A feedforward compensation device and a feedforward coefficient self-tuning method for an unknown servo system

Abstract

The application discloses a feedforward compensation device and a feedforward coefficient self-tuning method of an unknown servo system, and belongs to the technical field of motor servo control. The device comprises a position signal differential module, a torque feedforward module, a speed feedforward module, a PID controller, a model identification module and a feedforward coefficient calculation module. The method comprises the following steps: judging the inertia type of the unknown servo system; identifying model parameters; and tuning and calculating torque feedforward coefficients and speed feedforward coefficients. The application can realize the judgment of the load inertia type of the servo system and the accurate identification of model parameters in a very short time, realize high-order accurate approximation of the controlled object model, consider the system delay factor, obtain accurate feedforward compensation signals by using high-order approximation, and solve the problems of long feedforward tuning time, poor compatibility of the load inertia model of the system, and insufficient feedforward compensation precision.

Description

TECHNICAL FIELD

[0001] The application relates to a feedforward compensation device and a feedforward coefficient self-tuning method for an unknown servo system and belongs to the technical field of motor servo control. BACKGROUND

[0002] In a motor position servo system, in order to enable the system output to quickly and accurately track a given signal, feedforward signal compensation is often needed to eliminate reference trajectory tracking error and improve system response speed and accuracy.

[0003] At present, most servo feedforward compensation mainly adopts speed and acceleration signal feedforward, which is essentially a second-order approximation of the controlled object model. However, this low-order approximation strategy has obvious limitations when facing complex systems, especially when dealing with double-inertia load models in flexible transmission systems. Specifically, the second-order approximation method ignores the adverse effects of resonance on low-frequency tracking performance and the signal lag problem caused by system delay, which severely limits the corresponding feedforward compensation effect and makes it difficult to achieve the expected precise tracking control goal.

[0004] To overcome this limitation, theoretically, higher-order differential signals can be considered for feedforward compensation to more accurately approximate the dynamic characteristics of the controlled object. However, high-order differential feedforward devices mean an increase in feedforward coefficients, which not only increases the complexity of system design but also greatly increases the difficulty of parameter tuning. Especially when the model parameters of the servo system are unknown and need to be identified, the parameter tuning process of the high-order feedforward link will become particularly tedious and time-consuming. Not only does it prolong the overall feedforward link tuning time, but it can also reduce the control performance of the system due to improper parameter tuning, thereby adversely affecting the realization of fast and accurate feedforward compensation tracking control of the servo system.

[0005] Therefore, how to effectively simplify the design and implementation of high-order feedforward devices while ensuring the accuracy of feedforward compensation, and reduce the difficulty of parameter tuning, has become a problem to be solved in the current motor position servo system feedforward compensation technology. SUMMARY

[0006] To solve the problems in the background art, the application provides a feedforward compensation device and a feedforward coefficient self-tuning method for an unknown servo system.

[0007] To achieve the above-mentioned purpose, the application adopts the following technical solution: a feedforward compensation device for an unknown servo system, comprising

[0008] a position signal differential module for obtaining high-order differential signals of a given position trajectory, including first-order to fifth-order differential signals;

[0009] A torque feedforward module is configured to generate a torque feedforward signal according to second-order to fifth-order differential signals of the given position trajectory, and input the torque feedforward signal into a current loop of the PID controller;

[0010] A speed feedforward module is configured to generate a speed feedforward signal according to first-order to fourth-order differential signals of the given position trajectory, and input the speed feedforward signal into a speed loop of the PID controller;

[0011] A PID controller is configured to control the servo motor according to the torque feedforward signal, the speed feedforward signal, and a position closed-loop given signal generated according to the given position trajectory;

[0012] A model identification module is configured to acquire a frequency response characteristic curve of a controlled object in a speed loop of the servo system and model parameters of the servo system;

[0013] A feedforward coefficient calculation module is configured to acquire each feedforward coefficient according to the model parameters of the servo system.

[0014] The method comprises the following steps:

[0015] S1: inertia type judgment of the unknown servo system;

[0016] S101: a frequency response characteristic curve of a controlled object in a speed loop of the servo system, i.e., a load characteristic curve, is acquired by means of frequency sweep injection;

[0017] The frequency sweep injection signal is a Chirp current signal with a rated amplitude and a total time length of 0.2 s, the frequency sweep injection signal is injected into a q-axis current given value, and a frequency sweep output signal is a motor terminal speed signal of the system.

[0018] The frequency range of the frequency sweep injection signal and the frequency sweep output signal is 0-1250 Hz.

[0019] S102: the original load characteristic curve acquired is smoothed by means of a Savitzky-Golay filtering method;

[0020] S103: an ideal transfer function of a motor load model of a double-inertia load servo system is defined as follows:

[0021]

[0022] In formula (1):

[0023] J = J l + J m is the total rotational inertia of the motor transmission system;

[0024] J l is the load rotational inertia;

[0025] J m J is the moment of inertia of the motor;

[0026] K t K is the torque coefficient of the motor drive system;

[0027] K s K is the torsional elastic coefficient of the motor drive system;

[0028] s is the Laplace operator of the transfer function;

[0029] S104: According to the obtained load characteristic curve, the existence of a pair of resonance frequency points in the detection motor load model is automatically detected through the extreme value search of the amplitude-frequency characteristic curve and the phase mutation of the phase-frequency characteristic curve, and the load inertia type is judged according to the detection result, that is, a single-inertia load servo system or a double-inertia load servo system;

[0030] If a pair of resonance frequency points exist, record the system anti-resonance frequency and the system resonance frequency Otherwise, go to S2.

[0031] S2: Model parameter identification;

[0032] S201: According to the least square fitting of the low frequency band data in the amplitude-frequency characteristic curve in the obtained load characteristic curve, the equivalent system total moment of inertia J / K t is obtained.

[0033] S202: According to the least square fitting of the low frequency band data in the phase-frequency characteristic curve in the obtained load characteristic curve, the equivalent system delay T s of the speed loop is obtained.

[0034] S203: According to the obtained load characteristic curve and the written speed loop controller parameters, the speed loop closed loop frequency response characteristic curve and the position loop controlled object frequency response characteristic curve are calculated and converted:

[0035]

[0036] In formula (2):

[0037] G p-plant (s) is the position loop controlled object transfer function;

[0038] G s (s) is the speed loop closed loop transfer function;

[0039] G c (s) is the speed loop controller transfer function;

[0040] G s-plant (s) is the speed loop controlled object transfer function;

[0041] S204: When dealing with the outer loop position loop parameter setting problem, the inner loop speed loop closed loop characteristic is regarded as a whole time delay link, so the position loop controlled object transfer function expression is converted to Therefore, the least square fitting of the low frequency band data in the phase frequency characteristic curve in the position loop controlled object frequency response characteristic curve is obtained p .

[0042] S3: Torque feedforward coefficient setting calculation;

[0043] S301: Define the transfer function from the torque feedforward signal compensation to the load end angle output:

[0044]

[0045] In formula (3):

[0046] G ω (s) is the system transfer function from the motor end speed to the load end speed;

[0047] S302: First order inertia approximation equivalent is carried out on the system time delay link of the speed loop, and the final transfer function is obtained:

[0048]

[0049] S303: The servo system is subjected to feedforward control, and the system transfer function in the torque feedforward module is:

[0050]

[0051] S304: Correspondingly, each order torque feedforward coefficient is obtained.

[0052] S30401: According to the obtained equivalent system total moment of inertia J / K t Determine the torque feedforward second order differential coefficient

[0053] S30402: According to the obtained equivalent system total moment of inertia J / K t And the equivalent system time delay T s of the speed loop, the torque feedforward third order differential coefficient is determined

[0054] S30403: According to the obtained equivalent system total moment of inertia J / K t And the system resonance frequency f N of the resonance frequency, the torque feedforward fourth order differential coefficient is determined

[0055] S30404: According to the equivalent system total moment of inertia J / K obtained t , resonance frequency system resonance frequency f N And speed loop equivalent system delay T s Determine the torque feedforward five-order differential coefficient

[0056]

[0057] S30405: If the load characteristic curve obtained when judging the load inertia type does not detect the resonance frequency point, that is, it is judged as a single inertia load, K a4 = K a5 = 0.

[0058] S4: Tuning calculation of speed feedforward coefficient.

[0059] S401: Define the transfer function of the speed feedforward signal compensation to the load end angle output:

[0060]

[0061] S402: First-order inertia approximation equivalent is carried out on the equivalent system delay link of the position loop, and the final transfer function is obtained:

[0062]

[0063] S403: The system transfer function in the speed feedforward module is:

[0064]

[0065] S404: Correspondingly, each order speed feedforward coefficient is obtained.

[0066] S40401: Take the speed feedforward first-order differential coefficient K v1 = 1;

[0067] S40402: According to the obtained position loop equivalent system delay T p Determine the speed feedforward second-order differential coefficient K v2 = T p ;

[0068] S40403: According to the obtained resonance frequency system anti-resonance frequency f A Determine the speed feedforward third-order differential coefficient

[0069] S40404: According to the obtained resonance frequency system anti-resonance frequency f A And the position loop equivalent system delay T p Determine the speed feedforward fourth-order differential coefficient

[0070] S40405: if the load characteristic curve obtained when judging the load inertia type does not detect a resonance frequency point, i.e. it is judged as a single-inertia load, K v3 = K v4 = 0.

[0071] Compared with the prior art, the present application has the following beneficial effects:

[0072] The present application can realize the judgment of the load inertia type of a servo system and the accurate identification of model parameters in a very short time, is compatible with various load inertia models, and thus realizes the high-order accurate approximation of the model of the controlled object, considers the system delay factor when setting the feedforward coefficient, and uses high-order approximation to obtain an accurate feedforward compensation signal. In addition, the present application provides setting formulas for the coefficients of the torque feedforward and the speed feedforward, avoids complicated manual debugging, greatly shortens the setting time of the feedforward link, and completes the self-setting of the feedforward link parameters of an unknown servo system in 0.2 seconds, effectively solving the problems of long feedforward setting time, poor compatibility of the load inertia model of the system, and insufficient feedforward compensation accuracy. BRIEF DESCRIPTION OF DRAWINGS

[0073] Figure 1 is a structural schematic diagram of the feedforward compensation device of the unknown servo system of the present application;

[0074] Figure 2 is a schematic diagram of frequency sweep injection;

[0075] Figure 3 is a schematic diagram of the original frequency response characteristic curve and the smoothed curve after filtering processing obtained by frequency sweep;

[0076] Figure 4 is a schematic diagram of the identification of the speed loop model information;

[0077] Figure 5 is a schematic diagram of the fitting of the position loop equivalent system delay;

[0078] Figure 6 is a flow chart of the feedforward coefficient self-setting method of the feedforward compensation device of the unknown servo system of the present application;

[0079] Figure 7 is a schematic diagram of the comparison results of the feedforward effect. DETAILED DESCRIPTION

[0080] Clearly, the described embodiments are only a part of the embodiments of the application, and not all the embodiments. Based on the embodiments in the application, all other embodiments obtained by those of ordinary skill in the art without creative labor fall within the protection scope of the application.

[0081] A feedforward compensation device of an unknown servo system comprises

[0082] A position signal differential module is configured to obtain high-order differential signals of a given position trajectory, including first-order to fifth-order differential signals.

[0083] A torque feedforward module is configured to generate a torque feedforward signal according to the second-order to fifth-order differential signals of the given position trajectory, and input the torque feedforward signal into a current loop of a PID controller.

[0084] A speed feedforward module is configured to generate a speed feedforward signal according to the first-order to fourth-order differential signals of the given position trajectory, and input the speed feedforward signal into a speed loop of the PID controller.

[0085] The PID controller is configured to control a servo motor according to the torque feedforward signal, the speed feedforward signal, and a position closed-loop given signal generated by the given position trajectory.

[0086] A model identification module is configured to obtain a frequency response characteristic curve of a controlled object in a speed loop of the servo system and model parameters of the servo system.

[0087] A feedforward coefficient calculation module is configured to obtain each feedforward coefficient according to the model parameters of the servo system.

[0088] A feedforward coefficient self-tuning method of a feedforward compensation device of an unknown servo system provided by the application comprises the following steps:

[0089] S1: Inertia type judgment of the unknown servo system.

[0090] S101: Obtain a frequency response characteristic curve of a controlled object in a speed loop of the servo system, i.e., a load characteristic curve, by means of frequency sweep injection.

[0091] The frequency sweep injection signal in S101 is a Chirp current signal with a rated amplitude and a total time length of 0.2s. The frequency sweep injection signal is injected into a q-axis current given value. A frequency sweep output signal is a motor end speed signal. The overall frequency sweep injection schematic diagram is shown in Figure 2 .

[0092] For the convenience of subsequent fitting system model parameters, the frequency range of the sweep injection signal and the sweep output signal is 0-1250Hz. For a permanent magnet synchronous motor double inertia load servo system with unknown model parameters, the original frequency response characteristic curve obtained by sweep and its smoothed curve after filtering are shown in Figure 3

[0093] S102: Smooth the obtained original load characteristic curve by using Savitzky-Golay filtering method to eliminate its sawtooth characteristics;

[0094] S103: Define the ideal transfer function of the motor load model of the double inertia load servo system:

[0095]

[0096] In formula (1):

[0097] J = J l + J m is the total rotational inertia of the motor transmission system;

[0098] J l is the load rotational inertia;

[0099] J m is the motor rotational inertia;

[0100] K t is the torque coefficient of the motor transmission system;

[0101] K s is the torsional stiffness coefficient of the motor transmission system;

[0102] s is the Laplace operator of the transfer function;

[0103] S104: According to the obtained load characteristic curve, the existence of paired resonance frequency points in the motor load model is automatically detected by the extreme value search of the amplitude-frequency characteristic curve and the phase mutation of the phase-frequency characteristic curve, and the load inertia type is judged according to the detection result, that is, single inertia load servo system or double inertia load servo system;

[0104] If there are paired resonance frequency points, record the system anti-resonance frequency (small value) and the system resonance frequency (large value) Otherwise, S2 is performed.

[0105] S2: Fast and accurate model parameter identification;

[0106] S201: According to the least square fitting of the low frequency band data in the amplitude-frequency characteristic curve in the obtained load characteristic curve, the equivalent system total rotational inertia J / K​t .

[0107] S202: Obtain the speed loop equivalent system delay T according to the least square fitting of the low frequency band data in the phase frequency characteristic curve in the obtained load characteristic curve s ;

[0108] S203: Calculate the speed loop closed loop frequency response characteristic curve and the position loop controlled object frequency response characteristic curve (the difference between the two is an integral element) according to the obtained load characteristic curve and the written speed loop controller parameters:

[0109]

[0110] In formula (2):

[0111] G p-plant (s) is the position loop controlled object transfer function;

[0112] G s (s) is the speed loop closed loop transfer function;

[0113] G c (s) is the speed loop controller transfer function;

[0114] G s-plant (s) is the speed loop controlled object transfer function;

[0115] S204: When processing the position loop parameter setting problem of the outer loop, the inner loop speed loop closed loop characteristic is regarded as a whole delay element, so the position loop controlled object transfer function expression is converted to Therefore, the position loop equivalent system delay T is obtained according to the least square fitting of the low frequency band data in the phase frequency characteristic curve in the obtained position loop controlled object frequency response characteristic curve p .

[0116] S3: Torque feedforward coefficient setting calculation;

[0117] S301: For a permanent magnet synchronous motor double inertia load servo system, the transfer function of the torque feedforward signal compensation to the load end angle output is defined as:

[0118]

[0119] In formula (3):

[0120] G ω (s) is the system transfer function from the motor end speed to the load end speed;

[0121] S302: The system delay element of the speed loop is approximated to a first order inertia equivalent, and the final transfer function is obtained:

[0122]

[0123] S303: Feedforward control is performed on the servo system, which essentially compensates for the inverse model from the feedforward signal injection point to the system output. The system transfer function in the torque feedforward module is:

[0124]

[0125] S304: Corresponding torque feedforward coefficients of each order are obtained respectively.

[0126] S30401: According to the obtained equivalent system total moment of inertia J / K t Determine the torque feedforward second-order differential coefficient

[0127] S30402: According to the obtained equivalent system total moment of inertia J / K t And the speed loop equivalent system delay T s Determine the torque feedforward third-order differential coefficient

[0128] S30403: According to the obtained equivalent system total moment of inertia J / K t And the resonant frequency of the system resonant frequency f N Determine the torque feedforward fourth-order differential coefficient

[0129] S30404: According to the obtained equivalent system total moment of inertia J / K t , the resonant frequency of the system resonant frequency f N And the speed loop equivalent system delay T s Determine the torque feedforward fifth-order differential coefficient

[0130]

[0131] S30405: If the load characteristic curve obtained when judging the load inertia type does not detect the resonant frequency point, that is, it is judged as a single-inertia load, then K a4 = K a5 = 0.

[0132] S4: Tuning calculation of speed feedforward coefficient.

[0133] S401: Define the transfer function from the speed feedforward signal compensation point to the load end angle output:

[0134]

[0135] S402: The equivalent system delay link of the position loop is approximated to first-order inertia, and the final transfer function is obtained:

[0136]

[0137] S403: The feedforward control is performed on the servo system, which is essentially compensating the inverse model from the feedforward signal injection to the system output, and the system transfer function in the speed feedforward module is:

[0138]

[0139] S404: Corresponding speed feedforward coefficients of each order are obtained respectively.

[0140] S40401: The speed feedforward first-order differential coefficient K v1 = 1 is taken.

[0141] S40402: The speed feedforward second-order differential coefficient K p is determined according to the obtained position loop equivalent system delay T v2 = T p .

[0142] S40403: The speed feedforward third-order differential coefficient K A is determined according to the obtained resonance frequency system anti-resonance frequency f

[0143] S40404: The speed feedforward fourth-order differential coefficient K A is determined according to the obtained resonance frequency system anti-resonance frequency f p and the position loop equivalent system delay T v3 .

[0144] S40405: If the load characteristic curve obtained when judging the load inertia type does not detect the resonance frequency point, that is, it is judged as a single-inertia load, K v3 = K v4 = 0.

[0145] Thus, the whole process of the feedforward signal compensation and coefficient self-tuning of the unknown permanent magnet synchronous motor servo system is completed, and the flowchart of the overall scheme is shown in Figure 6 .

[0146] Finally, the feedforward compensation and coefficient self-tuning method disclosed in the application is compared with the traditional second-order method for experimental verification, and the results are shown in Figure 7 . It can be seen that the disclosed method has smaller tracking error, higher tracking accuracy and no steady-state tracking error under the same position tracking trajectory command. At the same time, the overall method does not require additional feedforward coefficient tuning time and tuning running trajectory in the implementation process except for the system model identification process, and has the significant advantages of fast speed and short running trajectory in the overall process. The feedforward signal compensation and feedforward coefficient self-tuning of the unknown servo system can be realized quickly and accurately within 0.2s.

[0147] It will be apparent to those skilled in the art that the application is not limited to the details of the above-exemplified embodiments and that the present application can be implemented in other forms without departing from the spirit or essential characteristics thereof. The embodiments should therefore be considered in all respects as illustrative and not restrictive, the scope of the application being indicated by the appended claims rather than by the above description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein. No reference signs in the claims should be considered as limiting the scope of the claims to the identity of the reference signs therein.

[0148] Furthermore, it should be understood that although the description is made on embodiments, not every embodiment contains only one independent technical solution, and the description of the specification is only for the sake of clarity, and those skilled in the art should consider the specification as a whole, and the technical solutions in each embodiment can also be appropriately combined to form other embodiments that those skilled in the art can understand.

Claims

1. A method for self-tuning the feedforward coefficient of a feedforward compensation device of an unknown servo system, characterized by: The method comprises the following steps: S1: Determine the inertia type of the unknown servo system; The S1 comprises the following steps: S101: obtaining a frequency response characteristic curve of a controlled object of a speed loop of a servo system, that is, a load characteristic curve, by means of a swept frequency injection method; S102: Smoothing the obtained original load characteristic curve using a Savitzky-Golay filtering method; S103: Define the ideal transfer function of the motor load model for the dual-inertia load servo system: （1） In formula (1): is the total moment of inertia of the motor drive system; is the load moment of inertia; is the motor moment of inertia; is the torque coefficient of the motor drive system; is the torsional elastic coefficient of the motor transmission system; s is the Laplace operator of the transfer function; S104: Based on the acquired load characteristic curve, automatic detection is performed to detect whether there are paired resonant frequency points in the motor load model by performing extreme value search of the amplitude-frequency characteristic curve and phase mutation verification of the phase-frequency characteristic curve. The load inertia type is determined based on the detection result, i.e., a single-inertia load servo system or a dual-inertia load servo system. If there are paired resonant frequency points, record the system anti-resonant frequency and the system resonant frequency ; Otherwise, proceed to S2; S2: Model parameter identification; The S2 comprises the following steps: S201: Obtain the total moment of inertia of the equivalent system in J / K based on the least squares fitting of the low-frequency data in the amplitude-frequency characteristic curve of the acquired load characteristic curve t ; S202: Obtain the speed loop equivalent system delay T according to the least square fitting of the low-frequency data in the phase-frequency characteristic curve of the acquired load characteristic curve s ; S203: Calculate and convert the speed loop closed-loop frequency response characteristic curve and the position loop controlled object frequency response characteristic curve based on the acquired load characteristic curve and the written speed loop controller parameters: （2） In formula (2): G p-plant (s) is the transfer function of the position loop controlled object; G s (s) is the closed-loop transfer function of the speed loop; G c (s) is the transfer function of the speed loop controller; G s-plant (s) is the transfer function of the speed loop controlled object; S204: When dealing with the outer position loop parameter tuning problem, the inner speed loop closed loop characteristic is regarded as the overall delay link, so the position loop controlled object transfer function expression is converted to Therefore, the position loop equivalent system delay T is obtained based on the least square fitting of the low-frequency data in the phase-frequency characteristic curve of the obtained frequency response characteristic curve of the position loop controlled object. p ; S3: Torque feedforward coefficient setting calculation; The S3 comprises the following steps: S301: Define the transfer function from the torque feedforward signal compensation to the load end angle output: （3） In formula (3): is the system transfer function from the motor end speed to the load end speed; S302: Perform first-order inertia approximation equivalence on the system delay link of the speed loop to obtain the final transfer function: （4） S303: Perform feedforward control on the servo system. The system transfer function in the torque feedforward module is: （5） S304: Obtaining torque feedforward coefficients of various orders respectively; The S304 includes the following steps: S30401: Based on the obtained total moment of inertia of the equivalent system J / K t Determine the second-order differential coefficient of torque feedforward ; S30402: Based on the obtained total moment of inertia of the equivalent system J / K t And the speed loop equivalent system delay T s Determine the third-order differential coefficient of torque feedforward ; S30403: Based on the obtained total moment of inertia of the equivalent system J / K t and resonant frequency system resonant frequency f N Determine the fourth-order differential coefficient of torque feedforward ; S30404: Based on the obtained total moment of inertia of the equivalent system J / K t , resonant frequency system resonant frequency f N And the speed loop equivalent system delay T s Determine the fifth-order differential coefficient of torque feedforward ; S30405: If the load characteristic curve obtained when judging the load inertia type does not detect the resonant frequency point, that is, it is judged to be a single inertia load, then ; S4: Calculation of speed feedforward coefficient; The S4 comprises the following steps: S401: Define the transfer function from the speed feedforward signal compensation to the load end angle output: （6） S402: Perform first-order inertia approximation on the delay link of the position loop equivalent system to obtain the final transfer function: （7） S403: Perform feedforward control on the servo system. The system transfer function in the speed feedforward module is: S404: Obtaining corresponding speed feedforward coefficients of each order respectively; The S404 includes the following steps: S40401: Get the first-order differential coefficient of speed feedforward ; S40402: Based on the acquired position loop equivalent system delay T p Determine the second-order differential coefficient of the velocity feedforward ; S40403: Based on the obtained resonant frequency system anti-resonant frequency f A Determine the third-order differential coefficient of velocity feedforward ; S40404: Based on the obtained resonant frequency system anti-resonant frequency f A and the position loop equivalent system delay T p Determine the fourth-order differential coefficient of velocity feedforward ; S40405: If the load characteristic curve obtained when judging the load inertia type does not detect the resonant frequency point, that is, it is judged to be a single inertia load, then .

2. The method according to claim 1, wherein: The sweep frequency injection signal in S101 is a Chirp current signal of rated amplitude with a total duration of 0.2s. The sweep frequency injection signal is injected into the q-axis current setting point, and the sweep frequency output signal is the system motor end speed signal.

3. The method according to claim 2, wherein: The frequency range of the swept frequency injection signal and the swept frequency output signal is 0-1250 Hz.

4. A feedforward compensation device for an unknown servo system, used to implement the feedforward coefficient self-tuning method of the feedforward compensation device for an unknown servo system according to any one of claims 1 to 3, characterized in that: include Position signal differential module, used to obtain high-order differential signals of a given position trajectory, including first-order to fifth-order differential signals; A torque feedforward module is used to generate a torque feedforward signal based on the second-order to fifth-order differential signals of a given position trajectory, and input the torque feedforward signal into the current loop of the PID controller; A velocity feedforward module is used to generate a velocity feedforward signal based on the first to fourth order differential signals of a given position trajectory, and input the velocity feedforward signal into the velocity loop of the PID controller; PID controller, used to control the servo motor based on the torque feedforward signal, velocity feedforward signal and position closed-loop given signal generated by the given position trajectory; Model identification module, used to obtain the frequency response characteristic curve of the servo system speed loop controlled object and the servo system model parameters; The feedforward coefficient calculation module is used to obtain various feedforward coefficients according to the servo system model parameters.

Images