drive command for controlling the operation of the joint module, thereby suppressing the full-band jitter of the robot joint module.

[0148] Computer program code for performing the operations of this application can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0149] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation that may be implemented in systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0150] The modules described in the embodiments of this application can be implemented in software or hardware. The names of the modules do not necessarily limit the functionality of the unit itself.

[0151] The readable storage medium provided in this application is a computer-readable storage medium that stores computer-readable program instructions (i.e., a computer program) for executing the above-described control method for suppressing robot joint module jitter, thereby solving the technical problem of poor robot joint anti-shaking effect. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in this application are the same as the beneficial effects of the control method for suppressing robot joint module jitter provided in the above embodiments, and will not be repeated here.

[0152] The above description is only a part of the embodiments of this application and does not limit the patent scope of this application. All equivalent structural transformations made under the technical concept of this application and using the contents of the specification and drawings of this application, or direct / indirect applications in other related technical fields, are included in the patent protection scope of this application.

Claims

1. A control method for suppressing vibration of a robot joint module, characterized in that, The method includes: Based on the robot's operating state, the multi-source motion signals are correlated and aligned according to the clock reference to obtain the original jitter signal set; The original jitter signal set is decoupled according to the boundary frequency spectrum to separate the high-frequency resonant component, the intermediate frequency error component and the low-frequency disturbance component. The high-frequency resonant component, the mid-frequency error component, and the low-frequency disturbance component are compensated in layers according to the full-band compensation strategy, and the layered compensation signal is obtained by sorting. The layered compensation signal is superimposed and fused with the servo three-loop control signal to generate drive commands for controlling the operation of the joint module, thereby suppressing the full-frequency jitter of the robot joint module.

2. The control method for suppressing robot joint module vibration as described in claim 1, characterized in that, The step of correlating and aligning multi-source motion signals according to a clock reference based on the robot's operating state to obtain the original jitter signal set includes: Based on the robot's operating status data and the clock reference, a timing alignment interval is defined to determine the correction reference for the multi-source motion signal; The multi-source motion signal is calibrated for timing deviation based on the calibration benchmark, and the time-calibrated motion signal is obtained by correlation. The time-calibrated motion signal is fused and arranged with the robot's posture operation signal to form a global motion signal group; Effective fluctuation signals with body sway and offset characteristics are selected from the global motion signal set, and the effective fluctuation signals are aligned according to the clock reference to obtain the original jitter signal set.

3. The control method for suppressing robot joint module vibration as described in claim 1, characterized in that, The step of decoupling the original jitter signal set according to the boundary frequency spectrum to separate the high-frequency resonant component, the intermediate frequency error component, and the low-frequency disturbance component includes: Based on the original jitter signal, and combined with the vibration law of robot mechanical transmission and the characteristics of interference in the field environment, a spectrum decoupling partitioning system including the boundary frequency is constructed. The various vibration signals in the original jitter signal set are coupled and correlated and decomposed using the aforementioned spectrum decoupling partitioning system to obtain layered spectrum signals; Based on the defined boundary frequency, the signal level and signal category of the layered spectrum signal are identified, and individual signals with different vibration properties are obtained. Individual signals of the same type and frequency band are uniformly grouped together to obtain the high-frequency resonant component, the mid-frequency error component, and the low-frequency disturbance component.

4. The control method for suppressing robot joint module jitter as described in claim 3, characterized in that, The steps for constructing a spectrum decoupling partitioning system, including the boundary frequency, based on the original jitter signal and combining the vibration law of robot mechanical transmission with the characteristics of interference in the field environment include: Extract the global vibration distribution features from the original jitter signal and distinguish the interference feature ranges corresponding to the body's own vibration and external environmental disturbances; Based on the range of the interference characteristics, and combined with the robot's full-scene operation data, the inherent vibration evolution law of the mechanical rotation result of the body is sorted out, and the basic frequency range of the body vibration is determined. Using the aforementioned basic frequency range as a benchmark, and combining it with the actual working conditions on site, we analyze the vibration characteristics caused by various external disturbances and define the vibration frequency fluctuation range corresponding to various environmental disturbances. By integrating the basic frequency range and the vibration frequency floating range, and uniformly setting hierarchical frequency boundary points, a spectrum decoupling partitioning system including the boundary frequencies is obtained.

5. The control method for suppressing robot joint module vibration as described in claim 1, characterized in that, The step of performing layered compensation of the high-frequency resonant component, the intermediate-frequency error component, and the low-frequency disturbance component according to a full-band compensation strategy, and then processing them to obtain a layered compensation signal includes: In accordance with the full-band compensation strategy, the hierarchical compensation execution criteria and compensation control intensity are matched to the high-frequency resonant component, the mid-frequency error component, and the low-frequency disturbance component. Based on the compensation adjustment intensity, appropriate vibration cancellation adjustment is applied to the high-frequency resonant component, the mid-frequency error component, and the low-frequency disturbance component respectively to obtain the compensation component; According to the hierarchical compensation execution criteria and the compensation components, the high-frequency resonance component, the mid-frequency error component, and the low-frequency disturbance component are processed by deviation offsetting to obtain the hierarchical compensation signal.

6. The control method for suppressing robot joint module jitter as described in claim 5, characterized in that, The step of processing the high-frequency resonant component, the mid-frequency error component, and the low-frequency disturbance component according to the hierarchical compensation execution criteria and the compensation components, and then processing them to obtain the hierarchical compensation signal includes: According to the hierarchical compensation execution criteria, the compensation component corresponding to the high-frequency resonance component is matched, and the high-frequency resonance component is subjected to structural adaptation deviation offsetting to obtain the high-frequency corrected vibration component. According to the hierarchical compensation execution criteria, the compensation component corresponding to the intermediate frequency error component is matched, and the intermediate frequency error component is subjected to operational state linkage deviation offsetting to obtain the intermediate frequency corrected vibration component. Based on the shaping state of the high-frequency corrected vibration component and the mid-frequency corrected vibration component, the compensation component corresponding to the low-frequency disturbance component is matched, and the low-frequency disturbance component is subjected to scene adaptive deviation offsetting to obtain the low-frequency corrected vibration component. The high-frequency corrected vibration component, the mid-frequency corrected vibration component, and the low-frequency corrected vibration component are systematically fused and aggregated according to the spectral arrangement rules to obtain the layered compensation signal.

7. The control method for suppressing robot joint module jitter as described in claim 1, characterized in that, The step of superimposing and fusing the layered compensation signal with the servo three-loop control signal to generate drive commands for controlling the operation of the joint module, thereby suppressing the full-band jitter of the robot joint module, includes: Based on the servo three-loop control signals during the operation of the robot joint module, the hierarchical compensation signal is time-synchronized and adapted to the working condition dimension to obtain the calibrated hierarchical compensation signal. The calibrated hierarchical compensation signal is embedded into the servo three-loop control signal and superimposed to form a control signal; The control signal is logically normalized and transformed according to the robot's operating logic to obtain the drive command corresponding to the joint module; The drive commands control the joint module to complete corresponding actions, and the compensation signals within the drive commands suppress various types of body vibration.

8. The control method for suppressing robot joint module jitter as described in claim 1, characterized in that, After the step of superimposing and fusing the layered compensation signal with the servo three-loop control signal to generate drive commands for controlling the operation of the joint module to suppress the full-band jitter of the robot joint module, the control method for suppressing the jitter of the robot joint module further includes: During the operation of the robot joint module, the global vibration state feedback data generated during the operation of the joint module is collected. The full-range vibration state feedback data is compared with the preset stable operation standard data in all aspects to determine the actual implementation difference of the current full-band jitter suppression. Based on the actual implementation differences, there are control defects and adaptation loopholes due to the mismatch between the positioning-level compensation method and the on-site operating conditions. Based on the aforementioned control defects and adaptation loopholes, the full-band compensation strategy is adjusted, and the hierarchical compensation execution criteria and compensation control intensity corresponding to the vibration components are fine-tuned according to the strategy adjustment range, so as to optimize the vibration reduction effect of the next jitter suppression.

9. A control device for suppressing vibration of a robot joint module, characterized in that, The control device for suppressing robot joint module jitter includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the control method for suppressing robot joint module jitter as described in any one of claims 1 to 8.

10. A storage medium, characterized in that, The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, it implements the steps of the control method for suppressing robot joint module jitter as described in any one of claims 1 to 8.

Images