Robot joint motor dual-port two-degree-of-freedom fuzzy adaptive internal model control method

Abstract

The invention belongs to the technical field of robots, and particularly relates to a robot joint motor dual-port two-degree-of-freedom fuzzy adaptive internal model control method based on inertia identification. Comprising the following steps: step 1, designing a robot joint motor rotating speed ring dual-port two-degree-of-freedom internal model controller; 2, designing an improved recursive least square rotational inertia identification method with a forgetting factor; and 3, designing a fuzzy adaptive law to realize dual-port two-degree-of-freedom fuzzy adaptive internal model control. Double-port two-degree-of-freedom internal model control is adopted for the rotating speed ring, the rotating speed falling amplitude after sudden load adding is reduced, and the return time after load disturbance is shortened. The invention provides an identification method capable of accurately identifying the rotational inertia under the condition that the rotating speed of the motor is steady, and automatic setting of controller parameters is realized by combining a fuzzy adaptive law.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of robots, and particularly relates to a robot joint motor double-port two-degree-of-freedom fuzzy adaptive internal model control method. BACKGROUND

[0002] To realize high-precision and high-dynamic force control of an industrial robot, the bottom-layer performance optimization of a driving execution unit must be fundamentally emphasized. As a power core, the performance of a speed regulation system of a robot servo joint motor is the cornerstone of force control ability. Specifically, in the speed regulation system of the robot joint motor, the performance of a controller directly affects the dynamic response, anti-disturbance ability and steady-state precision of the system. At present, a widely used PI controller can quickly restore the steady-state speed when a sudden load disturbance occurs, but its inherent characteristics cause a significant speed drop of the system. In comparison, a standard internal model control strategy performs excellently in inhibiting the influence of a load disturbance on speed and can significantly improve the anti-load disturbance performance of the system, but it takes a long time to restore to the steady state after a sudden load occurs, which is difficult to meet the demand of high dynamic response occasions. Therefore, in scenes with high requirements for processing precision and efficiency, designing a speed controller that has excellent anti-load disturbance ability and fast disturbance recovery ability has become a key challenge to improve the performance of the robot joint motor driving system. In addition, during the operation of the motor, its physical parameters will change due to load changes or mechanical structure characteristics, resulting in the deterioration of the performance of the controller based on fixed parameter tuning. Therefore, developing an adaptive control strategy that can automatically adjust the parameters of the controller in real time based on the changes of key physical parameters can not only simplify the parameter tuning process and significantly reduce the debugging workload, but also effectively improve the robustness and control quality of the system under variable parameter conditions. The purpose of optimizing the force control precision of the industrial robot from the level of the robot servo joint motor is achieved. SUMMARY

[0003] To solve the above problems, the application provides a robot joint motor double-port two-degree-of-freedom fuzzy adaptive internal model control method based on inertia identification, which adopts double-port two-degree-of-freedom internal model control for the speed loop to reduce the speed drop amplitude after a sudden load occurs and shorten the recovery time after a load disturbance. An identification method that can accurately identify the moment of inertia in the case of stable motor speed is proposed, and the automatic tuning of the controller parameters is realized in combination with a fuzzy adaptive law.

[0004] The technical scheme of the application is explained in combination with the drawings as follows:

[0005] The application provides a robot joint motor double-port two-degree-of-freedom fuzzy adaptive internal model control method based on inertia identification, which comprises the following steps:

[0006] Step 1: Design a double-port two-degree-of-freedom internal model controller for the speed loop of the robot joint motor.

[0007] Step two, design improved recursive least square inertia identification method with forgetting factor;

[0008] Step three, design fuzzy adaptive law to realize two-port two-degree-of-freedom fuzzy adaptive internal model control.

[0009] Further, the specific method of step one is as follows:

[0010] 11) Let and be designed as two-degree-of-freedom internal model controller based on internal model control principle, as follows:

[0011]

[0012]

[0013] In the formula, is a low-pass filter; is a low-pass filter time constant; is a low-pass filter; is a low-pass filter time constant; is the internal model of the motor, ,

[0014] wherein, , , is the moment of inertia, is the damping coefficient;

[0015] 12) Let the feedback control term be designed as a proportional term, as follows:

[0016]

[0017] In the formula, is a proportional term.

[0018] Further, the specific method of step two is as follows:

[0019] 21) Let the estimation have been made times of observation, and the data of the previous times are defined as: is a vector composed of previous observation output; is a matrix composed of previous times of observation input; is a vector composed of estimation values obtained by times of observation; according to the definition, the following expression is obtained:

[0020]

[0021]

[0022] Proceeding The second measurement is written as:

[0023]

[0024]

[0025]

[0026] From the above formula:

[0027]

[0028] 22) Definition The second observation:

[0029]

[0030]

[0031] 23) Introduce the forgetting factor, and the recursive least squares recursive formula with forgetting factor is derived as follows:

[0032]

[0033]

[0034]

[0035] In the formula, is the parameter vector to be estimated; is the error covariance matrix; is the input regression vector; is the system output; is the gain vector; is the forgetting factor;

[0036] 24) The motor motion equation is discretized by backward difference method, and the discretized equation is:

[0037]

[0038] In the formula, is the motor speed; is the sampling period; is the load torque; is the load torque; is the moment of inertia; is the damping coefficient;

[0039] 25) Order:

[0040]

[0041] Summarized as follows:

[0042]

[0043] Substituting into the recursive formula, we get:

[0044]

[0045] Therefore, from the result The moment of inertia is obtained from the identification and estimation. ;

[0046] 26) Introduce a dynamic triggering mechanism, calculate the difference between the feedback speed and the desired speed, and set the speed change amount. ;in, This represents the change in rotational speed of the robot's joint motors. This is the motor feedback speed value. This represents the desired speed of the motor.

[0047] Based on the change in velocity The size dynamically triggers different recognition strategies: when Higher than the set threshold When the input stimulus is considered sufficient, recursive least squares method with forgetting factor is executed for inertia identification; when Below the set threshold When the input stimulus is deemed insufficient, an external stimulus mechanism is triggered, and the identified moment of inertia at the current moment is output. As the final estimated inertia value under steady-state conditions; when the load changes abruptly or the moment of inertia changes, the algorithm is re-initialized, all terms of the identification algorithm are set to zero, and the identification steps are re-executed; at the same time, during operation, a forgetting factor dynamic adjustment strategy is introduced to improve the identification accuracy and adapt to different operating conditions.

[0048] Set a dynamic adjustment strategy:

[0049] In the formula, The initial forgetting factor; This is the adjustment coefficient;

[0050] Adjust the threshold based on error feedback This enables the improved identification algorithm to accurately identify the moment of inertia when the motor is in steady-state operation.

[0051] Furthermore, the specific method for step three is as follows:

[0052] 31) Adjusting the parameters of the internal model controller by identified inertia ; wherein, is the torque coefficient;

[0053] 32) Introducing the fuzzy reasoning idea, establishing a relationship expression between and , assuming that the initial inertia of the system is , then the inertia ratio is defined as , the inertia ratio change rate is , a single-variable two-dimensional fuzzy controller is used, taking and as input, as output, so the expression of fuzzy reasoning is defined as: ; wherein, is the set parameter; is the parameter adjustment amount; is the weight factor.

[0054] The beneficial effects of the present application are:

[0055] 1) The double-port two-degree-of-freedom internal model controller structure designed in the present application effectively separates the tracking performance of the system to the set value command and the suppression performance to the load disturbance. When the load is suddenly added, the speed drop amplitude can be significantly reduced, greatly improving the anti-load disturbance ability of the system. After the load disturbance, the time required for the speed to recover to steady state can be significantly shortened, realizing the unified optimization of anti-interference and fast recovery.

[0056] 2) The improved recursive least squares method with forgetting factor proposed in the present application effectively solves the problem of significant distortion of identification accuracy at steady speed by introducing a trigger condition and setting a threshold value, ensuring that even during steady operation of the motor, effective inertia identification accuracy can be maintained.

[0057] 3) The present application realizes the automation and intelligent setting of the controller parameters through online inertia identification and fuzzy adaptive parameter adjustment, greatly simplifying the debugging workload of engineering implementation and later maintenance, and reducing the technical threshold of control system deployment and application. BRIEF DESCRIPTION OF DRAWINGS

[0058] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present application, and therefore should not be considered as limiting the scope. For those skilled in the art, other related drawings can also be obtained without creative labor on the basis of these drawings.

[0059] Figure 1Structure block diagram of control system of the present application;

[0060] Figure 2 Structure block diagram of double-port two-degree-of-freedom internal model control;

[0061] Figure 3 Fuzzy rule chart;

[0062] Figure 4 Robot joint motor pair drag experiment platform;

[0063] Figure 5 Rotational speed contrast image;

[0064] Figure 6 Rotational inertia identification image;

[0065] Figure 7 Rotational speed contrast image with fuzzy adaptive law. DETAILED DESCRIPTION

[0066] The present application will be further described below in conjunction with the accompanying drawings and embodiments. It can be understood that the specific embodiments described herein are only used to explain the present application, but not to limit the present application. In addition, it should be noted that, for the convenience of description, only the parts related to the present application are shown in the drawings, but not all the structures.

[0067] Embodiment one

[0068] Referring to Figure 1 , the embodiment provides a robot joint motor double-port two-degree-of-freedom fuzzy adaptive internal model control method based on inertia identification, comprising the following steps:

[0069] Step one, design a robot joint motor rotational speed ring double-port two-degree-of-freedom internal model controller, and the specific method is as follows:

[0070] 11) design and as a two-degree-of-freedom internal model controller based on the internal model control principle, as shown below:

[0071]

[0072]

[0073] In the formula, is a low-pass filter; is a low-pass filter time constant; is a low-pass filter; is a low-pass filter time constant; is a motor internal model, ,

[0074] wherein, , , is the moment of inertia, is the damping coefficient;

[0075] 12) the feedback control term is designed as a proportional term as follows:

[0076]

[0077] wherein, is the proportional term.

[0078] After introducing the feedback control term , the output response can be affected by properly adjusting to ensure that the steady state can be quickly restored in the case of sudden load. After introducing the two-degree-of-freedom internal model controller , the following can be completely independent of the adjustment of the system's following performance and disturbance rejection performance. For the following performance, the parameters of the internal model controller can be adjusted to optimize the response speed of the system. And for the steady-state performance, the parameters of the controller can be independently adjusted to maintain the stable output of the system under disturbance.

[0079] Step two, design an improved recursive least square moment of inertia identification method with forgetting factor, the specific method is as follows:

[0080] 21) The improved recursive least square moment of inertia identification method with forgetting factor proposed in this paper uses recursive calculation to correct the parameters to be identified according to the last parameter estimation value and the latest observation data. Suppose that observation estimation has been carried out, the previous data is defined as: is the vector composed of the previous observation output; is the matrix composed of the previous observation input; is the vector composed of the estimation value obtained by observation; According to this definition, the following expression is obtained:

[0081]

[0082]

[0083] When measurement is carried out, it is written as:

[0084]

[0085]

[0086]

[0087] From the above equation, we get:

[0088]

[0089] 22) Definition Subsequent observations:

[0090]

[0091]

[0092] 23) Introduce the forgetting factor, and derive the recursive least squares method with forgetting factor recursive formula as follows:

[0093]

[0094]

[0095]

[0096] where, is the parameter vector to be estimated; is the error covariance matrix; is the input regression vector; is the system output; is the gain vector; is the forgetting factor, the range of old data is affected by it;

[0097] 24) For the motor motion equation, backward difference method is used for discretization, and the discretized equation is:

[0098]

[0099] where, is the motor speed; is the sampling period; is the load torque; is the load torque; is the moment of inertia; is the damping coefficient;

[0100] 25) Because , for the above difference equation,

[0101]

[0102] After arrangement, we get:

[0103]

[0104] Substitute the recursive formula:

[0105]

[0106] Therefore, from the results The identified moment of inertia ;

[0107] 26) Introduce a dynamic triggering mechanism, which is the difference between the feedback speed and the expected speed, and set the speed change ; Where, is the speed change of the robot joint motor, is the motor feedback speed value, is the motor expected speed value;

[0108] According to the size of the speed change , different identification strategies are triggered dynamically: when is higher than the set threshold , it is considered that the input excitation is sufficient, and the recursive least squares method with forgetting factor is executed to identify the inertia; when is lower than the set threshold , it is considered that the input excitation is insufficient, and the external excitation mechanism is triggered, and the identified moment of inertia at the current time is output as the final estimated inertia value in the steady state to prevent parameter update stagnation. When the load changes or the moment of inertia changes, the algorithm is reinitialized, and the identification algorithm is set to zero, and the identification step is executed again to avoid the negative impact of historical data on the subsequent identification accuracy. At the same time, in the running, in order to prevent the forgetting factor from being too large, causing the identification to converge too quickly and affecting the estimation accuracy, a forgetting factor dynamic adjustment strategy is introduced to improve the identification accuracy and adapt to different operating conditions.

[0109] Set the dynamic adjustment strategy:

[0110] Where, is the initial forgetting factor; is the adjustment coefficient;

[0111] The threshold is adjusted through error feedback, so that the improved identification algorithm can identify the accurate moment of inertia in the steady state of the motor operation.

[0112] Referring to Figure 2 , step three, design a fuzzy adaptive law to realize a two-port two-degree-of-freedom fuzzy adaptive internal model control, the specific method is as follows:

[0113] 31) the identified inertia is obtained after completing inertia identification The parameters of the internal model controller can be adjusted by the identified inertia wherein, is the torque coefficient;

[0114] 32) Due to the existence of various disturbances in actual working conditions, the relationship between and is no longer a simple linear relationship in theory. Therefore, the fuzzy reasoning idea is introduced, and some prior experimental tests are carried out, so as to establish a practical relationship expression between and . Assuming that the initial inertia of the system is , the inertia ratio can be defined, and the inertia ratio change rate is . A single-variable two-dimensional fuzzy controller is used, taking and as inputs, and as output. Therefore, the expression of fuzzy reasoning can be defined as: wherein: is the set parameter, and is the parameter adjustment amount, and is the weight factor.

[0115] As shown in Figure 3 , the range of the tested inertia ratio is defined as (1, 16), and the inertia ratio change rate is defined as (0, 4). As shown in Table 1, the fuzzy set of is taken as {FX ZX ZZ ZD FB}, and the fuzzy set of is taken as {LBBX BD}.

[0116] Table 1

[0117] FX ZX ZZ ZD FB LB BB SZ ZJ ZJ XZ BX BB SZ ZJ XZ XZ BD SZ ZJ ZJ XZ GD

[0118] Example Two

[0119] Experiments are carried out on a permanent magnet synchronous pair-dragging experimental platform as shown in Figure 4 , wherein the motor parameters are: the number of pole pairs ; the stator inductance , ; the stator resistance ; the rotational inertia ; the damping coefficient ; the reference speed is set to ; a load is suddenly added at . The speed comparison image is as shown in Figure 5As shown in the figure, it can be seen that the double-port two-degree-of-freedom inner model controller has good effect on disturbance. The image of moment of inertia identification is as shown in Figure 6 The external inertia is changed to five times of the original inertia, i.e. , and the moment of inertia identification result is good. The image of speed comparison with and without fuzzy adaptive law is as shown in Figure 7 As can be seen, without fuzzy adaptive law, when the external inertia changes greatly, the speed control will produce great overshoot and fluctuation, and with fuzzy adaptive law, the control effect is excellent, verifying that the application is effective.

[0120] Although the embodiments of the present application have been shown and described, it can be understood by those skilled in the art that various changes, modifications, replacements and variations can be made to the embodiments without departing from the principles and spirits of the present application, and the scope of the present application is defined by the appended claims and their equivalents.

Claims

1. A two-port, two-degree-of-freedom fuzzy adaptive internal model control method for robot joint motors, characterized in that, Includes the following steps: Step 1: Design a two-port, two-DOF internal model controller for the robot joint motor speed loop; Step 2: Design an improved recursive least squares rotational inertia identification method with a forgetting factor; Step 3: Design a fuzzy adaptive law to implement dual-port two-degree-of-freedom fuzzy adaptive internal model control.

2. The method for dual-port two-degree-of-freedom fuzzy adaptive internal model control of a robot joint motor according to claim 1, characterized in that, The specific method for step one is as follows: 11) and The design is a two-degree-of-freedom internal model controller based on the internal model control principle, as shown below: ； ； In the formula, It is a low-pass filter; For low-pass filter Time constant; It is a low-pass filter; For low-pass filter Time constant; For the internal model of the motor, , in, , , For rotational inertia, The damping coefficient; 12) Feedback control items It is designed as a proportional term, as shown below: ； In the formula, This is a proportional term.

3. The method for dual-port two-degree-of-freedom fuzzy adaptive internal model control of a robot joint motor according to claim 1, characterized in that, The specific method for step two is as follows: 21) Assume that it has been carried out The first observation estimate, the previous The data for this time is defined as: This is a vector composed of the output values ​​from the previous observations; For the previous A matrix composed of the input quantities from each observation; for A vector composed of estimates obtained from several observations; based on this definition, the following expression is derived: ； ； conduct For this measurement, it was written as: ； ； ； From the above formula, we get: ； 22) Definition Second observation: ； ； 23) Introducing a forgetting factor, the recursive formula for the recursive least squares method with the forgetting factor is derived as follows: ； ； ； In the formula, The vector of parameters to be estimated; Here is the error covariance matrix; The input regression vector; For system output; It is the gain vector; Forgetting factor; 24) The motor motion equations are discretized using the backward difference method. The discretized equations are as follows: ； In the formula, This refers to the motor speed; The sampling period; This is the load torque; This is the load torque; It is the moment of inertia; The damping coefficient; 25) Order: ； Summarized as follows: ； Substituting into the recursive formula, we get: ； Therefore, from the result The moment of inertia is obtained from the identification and estimation. ; 26) Introduce a dynamic triggering mechanism, calculate the difference between the feedback speed and the desired speed, and set the speed change amount. ;in, This represents the change in rotational speed of the robot's joint motors. This is the motor feedback speed value. This represents the desired speed of the motor. Based on the change in velocity The size dynamically triggers different recognition strategies: when Higher than the set threshold When the input stimulus is considered sufficient, recursive least squares method with forgetting factor is executed for inertia identification; when Below the set threshold When the input stimulus is deemed insufficient, an external stimulus mechanism is triggered, and the identified moment of inertia at the current moment is output. As the final estimated inertia value under steady-state conditions; when the load changes abruptly or the moment of inertia changes, the algorithm is re-initialized, all terms of the identification algorithm are set to zero, and the identification steps are re-executed; at the same time, during operation, a forgetting factor dynamic adjustment strategy is introduced to improve the identification accuracy and adapt to different operating conditions. Set a dynamic adjustment strategy: ; In the formula, The initial forgetting factor; This is the adjustment coefficient; Adjust the threshold based on error feedback This enables the improved identification algorithm to accurately identify the moment of inertia when the motor is in steady-state operation.

4. The method for dual-port two-degree-of-freedom fuzzy adaptive internal model control of a robot joint motor according to claim 1, characterized in that, The specific method for step three is as follows: 31) Adjusting the parameters of the internal model controller by identifying the inertia. ;in, This is the torque coefficient; 32) Introduce the concept of fuzzy reasoning to establish a... and The relationship expression, assuming the initial inertia of the system is... Then the inertia ratio is defined. The rate of change of inertia ratio A single-variable two-dimensional fuzzy controller is used to... and As input, As the output, the expression for fuzzy inference is defined as follows: ;in, These are the adjusted parameters; For parameter adjustment amount; This is the weighting factor.

Images