Proxy model construction method based on extended state observer

Abstract

The invention belongs to the field of agent models for fast decision support, and particularly discloses an agent model construction method based on an extended state observer, which comprises the following steps of: based on understanding of a modeled process, firstly determining input information and output information of the modeled process, and determining a relative order according to a physical principle; based on the obtained input information, output information and relative order, constructing a state space dynamic system equation of the modeled process to obtain an output equation; based on the state space dynamic system equation of the current relative order, designing an expansion state observer of the next order of the current relative order; obtaining an adjustable parameter based on the next-order extended state observer of the current relative order; an original system dynamic equation is rewritten into an integrator series type equation formula by introducing adjustable parameters; and directly calculating the output information of the next moment according to the input information of the current moment and the relative order of the system.

Description

Technical Field

[0001] The present invention relates to the field of agent models for rapid decision support, and in particular to a method for constructing an agent model based on an extended state observer. Background Art

[0002] Building agent models that can predict one step ahead. After mastering the current state information, such models can quickly predict future state changes or action effects. They are core tools in many fields such as robot control, game AI strategy formulation, and fault warning. Its core value lies in providing fast decision support, optimizing control strategies, precise planning, and forward-looking fault prevention. The key to realizing such models lies in the integration of advanced machine learning technologies, including but not limited to reinforcement learning algorithms, various prediction models (random forests, support vector machines, deep learning, etc.), and time series analysis strategies. These technologies give models the ability to predict the future through in-depth analysis and learning of historical data. However, in some special applications, there is not enough historical data for model training, or it is necessary to consider simplifying complex models or choosing methods with higher computational efficiency.

[0003] Surrogate models based on linear equations are models that approximate the behavior of a system or predict future states through linear relationships. These models use linear equations for modeling and decision-making, and usually have low computational complexity and good interpretability. For example: linear regression, linear dynamic systems, etc. In general, surrogate models based on linear equations are suitable for situations where the relationship between features and target variables is simple and computing resources are limited. They can serve as benchmarks or preliminary models for more complex models and play an important role in many practical applications.

[0004] The advantages of the surrogate model based on linear equations are low computational cost, strong interpretability and simple implementation, which are applicable to many basic tasks. However, it has shown great limitations in use for two reasons. First, the expressive power is limited. Linear equations cannot capture complex nonlinear relationships and perform poorly for nonlinear problems. Second, the assumption is simplified, assuming that there is a linear relationship between features and target variables, which is often not true in practical applications. Summary of the invention

[0005] The purpose of the present invention is to provide a method for constructing a proxy model based on an extended state observer to solve the problem that the existing linear proxy model cannot capture complex nonlinear relationships.

[0006] To achieve the above object, the present invention provides the following technical solution: a method for constructing a proxy model based on an extended state observer, comprising the following steps:

[0007] Step S1: Based on the understanding of the modeled process, first determine the input information and output information of the modeled process, the modeled process includes a physical system, an economic system, a biological system, and a signal processing system;

[0008] Step S2: According to physical principles, determining the constraint conditions of the modeled process from input information to output information, wherein the constraint conditions include relative order;

[0009] Step S3: Based on the obtained input information, output information and relative order, construct the state space dynamic system equation of the modeled process, take the current relative order, output information and the previous order derivative of the current relative order as state variables, and then substitute the state variables into the state equation to obtain the output equation;

[0010] Step S4: Based on the state space dynamic system equation of the current relative order, design the next-order extended state observer of the current relative order;

[0011] Step S5: obtaining adjustable parameters based on the next-order extended state observer of the current relative order;

[0012] Step S6: rewriting the original system dynamic equation into an integrator series equation formula by introducing adjustable parameters;

[0013] Step S7: directly calculate the output information at the next moment according to the input information at the current moment and the relative order of the system.

[0014] As a preferred technical solution of the first aspect of the present invention, in step S1, the input information and output information of the modeled process are measurement information, and the measurement information is directly acquired by a measurement tool.

[0015] As a preferred technical solution of the first aspect of the present invention, a transfer function or state-space model of the system is established and analyzed based on physical principles. By observing the numerator and denominator polynomials of the transfer function, or analyzing the state variables and output equations in the state-space model, the relative order of the system can be calculated.

[0016] As a preferred technical solution of the first aspect of the present invention, in step S2, when the relative order information of the modeled process cannot be calculated and obtained, the relative order information is set to a low-order relative order, and the low-order relative order includes two and three.

[0017] As a preferred technical solution of the first aspect of the present invention, in step S3, the output equation is used to describe the factors affecting the prediction of the next output by the input information, system dynamics and system uncertainty dynamics during the modeling process;

[0018] As a preferred technical solution of the first aspect of the present invention, in step S4, the extended state observer includes a linear extended state observer and a nonlinear extended state observer, which are used to obtain the total disturbance information of the system from the measurable input and output information, and dynamically expand the uncertainty into a system state.

[0019] As a preferred technical solution of the first aspect of the present invention, in step S4, when n>3, the relative order n is high-order; when n≤3, the relative order n is low-order; when the relative order information of the modeled process is high-order, a low-order extended state observer is selected, and the low-order extended state observer includes a second-order extended state observer and a third-order extended state observer.

[0020] As a preferred technical solution of the first aspect of the present invention, in step S6, the output equation formula is updated to obtain a standard integrator series configuration, the observer is used to estimate the total disturbance of the system, the estimated value is added to the product of the system modelable dynamics and the input signal and the adjustable gain to obtain an intermediate variable; the intermediate variable is relatively integrated to obtain an estimated value of the output signal at the next moment.

[0021] In a second aspect, the present invention provides an electronic device, comprising: a processor and a memory, wherein the memory stores a computer program that can be called by the processor;

[0022] The processor executes the method for constructing an agent model based on an extended state observer described in the first aspect by calling the computer program stored in the memory.

[0023] In a third aspect, the present invention provides a computer-readable storage medium storing instructions, which, when executed on a computer, enables the computer to execute the method for constructing an agent model based on an extended state observer as described in the first aspect.

[0024] Compared with the prior art, the present invention has the following beneficial effects:

[0025] The invention is based on the control signal, the system output signal and the state space dynamic equation of the modeled process obtained by real-time measurement of the system, and obtains the estimated value of the total disturbance of the system through the designed extended state observer, and adds the estimated value to the system modelable dynamics and the product of the input signal and the adjustable gain to obtain the intermediate variable, and then performs n times of integration to obtain the estimated value of the output signal at the next moment;

[0026] The present invention can capture complex nonlinear relationships and has no strict requirements on the accuracy of the state space dynamic equations of the modeled process. If there is no modelable dynamics in the system, the modelable dynamics can be regarded as part of the system's uncertain dynamics and estimated through an extended state observer. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 It is a flow chart of the agent modeling method based on the extended state observer of the present invention;

[0028] Figure 2 It is a principle block diagram of the agent modeling method based on the extended state observer of the present invention. DETAILED DESCRIPTION

[0029] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0030] See also Figure 1-2 The present invention provides a technical solution: a method for constructing a proxy model based on an extended state observer, comprising the following steps:

[0031] Step S1: Based on the understanding of the process to be modeled, first determine the input information and output information of the process to be modeled;

[0032] It should be noted that input information refers to external factors or variables that can affect the state or behavior of the process being modeled, and the input variable information that changes over time;

[0033] Output information is the response of the modeled process to input information and internal state changes, which reflects the performance, state or behavior results of the modeled process according to time changes;

[0034] The modeled process refers to the actual process or system described, predicted or controlled by a mathematical model, which can be physical (such as mechanical systems), economic (such as market behavior), biological (such as physiological processes in organisms), technical (such as signal processing) or any other observable and measurable system; that is, the modeled process includes physical systems, economic systems, biological systems and signal processing systems; based on the understanding of the modeled process, the predictive ability and application scope of the input and output information for building an effective agent model are clarified, so as to effectively predict or control.

[0035] Specifically, the input information and output information of the modeled process are measurement information, and the measurement information is directly obtained by the measurement tool.

[0036] Step S2: According to physical principles, determining the constraint conditions of the modeled process from input information to output information, wherein the constraint conditions include relative order;

[0037] It should be noted that physical principles provide the basic rules and constraints for the behavior of the process being modeled. By understanding physical principles, a mathematical model describing the dynamics of the process being modeled can be established.

[0038] Specifically, relative order is an important parameter that describes the dynamic characteristics of a system. It reflects the response speed of the system output to changes in the input. When the relative order is determined, the transfer function or state-space model of the system is established and analyzed based on physical principles. The relative order of the system can be calculated by observing the numerator and denominator polynomials of the transfer function, or analyzing the state variables and output equations in the state-space model.

[0039] Step S3: Based on the input information, output information and relative order obtained, the state space dynamic system equation of the modeled process is constructed, the current relative order, output information and the previous order derivative of the current relative order are used as state variables, and then the state variables are substituted into the state equation to obtain the output equation; the output equation is shown in formula (1):

[0040]

[0041] Where x1(t), x2(t), …, x n (t) represents the system state variable, n represents the number of system state variables, n = 1, 2, ... N, N is an integer greater than or equal to 1, representing the total number of different variables, g(·, t) represents the known system modelable object, f(·, t) represents the uncertain dynamics of the system, and b represents the unknown gain; u(t) represents the input information of the controlled process, and y(t) represents the output information of the controlled process;

[0042] In the modeled process described by formula (1), the input information u(t), system dynamics g(·, t) and system uncertainty dynamics f(·, t) are the main factors affecting the next output y(t+1) of the system.

[0043] Since b is an unknown gain, in order to facilitate the design of the extended state observer, b0 is taken as an adjustable parameter to approximate the true value of b, so the dynamic system equation (1) is rewritten as:

[0044]

[0045] It should be noted that the state space model is a method to describe the dynamic behavior of a system. It uses a set of state variables to represent the internal state of the system and describes how these state variables change over time and how they are related to inputs and outputs through state equations and output equations.

[0046] Step S4: Based on the state space dynamic system equation of the current relative order, design the next-order extended state observer of the current relative order;

[0047] By dynamically expanding the uncertainty in formula (2) into a system state, formula (2) can be rewritten as:

[0048]

[0049] In the formula, is an unknown disturbance, then f(·,t)+(b-b0)u(t) can be estimated by the state observer.

[0050] Specifically, if the relative order is marked as n, then the next order of the current relative order is n+1;

[0051] n+1-order extended state observer:

[0052]

[0053] Where y(t) represents the output information of the controlled process, e(t) represents the estimated deviation of the output y(t), Z1(t), Z2(t), ..., Z n (t) represent the states x1(t), x2(t), …, x n The estimated values ​​of (t), β1,…,β n and β n+1 represents the system parameters of the controlled process; appropriately adjust the parameters β1,…,β n , β n+1 and b0, Z1(t) can approach x1(t), Z2(t) can approach x2(t), ..., Z n+1 (t) approaches f(·,t)+(b-b0)u(t).

[0054] Let e2(t)=Z2(t)-x2(t),…,e n+1 (t) = Z n+1 (t)-x n+1 (t), then the error system of the observer is:

[0055]

[0056] The error e(t) is determined by formula (6):

[0057]

[0058] Coefficients β1,…,β n , β n+1 It can be specially taken that the characteristic equation (7) is:

[0059] p o (s)=s (n+1) +β1s (n) +...+β n s+βn+1 (7)

[0060] The roots of the characteristic equation (7) are located on the left side of the complex plane and far enough away from the imaginary axis. Reasonable selection of the coefficients of the extended state observer can make Z n+1 Approximate the total disturbance of the controlled process f(·,t)+(b-b0)u(t);

[0061] Taking the fourth-order extended state observer as an example, for the convenience of parameter adjustment, the roots of the observer can be configured at ω o Place, that is

[0062] p o (s)=s 4 +β1s 3 +β2s 2 +β3s+β4=(s+ω o ) 4 (8)

[0063] Among them, ω o is called the bandwidth of the observer. Therefore, the four parameters of the observer can be expressed as β1 = 4ω o , In other words, properly adjusting ω o , Z1(t) can be made to approach x1(t), Z2(t) to approach x2(t), …, Z4(t) to approach f(·,t)+(b-b0)u(t).

[0064] The observer error formula (5) can be rewritten as:

[0065]

[0066] When the error system enters a steady state,

[0067]

[0068] Therefore, the steady-state error of the system is:

[0069]

[0070] It can be found that when ω o >>When k, the system steady-state error will be small enough.

[0071] It should be noted that a state observer is a dynamic system that derives estimated values ​​of state variables based on the measured values ​​of the system's external variables (input variables and output variables); the observer's error system is determined by specific coefficients, which can be specifically valued to ensure the stability of the system.

[0072] In practical applications, when the state variables of the state space model of the modeled process cannot be directly measured using sensors, a state observer is used to estimate the actual state of the system. By building a new system and using the known external input and measurable system output of the original system to complete the reconstruction of the system model, the internal state variables of the system can be estimated.

[0073] More specifically, in step S4, the extended state observer includes a linear extended state observer and a nonlinear extended state observer, which are used to obtain the total disturbance information of the system from the measurable input and output information;

[0074] It should be noted that: when n>3, the relative order n is high-order; when n≤3, the relative order n is low-order; when the relative order information of the modeled process is high-order, a low-order extended state observer is selected, and the low-order extended state observer includes a second-order extended state observer or a third-order extended state observer.

[0075] Step S5: Based on the estimated value Z of the currently obtained system uncertainty dynamics f(·,t)+(b-b0)u(t) n+1 (t), combined with the known system dynamics g(·,t) and system input information u(t), the formula is g(·,t)+Z n+1 (t)+b0u(t)=u0(t).

[0076] It should be noted that b0 is an adjustable parameter of the extended state observer, which is used to approximate the true value of b in formula (1).

[0077] Step S6: By introducing adjustable parameters, the reconstructed Information u0(t), so the original system dynamic equation (1) can be rewritten as the integrator series equation formula (12):

[0078]

[0079] Specifically, the formula (12) obtained by rewriting the original system dynamic equation (1) is a standard integrator series configuration, that is, u0(t) is the nth-order derivative of the system output information y(t+1).

[0080] Step S7: directly calculate the system's t+1th step output y(t+1) based on the input information u0(t) of the tth step and the relative order n of the system.

[0081] Specifically, an observer is used to estimate the total disturbance of the system, and the estimated value is added to the product of the system's modelable dynamics and the input signal and the adjustable gain to obtain an intermediate variable; the intermediate variable is integrated for a relative order nth time to obtain an estimated value of the output signal at the next moment, which can be used for real-time measurement and real-time adjustment of the system dynamics.

[0082] Embodiment 2: This embodiment obtains wind speed information and wind power information through real-time measurement, and combines the proxy model construction method based on the extended state observer described in Embodiment 1 to construct a wind power prediction model for the next sampling moment, including the following steps:

[0083] Step S1: The input information u(t) of this process is wind speed information, and the system output information y(t) is wind power information;

[0084] Step S2: taking the control process of wind speed as the main input factor to determine the wind power output as the research object, assuming that the relative order of the control process is 2;

[0085] Step S3: Construct the state space dynamic system equation of the controlled process as:

[0086]

[0087] Among them, x1(t) and y(t) are the system output information, i.e., wind power information, x2(t) is the wind power change rate, f(·,t) is the uncertainty dynamics of the system, b is the unknown gain, and b0 is an adjustable parameter;

[0088] Step S4: Design a third-order linear extended state observer as:

[0089]

[0090] β1, β2, β3 follow β1=3ω o , Choose parameters and choose ω reasonably o and b0, Z1(t) can be approximated to y(t), and Z3(t) can be approximated to the total disturbance of the system f(·,t)+(b-b0)u(t);

[0091] Alternatively, a third-order nonlinear extended state observer can be designed as:

[0092]

[0093] By properly choosing β1, β2, β3, δ, and b0, Z1(t) can approach y(t), and Z3(t) can approach the total disturbance of the system f(·,t)+(b-b0)u(t);

[0094] Step S5: Let Z3(t)+b0u(t)=u0(t) (16)

[0095] Step S6: The original dynamic system equation (13) is rewritten as:

[0096]

[0097] Step S7: According to the u0(t) information obtained from formula (17) and the relative order n=2, the next step y(t+1) information can be obtained by integrating u0(t) twice, which is the predicted information of wind power at the next moment.

[0098] Based on the control signal u(t) obtained by real-time measurement of the system, the system output signal y(t) and the state space dynamic equation of the modeled process, the estimated value Z of the total disturbance f(·,t)+(b-b0)u(t) of the system is obtained through the designed extended state observer. n+1 , add the estimated value to the system modelable dynamic g(·, t) and the product of the input signal u(t) and the adjustable parameter to obtain the intermediate variable u0(t), and then integrate u0(t) n times to obtain the estimated value y(t+1) of the output signal at the next moment; the present invention has no strict requirements on the accuracy of the state space dynamic equation of the modeled process. If the system does not have modelable dynamic g(·, t), g(·, t) can be regarded as a part of the system uncertainty dynamic f(·, t) and estimated by the extended state observer; the present invention can capture complex nonlinear relationships.

[0099] Example 3

[0100] An electronic device according to an exemplary embodiment includes: a processor and a memory, wherein the memory stores a computer program that can be called by the processor;

[0101] The processor executes the above-mentioned method for constructing an agent model based on an extended state observer by calling the computer program stored in the memory.

[0102] The electronic device may have relatively large differences due to different configurations or performances, and may include one or more processors (Central Processing Units, CPU) and one or more memories, wherein the memory stores at least one computer program, and the at least one computer program is loaded and executed by the processor to implement the agent model construction method based on the extended state observer provided by the above-mentioned various method embodiments. The electronic device may also include other components for realizing the functions of the device, for example, the electronic device may also have components such as a wired or wireless network interface and an input and output interface for input and output. The embodiments of this application are not described in detail here.

[0103] Example 4

[0104] According to an exemplary embodiment, a computer-readable storage medium is shown, on which a rewritable computer program is stored;

[0105] When the computer program is executed on a computer device, the computer device is enabled to execute the above-mentioned method for constructing an agent model based on an extended state observer.

[0106] In an exemplary embodiment, a computer-readable storage medium is also provided, such as a memory including at least one computer program, wherein the at least one computer program is executable by a processor to implement a method for constructing an agent model based on an extended state observer in the above embodiment. For example, the computer-readable storage medium can be a read-only memory (ROM), a random access memory (RAM), a compact disc read-only memory (CD-ROM), a magnetic tape, a floppy disk, an optical data storage device, etc.

[0107] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0108] It should be understood that determining B based on A does not mean determining B only based on A. B can also be determined based on A and / or other information.

[0109] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented by software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website site, computer, server or data center to another website site, computer, server or data center by wired (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state hard disk.

[0110] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0111] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0112] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of the units is only one, and there may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0113] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0114] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0115] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in the various embodiments of the present application. The aforementioned storage media include: various media that can store program codes, such as USB flash drives, mobile hard disks, read-only memories (ROM), random access memories (RAM), magnetic disks or optical disks.

[0116] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

[0117] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for constructing a proxy model based on an extended state observer, characterized in that: The following steps are involved: Step S1: Based on the understanding of the modeled process, first determine the input information and output information of the modeled process, the modeled process includes a physical system, an economic system, a biological system, and a signal processing system; Step S2: According to physical principles, determining the constraint conditions of the modeled process from input information to output information, wherein the constraint conditions include relative order; Step S3: Based on the input information, output information and relative order obtained, a state space dynamic system equation of the modeled process is constructed, the current relative order, output information and the previous order derivative of the current relative order are used as state variables, and the state variables are substituted into the state equation to obtain the output equation; Step S4: Based on the state space dynamic system equation of the current relative order, design the next-order extended state observer of the current relative order; Step S5: obtaining adjustable parameters based on the next-order extended state observer of the current relative order; Step S6: rewriting the original system dynamic equation into an integrator series equation formula by introducing adjustable parameters; Step S7: directly calculate the output information at the next moment according to the input information at the current moment and the relative order of the system.

2. The method for constructing a proxy model based on an extended state observer according to claim 1, characterized in that: In step S1 , the input information and output information of the modeled process are measurement information, and the measurement information is directly acquired by a measurement tool.

3. The method for constructing a proxy model based on an extended state observer according to claim 1, characterized in that: Based on physical principles, the transfer function or state-space model of the system is established and analyzed. By observing the numerator and denominator polynomials of the transfer function, or analyzing the state variables and output equations in the state-space model, the relative order of the system can be calculated.

4. The method for constructing a proxy model based on an extended state observer according to claim 3, characterized in that: In step S2, when the relative order information of the modeled process cannot be calculated, the relative order information is set to a low-order relative order, and the low-order relative order includes two and three.

5. The method for constructing a proxy model based on an extended state observer according to claim 1, characterized in that: In step S3, the output equation is used to describe the factors affecting the prediction of the next output by the input information, system dynamics and system uncertainty dynamics in the modeled process.

6. The method for constructing a proxy model based on an extended state observer according to claim 1, characterized in that: In step S4, the extended state observer includes a linear extended state observer and a nonlinear extended state observer, which are used to obtain the total disturbance information of the system from the measurable input and output information, and dynamically expand the uncertainty into a system state.

7. The method for constructing a proxy model based on an extended state observer according to claim 1, characterized in that: In step S4, when n>3, the relative order n is high-order; when n≤3, the relative order n is low-order; when the relative order information of the modeled process is high-order, a low-order extended state observer is selected, and the low-order extended state observer includes a second-order extended state observer and a third-order extended state observer.

8. The method for constructing a proxy model based on an extended state observer according to claim 1, characterized in that: In step S6, the output equation formula is updated to obtain a standard integrator series configuration, the total disturbance of the system is estimated using an observer, and the estimated value is added to the product of the system modelable dynamics and the input signal and the adjustable gain to obtain an intermediate variable; Integrate the intermediate variable for a relative number of times to obtain an estimate of the output signal at the next moment.

9. An electronic device, characterized in that: include: A processor and a memory, wherein the memory stores a computer program that can be called by the processor; The processor executes the method for constructing an agent model based on an extended state observer according to any one of claims 1 to 8 by calling the computer program stored in the memory.

10. A computer-readable storage medium, characterized in that: Instructions are stored, and when the instructions are executed on a computer, the computer executes the agent model construction method based on an extended state observer as described in any one of claims 1 to 8.

Images