Intelligent servo tuning and diagnostics method for enhancing dynamic path accuracy of robotic manipulators

TWI946004BActive Publication Date: 2026-11-11NAT YANG MING CHIAO TUNG UNIV
View PDF 3 Cites 0 Cited by

Patent Information

Application Number
TW114131111
Authority / Receiving Office
TW · TW
Patent Type
Patents
Current Assignee / Owner
Filing Date
2025-08-14
Publication Date
2026-11-11
Estimated Expiration
2045-08-13

Smart Images

  • Figure TWG2TB001920756_001
    Figure TWG2TB001920756_001
  • Figure TWG2TB001920756_002
    Figure TWG2TB001920756_002
  • Figure TWG2TB001920756_003
    Figure TWG2TB001920756_003
Patent Text Reader

Abstract

This invention provides an intelligent servo tuning and diagnostic method for improving the dynamic trajectory accuracy of robotic arms. It addresses servo mismatch issues in multi-joint axis robotic arms, thereby improving dynamic trajectory accuracy. A virtual servo drive model can be used to simulate and evaluate the performance of the robotic arm's servo drive system. For servo mismatch diagnosis, a servo diagnostic model is used to analyze trajectory tracking errors and identify the worst-performing axis of the robotic arm. Finally, a servo parameter adjustment algorithm based on Lagrange Polynomial Interpolation (LPI) enables effective servo parameter adjustment without requiring large datasets or long convergence times. This method significantly enhances the dynamic response of each joint axis servo and improves the overall system path accuracy. In circular motion, it can improve the average dynamic trajectory accuracy by more than 67.38%.
Need to check novelty before this filing date? Find Prior Art

Claims

1. A method for intelligent servo tuning and diagnostics to improve the dynamic trajectory accuracy of a robotic arm, comprising the following steps after being loaded into a device: Using a servo adjustment parameter algorithm based on Lagrange Polynomial Interpolation (LPI) as iterative servo adjustment parameters, establishing a correspondence between the gain of the servo parameters of a multi-joint robotic arm and the feedback signal characteristics of the robotic arm, and iteratively updating the servo parameters of the robotic arm; the servo parameters include the proportional gain of the position loop, the proportional gain of the velocity loop, and the integral time constant of the robotic arm; the characteristics include the gain deviation and phase difference of the robotic arm; the servo adjustment parameter algorithm adjusts the proportional gain of the position loop and the proportional gain of the velocity loop through the gain deviation, and adjusts the integral time constant through the phase difference; The robotic arm executes a motion trajectory, thereby causing the servo adjustment parameter algorithm to calculate the characteristics in the first iteration, and using the Newton-Raphson method. The method calculates the gain parameter for each joint axis and obtains the suggested servo parameters for each joint axis of the robotic arm. In subsequent iterations, new features are calculated, and the worst axis of the robotic arm is identified through a servo diagnostic model. The suggested servo gain for subsequent iterations is updated by interpolation of the previous order Lagrange polynomial. The servo parameters of the robotic arm are updated based on the suggested servo gain.

2. The method for intelligent servo tuning and diagnostics to improve the dynamic trajectory accuracy of a robotic arm as described in claim 1, further comprising a stopping condition including at least one of a gain deviation threshold and a phase difference threshold, and a maximum number of iterations.

3. The method for intelligent servo tuning and diagnostics as described in claim 1 to improve the dynamic trajectory accuracy of a robotic arm, wherein, The Newton-Raphson system is defined as shown in Equation 1: Equation 1; where is the proposed servo parameter for the first iteration; is the initial servo gain; is the feature; and are the upper and lower boundaries of the feature axis in the servo gain and feature; i is the axis of each joint.

4. The method for intelligent servo tuning and diagnostics as described in claim 1 to improve the dynamic trajectory accuracy of a robotic arm, wherein, The Lagrange polynomial interpolation system is defined as shown in Equation 2: Equation 2; where the index symbol indicates the value of the i-axis at the g-th iteration, n is the order of the current polynomial and the number of iterations; is the servo gain determined by the servo diagnostic model and corresponding to the selected feature; is the Lagrange basis function for feature input evaluation; and its value is gain deviation or phase difference; is the position loop proportional gain; is the velocity loop proportional gain; and is the reciprocal of the integral time constant.

5. The method for intelligent servo tuning and diagnostics to improve the dynamic trajectory accuracy of a robotic arm, as described in any one of claims 1 to 4, further includes a servo diagnostic model training step, the step comprising: simulating the control response of the robotic arm through a virtual servo drive model and generating a training dataset to train the servo diagnostic model; wherein the training dataset has time-domain motion data, the motion data includes position loops, velocity loops and current loops, and defines discrete controller models for the position loops and velocity loops, and correspondingly sets their velocity command filters and current command filters, and sets their respective constraint functions; and causing the virtual servo drive model to execute a motion trajectory within the range of the discrete controller models to capture the amplitude, radius and frequency of the joints, to respectively form samples corresponding to joint commands and their joint responses; the servo diagnostic model is constructed by inputting the samples of the joints into a one-dimensional convolutional neural network to output a vector, thereby presenting the possibility that each joint axis is the axis with the worst control performance.

6. The method for intelligent servo tuning and diagnostics as described in claim 5 to improve the dynamic trajectory accuracy of a robotic arm, wherein, The discrete model Kp(z) of the position loop controller and the discrete model Kv(z) of the speed loop controller are defined as shown in Equation 3: Equation 3; where is the proportional gain of the position loop, is the proportional gain of the speed loop, is the integral time constant, and is the discrete time control period.

7. The method for intelligent servo tuning and diagnostics as described in claim 6 to improve the dynamic trajectory accuracy of a robotic arm, wherein, The speed command filter Hω(z) and the current command filter HC(z) are defined as shown in Equation 4: Equation 4; where Mω and MC are both ≥1, and are the orders of the filters respectively.

8. The method for intelligent servo tuning and diagnostics as described in claim 7 to improve the dynamic trajectory accuracy of a robotic arm, wherein, The saturation limiting functions of the speed command filter and the current command filter are defined as shown in Equations 5 and 6 below: Equation 5; Equation 6; where, , , and are the upper and lower bounds of speed and current, respectively.

9. A method for improving the dynamic trajectory accuracy of a robotic arm using intelligent servo tuning and diagnostics as described in any one of claims 1 to 4, wherein, The trajectory of this motion is either a sinusoidal trajectory or a circular trajectory.

10. A method for improving the dynamic trajectory accuracy of a robotic arm using intelligent servo tuning and diagnostics as described in any one of claims 1 to 4, wherein, The robotic arm is a joint with six links.

Citation Information

Patent Citations

  • Robot fixed-point operation trajectory tracking optimization method

    CN114102598A

  • Mobile two-degree-of-freedom mechanical arm control method and system based on geometric PID (Proportion Integration Differentiation) control

    CN118123831A

  • Flexible joint mechanical arm tracking control method and device based on adaptive neural network

    CN120116212A