A Control Method for Industrial Robots Based on Command Filtering Compensation

Abstract

This invention discloses an industrial robot control method based on command filtering compensation, belonging to the field of robot control. The method includes: first, using a fuzzy logic system to estimate uncertain disturbances or unmodeled functions existing in the industrial robot's dynamic system; then, designing a Levant command filter between each subsystem; next, designing a corresponding filtering compensation system to address the filtering errors generated by the Levant command filter; and finally, constructing virtual control signals and actual control signals for driving the end joint movement using a dynamic surface framework. This invention, on the one hand, uses a filtering compensation system to reduce filtering errors, which further improves the tracking accuracy of traditional dynamic surface control algorithms; on the other hand, it uses a minimum parameter estimation mechanism to update the weight parameters in the fuzzy logic system, significantly reducing the computational load required for training the weight parameters, thereby improving the real-time performance of the control algorithm.

Description

Technical Field

[0001] This invention belongs to the field of robot control, and particularly relates to an industrial robot control method based on command filtering compensation. Background Technology

[0002] In the field of automated operation trajectory control for industrial robots, traditional dynamic surface control algorithms face significant technical bottlenecks when dealing with complex working conditions: the inherent high nonlinearity of the industrial robot's dynamic system and the lack of modeled dynamics can easily lead to insufficient trajectory tracking accuracy, especially in high-precision assembly and flexible processing scenarios, which can easily cause cumulative errors; at the same time, existing fuzzy approximation algorithms require online updates of all membership function parameters, resulting in an exponential increase in computational complexity, which severely restricts the real-time response capability of the control system.

[0003] While existing industrial robot control systems overcome the "differential explosion" problem by introducing a dynamic surface framework, the lack of modeling of the coupling effect between dynamics and filter errors still reduces system robustness. Furthermore, the lack of an effective mechanism to improve estimation efficiency during parameter estimation makes it difficult to balance control accuracy and computational efficiency. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention provides an industrial robot control method based on command filtering compensation, comprising:

[0005] Construct a self-learning fuzzy logic system based on uncertain disturbances or unmodeled functions existing in the dynamic system of industrial robots;

[0006] Construct a Levant command filter that includes a fourth-order filtering error compensation system;

[0007] The self-learning fuzzy logic system and the Levant command filter are introduced into the dynamic surface control framework to generate actual torque control signals for driving the motion of the end joint.

[0008] The robot arm is controlled based on the actual torque control signal of the drive terminal joint movement.

[0009] Optionally, the process of constructing the self-learning fuzzy logic system includes:

[0010] The sub-dynamic systems of the industrial robot dynamics model are organized to separate the uncertain parts that are difficult to model;

[0011] Multiple fuzzy logic systems are used to approximate and model the uncertain parts respectively;

[0012] Design a low-load adaptive update law model, and update the optimal weight parameters of the self-learning fuzzy logic system in real time based on the low-load adaptive update law model.

[0013] Optionally, the process of designing the low-load adaptive update law model includes:

[0014] Construct the coupling relationship between the estimated variables and the weight parameters;

[0015] The member function information of multiple self-learning fuzzy logic systems is introduced into the coupling relationship;

[0016] Construct a self-learning fuzzy logic system that approximates and models the uncertain components in each sub-dynamic.

[0017] Optionally, the process of constructing the Levant command filter including a fourth-order filtering error compensation system includes:

[0018] Three Levant command filters with two adjustment parameters are designed between the sub-dynamic systems of the industrial robot.

[0019] A fourth-order filtering error compensation system is constructed to address the filtering errors generated by the Levant command filter with the two adjustment parameters.

[0020] Optionally, the expression for the Levant command filter is:

[0021] ;

[0022] in, , Both represent the state variables of the filter. and They represent and The derivative of This indicates the virtual control signals that need to be designed subsequently. and All of these represent positive design parameters. Represents a symbolic function.

[0023] Optionally, the expression for the fourth-order filter compensation system is:

[0024] ;

[0025] in, , , , , , All of these represent positive design parameters. and Both represent the state variables of the filter compensation system; Represents state variables The derivative of Represents state variables The derivative of This indicates compensation for tracking errors.

[0026] Optionally, the process of generating the actual torque control signal for driving the end joint movement includes:

[0027] Design virtual control signals for low-order subsystems that do not contain actual control signals;

[0028] Based on the virtual control signal, the derivative of the previous virtual control signal is calculated using the Levant command filter, and the previous virtual control signal is used in the design process of the next virtual control signal.

[0029] Based on the virtual control signal and the state variables of the Levant command filter, the actual output torque control signal is generated.

[0030] Optionally, the process of generating the actual torque control signal for driving the end joint movement further includes:

[0031] Virtual control signals were designed for the first three sub-dynamic systems of the flexible manipulator.

[0032] Design the actual control signals for the last-stage subsystem;

[0033] The virtual control signal and the actual control signal are combined with the Levant command filter and the fourth-order filtering error compensation system to form a complete control signal chain.

[0034] On the other hand, the present invention also provides an electronic device including a memory, a processor, and a computing program stored in the memory and executable on the processor, wherein the processor implements the method when executing the computing program.

[0035] On the other hand, the present invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method.

[0036] Compared with the prior art, the present invention has the following advantages and technical effects:

[0037] This invention introduces a Levant command filter between each subsystem, which preprocesses the virtual control signals of the preceding subsystems to generate high-precision reference signals for input to subsequent subsystems. At the same time, the constructed filtering compensation system can dynamically compensate for residual filtering errors. This dual closed-loop correction mechanism effectively suppresses the impact of filtering errors on the tracking accuracy of the controller.

[0038] This invention innovatively proposes a fuzzy approximation mechanism based on the minimum weight parameter update criterion. This mechanism reduces the dimensionality of the traditional online adjustment of all parameters by reconstructing the parameter update rules of the fuzzy logic system. It adaptively updates only the maximum L2 norm of the weight parameters that dominate the dynamic characteristics of the control system, thus significantly reducing the amount of online computation. Attached Figure Description

[0039] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments and descriptions of this application are used to explain this application and do not constitute an undue limitation of this application. In the drawings:

[0040] Figure 1 This is a schematic diagram of the method flow according to an embodiment of the present invention;

[0041] Figure 2 This is a comparison diagram of angular displacement and reference trajectory in an embodiment of the present invention;

[0042] Figure 3 This is a comparison chart of tracking errors in an embodiment of the present invention;

[0043] Figure 4 This is a graph showing the changes in the estimated weight parameters according to an embodiment of the present invention.

[0044] Figure 5 This is a graph showing the changes in joint angular velocity and drive motor angular velocity according to an embodiment of the present invention;

[0045] Figure 6 This is a diagram showing the angular displacement variation of the drive motor according to an embodiment of the present invention. Detailed Implementation

[0046] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0047] It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0048] Example 1

[0049] like Figure 1 As shown, this embodiment provides an industrial robot control method based on command filtering compensation, including:

[0050] Construct a self-learning fuzzy logic system based on uncertain disturbances or unmodeled functions existing in the dynamic system of industrial robots;

[0051] A Levant command filter is constructed based on a fourth-order filtering error compensation system;

[0052] The actual torque control signal that drives the end joint motion is generated based on the self-learning fuzzy logic system and the Levant command filter.

[0053] The robot arm is controlled based on the actual torque control signal of the drive terminal joint movement.

[0054] Specifically:

[0055] Step 1: Construct a self-learning fuzzy logic system and use it to estimate uncertain disturbances or unmodeled functions in the dynamic system of an industrial robot.

[0056] Step 2: Design a Levant command filter and its filtering error compensation system;

[0057] Step 3: Introduce the self-learning fuzzy logic system and Levant command filter into the dynamic surface control framework to design an actual torque control signal that can be used to drive the motion of the end joint.

[0058] The specific method for constructing a self-learning fuzzy logic system and using it to estimate uncertain disturbances or unmodeled functions in the dynamics system of an industrial robot is as follows:

[0059] First, it is necessary to determine which data items in the dynamic system are difficult to model precisely, and thus these should be the targets that the fuzzy logic system needs to estimate. For example, consider the following type of dynamic model for an industrial robot:

[0060]

[0061] in, , These represent the angular displacement and angular velocity generated by the rotation of a link around its own joint, respectively. , This indicates the angular displacement and angular velocity generated by the drive motor installed at the joint during operation; This indicates the inherent torque transmission flexibility parameter of the connection components between the drive motor and the rotary joint; express The derivative of ; This represents the random, uncertain noise introduced when the sensor reports angular displacement. This refers to the random, uncertain noise generated by the sensor feedback on the angular displacement and angular velocity produced by the drive motor installed at the joint during operation; Indicates the mass of the connecting rod; Represents gravitational acceleration; Indicates the length of the link; This represents a frictional force model at a rotary joint, which is related to the rotational angular velocity of the joint; This represents the torque that needs to be provided for each joint, which can be considered as the control signal that needs to be designed. This represents the natural damping parameter of the drive motor; This represents the moment of inertia of the entire joint. This indicates the output signal.

[0062] On the one hand, the sensors inevitably introduce random and uncertain noise when feeding back information such as joint angles and motor output angles; on the other hand, the frictional force at the joint is also affected by many factors such as load and lubrication, making it difficult to obtain an accurate mathematical model of the frictional force. Therefore, it is necessary to modify the existing manipulator dynamics model. Further refinement is needed to distinguish between the deterministic and uncertain components. The deterministic component will serve as usable information in the controller design; the self-learning fuzzy logic system will perform approximate estimation of the uncertain component. Thus, the dynamic model... It is further organized into the form of the following formula (2):

[0063]

[0064] in, , Each of these represents an uncertain component in each sub-dynamic system that is difficult to model precisely. Referring to the theoretical dynamic model formula provided in this embodiment, it can be... It can be represented in the following form:

[0065]

[0066] However, in practical control environments, it is often impossible to obtain results using mathematical formulas. Precise expression Therefore, a self-learning fuzzy logic system is proposed to approximate these uncertain components. Then, these uncertain components... Nature is regarded as the approximation object of self-learning fuzzy logic systems.

[0067] Next, to approximate the uncertain part Taking this as an example, we will further illustrate the design process of a self-learning fuzzy logic system. First, it is necessary to construct a system with... A fuzzy logic system with fuzzy rules:

[0068]

[0069] in Represents the input signal vector. This is the output signal; ,and It is an important component of fuzzy sets. Please note the fuzzy logic system formula. The input signal is the uncertain part. input variables Therefore, it is necessary to experimentally determine which input variables are related to the uncertainties in each sub-dynamic system. Similarly, designing a fuzzy logic system for other uncertainties only requires replacing the input information. And determine the total number of fuzzy rules That's it. Next, the output signal of the fuzzy logic system can be... Further expressed in the following form:

[0070]

[0071] in, Represents the optimal weight parameter vector. Let represent the activation function vector. Where, the _i_th Activation function at each node It can be represented as:

[0072]

[0073] in, express The corresponding member functions, This represents the k-th input variable, where k=1 is... k=2 is k=3 is , This represents the total number of fuzzy rules. Indicates the cumulative multiplication symbol;

[0074] Following the design steps described above, the mathematical models of the other three fuzzy logic systems can be represented in the following form:

[0075]

[0076] in, express The corresponding output signal, express The corresponding fuzzy logic system, express The corresponding output signal, express The corresponding fuzzy logic system, express The corresponding output signal, express The corresponding fuzzy logic system.

[0077] It can be observed that the four fuzzy logic systems mentioned above all correspond to different optimal weight parameters. , , , However, traditional weighting parameter schemes involve designing adaptive parameter update laws to update the weight parameter values ​​in each fuzzy logic system, which inevitably leads to a huge computational load. Therefore, this embodiment, from the perspective of reducing computational load, further constructs a single estimation variable. The coupling relationship expression with the weight parameters in all fuzzy logic systems:

[0078]

[0079] in, Represents the optimal weight parameters The 2-norm.

[0080] That is, the estimated variables An equivalent estimate is made of the maximum value of the L2 norm of all weight parameters, such as... As shown. Next, this example will estimate the variables. An online self-learning update law was designed:

[0081]

[0082] in, , , Both represent the input signal vector. For the definition, please refer to the formula. Below; This indicates compensation for tracking error; , , All represent positive design parameters; express The derivative of . Therefore, the dynamic model Each uncertain component in the model can be approximated and modeled using the following self-learning fuzzy logic system:

[0083]

[0084] The specific method for designing a Levant command filter and its fourth-order filtering error compensation system is as follows:

[0085] The classic design process for dynamic surface controllers involves designing virtual or actual control signals for each subsystem, and then using state variables of filters to connect these control signals together to form the entire control system. However, the first-order subsystem has a definite control objective, namely, to make the output signal... Real-time tracking reference trajectory .then, The target can be manually set according to the actual work task, and therefore can be considered a known quantity. However, no clear tracking target can be found for the second to the last subsystem. Therefore, the solution proposed in this embodiment is to use the state variables of the Levant command filter as the tracking target to achieve cascaded control. Thus, a Levant command filter needs to be designed for each of the first to second, second to third, and third to fourth order subsystems:

[0086]

[0087] in , Both represent the state variables of the filter; and They represent and Their respective derivatives; This indicates the virtual control signals that need to be designed subsequently; and All represent positive design parameters. In the Levant command filter model... The sign function is defined in detail as follows:

[0088]

[0089] State variables in Levant command filters It can be regarded as the first The reference signal that each subsystem needs to track. When stable, it can be regarded as The estimated value of the derivative. Will be the first This is an important component of the virtual control signals in the subsystem. This design eliminates the need for... The complex differentiation operation solves the "differential explosion" problem in backstep control.

[0090] The approach of approximating the virtual control signal and its derivatives using the state variables of the Levant command filter inevitably introduces approximation errors. Therefore, to mitigate the impact of filtering errors on the tracking accuracy of the closed-loop system, this embodiment further designs the following type of fourth-order filtering compensation system:

[0091]

[0092] in , , , , , All of these represent positive design parameters. and Both represent the state variables of the filter compensation system; Represents state variables The derivative of Represents state variables The derivative of .

[0093] The method of introducing a self-learning fuzzy logic system and a Levant command filter into a dynamic surface control framework to design an actual torque control signal that can be used to drive the motion of the end joint is as follows;

[0094] The main function of the control system in industrial robots is to enable the terminal to track the specified reference trajectory signal. The dynamic surface controller design framework refers to designing virtual control signals for the first three sub-dynamic systems of a flexible manipulator. , and Design the actual control signals for the last-stage subsystem. These virtual control signals function in two ways: first, to ensure the stability of their respective sub-dynamic systems, for example... The main purpose is to ensure the first-order sub-dynamic system The stability of the filter is the firstly, and the secondly, it is the formula for the Levant command filter. Formula for filtering compensation system The input signals are provided. Next, the specific design process of these three virtual control signals and actual control signals will be introduced.

[0095] To quantize the output signal With reference trajectory signal The difference between them can be defined as the following tracking error:

[0096]

[0097] The filtering compensation system designed in this embodiment can further reduce tracking errors. Therefore, in order to quantitatively evaluate the effect of the state variables of the filtering compensation system on reducing tracking errors, we can focus on the first-order sub-dynamic system of the industrial robot (…). The following definition is used to compensate for tracking error. :

[0098]

[0099] in, These are the state variables of the filtering compensation system.

[0100] Next, it is necessary to compensate for the tracking error. and self-learning fuzzy logic systems Used for virtual control signals In the design process. Therefore, this embodiment will Designed in the following form:

[0101]

[0102] in , , , and All of these represent design parameters that are greater than zero; This represents the state variables corresponding to a robust compensation system; Represents the reference trajectory signal The derivative of . Introducing these two items The purpose is also to further improve the controller's anti-interference performance.

[0103] To ensure the state variables in the second-order dynamic subsystem of the industrial robot are stable The boundedness of the tracking error can be defined as follows:

[0104]

[0105] That is, let the virtual control signal Driving state variables Tracking the state variables corresponding to the bounded Levant command filter Of course, the state variable corresponding to the second subsystem in the fourth-order filter compensation system... Also used to reduce tracking error Therefore, for the second-order sub-dynamic system of industrial robots ( The following is a definition for compensating for tracking error:

[0106]

[0107] Next, it is necessary to compensate for the tracking error. and self-learning fuzzy logic systems Used for virtual control signals In the design process. Therefore, this embodiment will Designed in the following form:

[0108]

[0109] in , , and All of these represent design parameters that are greater than zero. Please note that... The state variables of the Levant command filter It is considered a tracking target. Therefore, it is first... Robust compensation item of The derivative of the state variable filtered by the Levant command Replace, and then introduce new design parameters. and Thus, virtual control signals are obtained. Robust compensation terms in .

[0110] To ensure the state variables in the third-order dynamic subsystem of the industrial robot are stable The boundedness of the tracking error can be defined as follows:

[0111]

[0112] That is, let the virtual control signal Driving state variables Tracking the state variables corresponding to the bounded Levant command filter Of course, the state variable corresponding to the third subsystem in the fourth-order filter compensation system... Also used to reduce tracking error Therefore, for the third-order sub-dynamic system of industrial robots ( The following is a definition for compensating for tracking error:

[0113]

[0114] Next, it is necessary to compensate for the tracking error. and self-learning fuzzy logic systems Used for virtual control signals In the design process. Therefore, this embodiment will Designed in the following form:

[0115]

[0116] in , , and All of these represent design parameters that are greater than zero.

[0117] Since the actual torque control signal has already appeared in the fourth-order dynamic subsystem. Therefore, it is possible to Further design is needed for the change function. The state variable corresponding to the Levant command filter... To provide a reference trajectory signal for the fourth-order dynamic subsystem, the tracking error can therefore be defined as follows:

[0118]

[0119] Of course, the state variable corresponding to the fourth subsystem in the fourth-order filter compensation system Also used to reduce tracking error Therefore, for the fourth-order sub-dynamic system of industrial robots ( The following is a definition for compensating for tracking error:

[0120]

[0121] Next, it is necessary to compensate for the tracking error. and self-learning fuzzy logic systems Used for actual control signals In the design process. Therefore, this embodiment will Designed in the following form:

[0122]

[0123] in , , and All of these represent design parameters that are greater than zero. Additionally, state variables need to be provided in real-time at each adoption time. The specific value is determined, therefore, the following robust update law is designed in this embodiment:

[0124] =

[0125] in, and All of these represent design parameters that are greater than zero. State variables The derivative of .

[0126] In this embodiment, the actual control signal with a filtering compensation system is input into the constructed industrial robot simulation model. Figure 2 This demonstrates that the robot's end effector can rapidly track a specified reference trajectory signal even with an initial angular displacement difference of 0.1 rad. To further illustrate the positive impact of the filtering compensation system on the tracking accuracy of the control algorithm, relevant simulation results of the traditional dynamic surface control algorithm are presented, such as... Figure 3 As shown. Traditional dynamic surface control algorithms lack a filtering compensation system, therefore, their filtering errors cannot be effectively suppressed, resulting in a decrease in tracking accuracy. In this embodiment, the tracking error of the controller has been reduced to within 0.005865 after the control system enters a stable period, as... Figure 3 As shown by the solid line in the middle. Figure 3 The solid line at the midpoint illustrates the tracking error of the traditional dynamic surface control algorithm, whose amplitude can only be controlled within the range of 0.02026, thus reducing the tracking accuracy by approximately [missing value]. . Figure 4 The adaptive parameter estimates are given. The real-time change curve shows that it has already shown regular changes after 8 seconds. Figure 5 The curves showing the changes in joint angular displacement and drive motor angular displacement are displayed separately. The drive motor has a relatively high initial speed to rotate the joint. Finally... Figure 6 The curve of the angular displacement of the drive motor as a function of time t is given.

[0127] On the other hand, this embodiment also provides an electronic device, including a memory, a processor, and a computing program stored in the memory and executable on the processor, wherein the processor implements the method when executing the computing program.

[0128] On the other hand, this embodiment also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method.

[0129] The above are merely preferred embodiments 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.

Claims

1. A control method for an industrial robot based on command filtering compensation, characterized in that, include: A self-learning fuzzy logic system is constructed based on the uncertain disturbances or unmodeled functions existing in the dynamic system of industrial manipulators. Construct a Levant command filter that includes a fourth-order filtering error compensation system; The self-learning fuzzy logic system and the Levant command filter are introduced into the dynamic surface control framework to generate actual torque control signals for driving the motion of the end joint. The robot arm is controlled based on the actual torque control signal of the drive terminal joint movement; The process of constructing the Levant command filter, which includes a fourth-order filtering error compensation system, includes: Three Levant command filters with two adjustment parameters are designed between the sub-dynamic systems of the industrial robot. A fourth-order filtering error compensation system is constructed to address the filtering errors generated by the Levant command filter with the two adjustment parameters. The expression for the Levant command filter is: ； in, , Both represent the state variables of the filter. and They represent and The derivative, This indicates the virtual control signals that need to be designed subsequently. and All of these represent positive design parameters. Represents a symbolic function; The expression for the fourth-order filtering error compensation system is: ； in, , , , , , All of these represent positive design parameters. and Both represent the state variables of the filtering compensation system; Represents state variables The derivative, Represents state variables The derivative, This indicates compensation for tracking errors.

2. The method according to claim 1, characterized in that, The process of constructing a self-learning fuzzy logic system includes: The sub-dynamic systems of the industrial robot dynamics model are organized to separate the uncertain parts that are difficult to model; Multiple fuzzy logic systems are used to approximate and model the uncertain parts respectively; Design a low-load adaptive update law model, and update the optimal weight parameters of the self-learning fuzzy logic system in real time based on the low-load adaptive update law model.

3. The method according to claim 2, characterized in that, The process of designing the low-load adaptive update law model includes: Construct the coupling relationship between the estimated variables and the weight parameters; The member function information of multiple self-learning fuzzy logic systems is introduced into the coupling relationship; Construct a self-learning fuzzy logic system that approximates and models the uncertain components in each sub-dynamic.

4. The method according to claim 1, characterized in that, The process of generating the actual torque control signal for driving the end joint movement includes: Design virtual control signals for low-order subsystems that do not contain actual control signals; Based on the virtual control signal, the derivative of the previous virtual control signal is calculated using the Levant command filter, and the previous virtual control signal is used in the design process of the next virtual control signal. Based on the virtual control signal and the state variables of the Levant command filter, the actual output torque control signal is generated.

5. The method according to claim 1, characterized in that, The process of generating the actual torque control signal for driving the end joint movement also includes: Virtual control signals were designed for the first three sub-dynamic systems of the flexible manipulator. Design the actual control signals for the last-stage subsystem; The virtual control signal and the actual control signal are combined with the Levant command filter and the fourth-order filtering error compensation system to form a complete control signal chain.

6. An electronic device comprising a memory, a processor, and a computing program stored in the memory and executable on the processor, characterized in that, When the processor executes the computing program, it implements the method of any one of claims 1-5.

7. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the method of any one of claims 1-5.

Images