Intelligent suspension robust control and state observation method based on LMI

Abstract

The invention discloses an LMI-based intelligent suspension robust control and state observation method, and relates to the technical field of suspension control, and the method comprises the following steps: S1, splitting a control system into a non-matching disturbance subsystem and a matching disturbance subsystem; s2, determining state feedback control of an LMID algorithm according to the weight matrix of state control; determining disturbance compensation of a disturbance observer part in an LMID algorithm based on a weight matrix of a matching disturbance subsystem and disturbance estimation, and realizing the LMID algorithm by using an LMIob observer; and S3, taking the sprung acceleration based on the IMU in the suspension system as the input quantity of the LMIob observer, and obtaining state estimation of the control system. The method is designed based on an LMID algorithm of LMI, the algorithm has disturbance observation properties, the robustness of the algorithm can be improved, a weight matrix is introduced, and flexible adjustment of state control and disturbance estimation is achieved.

Description

Technical Field

[0001] This invention relates to the field of suspension control technology, specifically to an intelligent suspension robust control and state observation method based on LMI. Background Technology

[0002] Intelligent suspension systems refer to a type of suspension that can dynamically adjust vehicle dynamics characteristics and improve typical ride performance such as ride comfort and handling stability through actuators and adapted advanced control algorithms.

[0003] The controller designed based on the Linear Matrix Inequality (LMI) and Lyapunov functions is a typical control method for suspension systems. However, the basic LMI control ignores disturbances during its development and cannot guarantee robustness well, necessitating algorithm improvement. Furthermore, to realize the control system, a state observer should be specifically designed to acquire the control system state. Summary of the Invention

[0004] To address the above problems, this invention proposes an intelligent suspension robust control and state observation method based on LMI.

[0005] The technical solution of this invention is: a robust control and state observation method for intelligent suspension based on LMI, comprising the following steps:

[0006] S1. The control system is split into a non-matched disturbance subsystem and a matched disturbance subsystem;

[0007] S2. Based on the weight matrix of the state control, determine the state feedback control of the LMID algorithm; based on the weight matrix of the matched disturbance subsystem and the disturbance estimation, determine the disturbance compensation of the disturbance observer part in the LMID algorithm, and implement the LMID algorithm using the LMIob observer.

[0008] S3. Use the sprung acceleration based on the IMU in the suspension system as the input to the LMIob observer to obtain the state estimate of the control system.

[0009] Furthermore, in S1, the expression for separating the unmatched perturbation subsystem and the matched perturbation subsystem is:

[0010] ;

[0011] in, Represents the state of the non-matched disturbance subsystem The derivative, Indicates the state of the matched disturbance subsystem The derivative, This represents the steady system matrix of the unmatched perturbation subsystem. This represents the steady system matrix of the matched perturbation subsystem. Represents the system's state variables. This represents the steady-state input matrix of the unmatched perturbation subsystem. This represents the steady-state input matrix of the matched perturbation subsystem. Indicates the system's control input, This represents the known terms of the non-matched perturbation subsystem. This represents the known terms of the matched perturbation subsystem. This represents the external disturbance of the unmatched disturbance subsystem, i.e., the unmatched disturbance. This represents the external disturbance of the matched disturbance subsystem, i.e., the matched disturbance.

[0012] Furthermore, in S21, the disturbance estimation The expression is:

[0013] ;

[0014] in, This represents the steady-state feedback gain of the perturbation observer part in the LMID algorithm. Represents the system's state variables. The system matrix represents the steady-state system matrix. Represents the system's constant input matrix. This represents the steady feedback gain of the feedback controller portion in the LMID algorithm. Further, in S22, the weight matrix for state control... The expression is:

[0015] ;

[0016] in, A simplified representation of a diagonal matrix. The elements in the matrix are the diagonal elements. This refers to the feedback controller part in the LMID algorithm, specifically for the state variables. The weights of the components in the first dimension. This refers to the feedback controller part in the LMID algorithm, specifically for the state variables. No. The weights of the components in each dimension. Represents the real number field. Represents state variables Dimensions.

[0017] In S22, the weight matrix for perturbation estimation The expression is:

[0018] ;

[0019] in, This refers to the perturbation observer part of the LMID algorithm, which is used to observe perturbations. The weights of the components in the first dimension. This refers to the perturbation observer part of the LMID algorithm, which is used to observe perturbations. No. The weights of the components in each dimension. A simplified representation of a diagonal matrix. The elements in the matrix are the diagonal elements.

[0020] Furthermore, in S3, the sprung acceleration The expression is:

[0021] ;

[0022] in, This represents the triaxial acceleration at the inertial measurement unit. This represents the triaxial angular acceleration at the IMU. This indicates the radius vector pointing from the IMU to the center of mass of the chassis. This represents the three-axis angular velocity at the IMU.

[0023] Furthermore, in S3, the expression for the state estimation of the control system is:

[0024] ;

[0025] in, Represents the state variables of the system The estimated value The derivative, Represents the state variables of the system The estimated value, Represents the state variables of the system The derivative, The system matrix represents the steady-state system matrix. Represents the system's constant input matrix. Indicates the system's control input, Represents the known terms of the system. This indicates external disturbances to the system.

[0026] The beneficial effects of this invention are:

[0027] (1) The present invention is based on the design of the LMID algorithm of LMI. The algorithm has the property of perturbation observation, which can improve the robustness of the algorithm. The weight matrix is ​​introduced to realize the flexible adjustment of state control and perturbation estimation.

[0028] (2) The present invention is based on the design of LMIob observer, which estimates the weight matrix for different states.

[0029] (3) In order to improve ride comfort, this invention uses the LMID algorithm as the controller for the control system design of the suspension system and uses the LMIob observer to obtain the state of the controller. Attached Figure Description

[0030] Figure 1 The flowchart shows the method for robust control and state observation of intelligent suspension based on LMI.

[0031] Figure 2 This is a schematic diagram of a 1 / 4 scale model with active suspension.

[0032] Figure 3 A schematic diagram of a suspension control system based on LMI design, including the LMID algorithm as the controller and the LMIob observer as the state observer;

[0033] Figure 4 This is a schematic diagram of a method for acquiring sprung acceleration information via a chassis domain IMU;

[0034] Figure 5 A schematic diagram showing the comparative control effects of the designed control system, passive suspension, and basic LMI algorithm on random road surfaces for three performance indicators: sprung acceleration, wheel dynamic deformation, and suspension dynamic deflection.

[0035] Figure 6 This diagram illustrates the observation performance of the LMID algorithm on matched perturbations and the estimation performance of the LMIob observer on the state of the on-spring system. Detailed Implementation

[0036] The embodiments of the present invention will be further described below with reference to the accompanying drawings.

[0037] like Figure 1 As shown, this invention provides a robust intelligent suspension control and state observation method based on LMI, comprising the following steps:

[0038] S1. The control system is split into a non-matched disturbance subsystem and a matched disturbance subsystem;

[0039] S2. Based on the weight matrix of the state control, determine the state feedback control of the LMID algorithm; based on the weight matrix of the matched disturbance subsystem and the disturbance estimation, determine the disturbance compensation of the disturbance observer part in the LMID algorithm, and implement the LMID algorithm using the LMIob observer.

[0040] S3. Use the sprung acceleration based on the IMU in the suspension system as the input to the LMIob observer to obtain the state estimate of the control system.

[0041] This invention designs two general algorithms, the LMID algorithm, and applies them to the ride comfort control of suspension systems. The LMID algorithm is a control algorithm comprising feedback control and a disturbance observer, aiming to stabilize the control system. Implementing this algorithm requires knowledge of the control system state. The LMIob observer is a state observer used to observe the control system state to implement the controller. In suspension ride comfort control, ride comfort can be reflected by sprung acceleration. In the designed control system, the LMID algorithm is used to improve ride comfort. Implementing the LMID algorithm requires obtaining the suspension dynamic deflection and sprung velocity, which are then estimated by the LMIob observer in conjunction with the sprung acceleration. This invention incorporates four weighting designs: the weights of the LMID algorithm's state feedback control component for each term of the control system state; the weights of the LMID algorithm's disturbance observer component for each term of disturbance; the weights of the disturbances in the LMIob observer for each term of estimation error; and the weights of the LMIob observer for each term of disturbance (process noise).

[0042] In this embodiment of the invention, in S1, the expression for splitting the unmatched perturbation subsystem and the matched perturbation subsystem is:

[0043] ;

[0044] in, Represents the state of the non-matched disturbance subsystem The derivative, Indicates the state of the matched disturbance subsystem The derivative, This represents the steady system matrix of the unmatched perturbation subsystem. This represents the steady system matrix of the matched perturbation subsystem. Represents the system's state variables. This represents the steady-state input matrix of the unmatched perturbation subsystem. This represents the steady-state input matrix of the matched perturbation subsystem. Indicates the system's control input, This represents the known terms of the non-matched perturbation subsystem. This represents the known terms of the matched perturbation subsystem. This represents the external disturbance of the unmatched disturbance subsystem, i.e., the unmatched disturbance. This represents the external disturbance of the matched disturbance subsystem, i.e., the matched disturbance.

[0045] In the system In each channel, for those containing control input By aggregating these channels, the unmatched perturbation subsystem is obtained. For applications that do not include control inputs The channels are aggregated to obtain the matched perturbation subsystem. . and For the original system state variables The weight, without loss of generality, is defined. ( (represents matrix transpose), and has , , and .

[0046] In this embodiment of the invention, in S21, the perturbation is estimated. The expression is:

[0047] ;

[0048] in, This represents the steady-state feedback gain of the perturbation observer part in the LMID algorithm. Represents the system's state variables. The system matrix represents the steady-state system matrix. Represents the system's constant input matrix. This represents the steady feedback gain of the feedback controller section in the LMID algorithm.

[0049] In this embodiment of the invention, in S22, the weight matrix for state control... The expression is:

[0050] ;

[0051] in, A simplified representation of a diagonal matrix. The elements in the matrix are the diagonal elements. This refers to the feedback controller part in the LMID algorithm, specifically for the state variables. The weights of the components in the first dimension. This refers to the feedback controller part in the LMID algorithm, specifically for the state variables. No. The weights of the components in each dimension. Represents the real number field. Represents state variables Dimensions.

[0052] In S22, the weight matrix for perturbation estimation The expression is:

[0053] ;

[0054] in, This refers to the perturbation observer part of the LMID algorithm, which is used to observe perturbations. The weights of the components in the first dimension. This refers to the perturbation observer part of the LMID algorithm, which is used to observe perturbations. No. The weights of the components in each dimension. A simplified representation of a diagonal matrix. The elements in the matrix are the diagonal elements.

[0055] In this embodiment of the invention, in S3, the spring acceleration... The expression is:

[0056] ;

[0057] in, This represents the triaxial acceleration at the inertial measurement unit. This represents the triaxial angular acceleration at the IMU. This indicates the radius vector pointing from the IMU to the center of mass of the chassis. This represents the three-axis angular velocity at the IMU.

[0058] In this embodiment of the invention, in S3, the expression for the state estimation of the control system is:

[0059] ;

[0060] in, Represents the state variables of the system The estimated value The derivative, Represents the state variables of the system The estimated value, Represents the state variables of the system The derivative, The system matrix represents the steady-state system matrix. Represents the system's constant input matrix. Indicates the system's control input, Represents the known terms of the system. This indicates external disturbances to the system.

[0061] The controller designed in this invention includes the following six components.

[0062] First, regarding the original system Split according to unmatched and matched perturbations;

[0063] Second, the LMID algorithm based on LMI is designed for controller design, and is introduced into the Lyapunov function. and The two weight matrices correspond to system state control and disturbance estimation, respectively;

[0064] Third, the proof of the consistent boundedness of the LMID algorithm;

[0065] Fourth, a method for acquiring sprung acceleration signals via a chassis domain IMU;

[0066] Fifth, the design of an LMIob observer based on LMI is used to acquire the state of the control system;

[0067] Sixth, proof of the uniform boundedness of LMIob observers.

[0068] Intelligent suspension systems refer to a type of suspension that dynamically adjusts vehicle dynamics through actuators and adapted advanced control algorithms to improve typical ride performance characteristics such as ride comfort and handling stability. They mainly include active actuators that directly generate force, such as electric hydraulic pumps or linear motors, collectively referred to as active suspension systems. Semi-active actuators that can adjust suspension dynamic parameters, such as CDC dampers and magnetorheological dampers that can adjust damping parameters, and air springs that can adjust vehicle height and spring stiffness, are also collectively referred to as semi-active suspension systems.

[0069] The development of advanced control algorithms for active suspension systems typically uses... Figure 2 The controller design is based on a 1 / 4 car model, which is then organized into state-space equations.

[0070] The 1 / 4 scale car model consists of two subsystems: the sprung subsystem and the unsprung subsystem, which are related to ride comfort and handling stability, respectively. The dynamic equations of the two subsystems are expressed as follows:

[0071] (1);

[0072] In the formula, The vertical displacement of the sprung mass; The vertical displacement of the unsprung mass; For vertical excitation of the road surface; For suspension dynamic deflection; For wheel-driven deformation; For the sprung mass; Unsprung mass; For spring stiffness; For tire stiffness; For damping of the shock absorber; For tire damping; It provides the power for the actuator.

[0073] The state-space equations of a control system are generally expressed as:

[0074] (2);

[0075] In the formula, For system state variables; Input for the system; These are known terms; For lumped disturbance, and ,in External disturbance; , , It is a constant matrix.

[0076] In the control of the sprung system related to ride comfort, the sprung system equation in equation (1) is usually processed, and the state variables are defined as follows: The control quantity is The disturbance term is And there are:

[0077] , , .

[0078] For the design of control systems in the state-space equation form mentioned above, including typical LQR, MPC, and H2 / H ∞ Control, etc. To implement the above control methods, it is also necessary to acquire the system status in real time. In a suspension system, the state that can usually be directly obtained is the acceleration of the sprung mass. Without considering the deployment of additional sensors, a state observer needs to be designed to achieve [the desired effect]. right The estimate. And because The inability to obtain this information is considered a disturbance in the design of state observers and controllers, requiring both designs to be robust.

[0079] Linear Matrix Inequality (LMI) and Lyapunov functions are typical controller design methods. In basic LMI control, [the following is a more detailed explanation of the LMI approach]. This system defines a symmetric positive definite matrix. Design steady feedback gain Control rate Design Lyapunov functions. Define scalar ,make ,get:

[0080] (3);

[0081] Equation (3) is about and For nonlinear systems, in order to solve the above equation, we define... , Multiplying both sides of the inequality by the left and right sides ,get:

[0082] (4);

[0083] Thus, it transforms into something about and The linear matrix inequalities, combined with the given... Solving this equation, we get .

[0084] Basic LMI control is for the system The algorithm's formulation ignores known terms. and disturbance terms When directly applied to the system form of equation (2), robustness cannot be guaranteed. Therefore, when the control quantity can be controlled... compensate And use the perturbation observer algorithm to compensate for the perturbation. This can further improve the robustness of the system.

[0085] The algorithm designed in this invention consists of two main parts: the controller LMID algorithm and the state observer LMIob. The controller structure is as follows: Figure 3 As shown. The LMID algorithm has disturbance compensation capability because it compensates through control inputs, meaning it can only handle matched disturbances. Therefore, before introducing the LMID algorithm, the system needs to be defined according to unmatched / matched disturbances. In the LMIob observer, the sprung mass acceleration, which is the observer input, is considered. The acquisition process will first be introduced through the chassis domain IMU. Therefore, the following will be introduced in six parts, including system definition, LMID algorithm, proof of uniform boundedness of LMID algorithm, acquisition of sprung acceleration signal, LMIob observer, and proof of uniform boundedness of LMIob observer.

[0086] ① Definition of unmatched / matched disturbance systems

[0087] For the system in equation (2), without loss of generality, we define ,in The channel in question is experiencing a mismatched disturbance. To match the perturbation, the matrix is ​​split along the same row dimension. , Known terms and aggregate disturbance ,get , , and The system (2) is split into unmatched / matched perturbation subsystems:

[0088] (5);

[0089] In the formula: unmatched / matched disturbances are generally classified according to whether the channel contains control variables, that is, generally there are .

[0090] ② Design of a robust control / disturbance observer algorithm based on LMI: LMID algorithm

[0091] definition , , To match the disturbance The estimated value, The estimation error is defined as... The design control rate is:

[0092] (6);

[0093] Define constant gain The design disturbance estimate is:

[0094] (7);

[0095] Substituting equation (6) into the first expression of equation (7), we get Combining this with the second equation, we obtain the disturbance estimate as follows:

[0096] (8);

[0097] This algorithm is named the LMID algorithm. To calculate the gain... and Taking the derivative of the second expression in equation (7), we get:

[0098] (9);

[0099] Let equation (9) be equivalent to the first expression of equation (7):

[0100] (10);

[0101] And by get ,thereby:

[0102] (11);

[0103] definition and They are respectively and The weights are defined as follows:

[0104] (12);

[0105] In the formula, ; .

[0106] definition , , Symmetric positive definite matrix Combining equation (11), we get:

[0107] (13);

[0108] definition And since all diagonal elements are positive, therefore:

[0109] (14);

[0110] Therefore, equation (13) can be rearranged as follows:

[0111] (15);

[0112] In the formula, , .

[0113] Define scalar and define The results are as follows:

[0114] (16);

[0115] In the formula, , , , .

[0116] By solving equation (16), two gains are obtained. and .

[0117] ③ Proof of the consistent boundedness of the LMID algorithm

[0118] To illustrate the uniform boundedness of the system, Substituting into equation (15), we get:

[0119] (17);

[0120] When satisfied That is, there exists a scalar When the boundary of this expression is indicated, the above expression can be rearranged as:

[0121] (18);

[0122] If defined as:

[0123] (19);

[0124] Differentiating the above equation and substituting it into equation (18), we get... ,thereby And the results were compiled as follows:

[0125] (20);

[0126] That is, non-negative numbers There is an upper limit.

[0127] In addition, definition , and obtain the system From what was obtained and The matrix can be obtained. The eigenvalues ​​of the system are such that if all eigenvalues ​​are negative real parts, it can be said that the system is bounded. Boundedness under [condition]. If defined... , ,but , i.e., matrix yes The two are similar transformations and have the same eigenvalues.

[0128] ④ Acquisition of sprung acceleration signals based on chassis domain IMU

[0129] Sprung acceleration is obtained based on chassis domain IMU signals. The process of the signal is as follows Figure 4 As shown. The specific calculation method is as follows: the three-axis acceleration is obtained through the chassis domain IMU. and angular velocity And differentiate to obtain angular acceleration Combined with the radius vector from the IMU to the vehicle's center of gravity calibration The triaxial acceleration at the center of mass of the vehicle body is calculated using the acceleration composition formula for a point on a rigid body. for:

[0130] (twenty one);

[0131] Since the angular velocity and angular acceleration are the same at different positions on the rigid body, that is, the three-axis angular velocity at the center of mass of the vehicle body is the same. angular acceleration Therefore, according to equation (21), the sprung accelerations of the vehicle body at the four wheels can be calculated. .

[0132] ⑤ Design of a state observer based on LMI: LMIob observer

[0133] For the system in equation (2), the observation system is designed as follows:

[0134] (twenty two);

[0135] In the formula, To The estimated value; The measured value, and ; The gain to be designed; , , It is a constant matrix.

[0136] Define observation error The error system is expressed as:

[0137] (twenty three);

[0138] definition The Lyapunov function is designed as follows:

[0139] (twenty four);

[0140] In the formula, Denotes the weight matrix, and , .

[0141] Differentiating the above equation, we get:

[0142] (25);

[0143] In the formula, ; ; .

[0144] definition , , Considering Therefore:

[0145] (26);

[0146] Define scalar Solve , obtain the observer gain .

[0147] Furthermore, due to the introduction of the observer, the system under actual control is equivalent to the system in equation (22), and the system in equation (2) can be rewritten as:

[0148] (27);

[0149] Even after using the observer, the actual equivalent perturbation processed by the perturbation observer algorithm is: And there is a relationship .

[0150] ⑥ Proof of the uniform boundedness of the LMID algorithm

[0151] Similarly, when the conditions are met Then, equation (27) can be rearranged as follows: .definition ,but ,have ,thereby non-negative numbers It has an upper bound; the observation system is uniformly bounded.

[0152] In the simulation model parameter settings, the sprung mass Take 345kg; unsprung mass Take 40.5 kg; spring stiffness Take 18000 Tire stiffness Take 192000 ; Vibration damper damping Take 1000 Tire damping Take 50 The operating force of the actuator The range is [-2500, 2500]. Speed Take 20 .

[0153] The comparison algorithm uses uncontrolled passive suspension and basic LMI control, on a Class C random road surface, regarding sprung acceleration. Wheel deformation Suspension dynamic deflection The simulation results for the three performance indicators are as follows: Figure 5 As shown in Table 1, the root mean square (rms) values ​​are as follows.

[0154] Table 1

[0155] Performance indicators passive suspension Basic LMI control LMID algorithm (LMIob observer) 0.459 0.324 0.269 1.761 1.991 2.774 5.169 5.075 6.889

[0156] For sprung acceleration The control algorithm designed for this system is superior to basic LMI control and passive suspension. A smaller RMS corresponds to a better improvement in smoothness, indicating that the algorithm has better robustness.

[0157] Deformation of the wheel and suspension dynamic deflection The designed algorithm deteriorates in both of these metrics because there is a conflict between smoothness and handling stability in the control itself.

[0158] Regarding the designed algorithm, the estimation of the matched perturbation by the LMID algorithm and the state estimation of the sprung system by the LMIob observer are as follows: Figure 6 As shown, the designed disturbance observer and state observer have small estimation errors and achieve good estimation, providing relatively accurate system information for the implementation of the control system.

[0159] Those skilled in the art will recognize that the embodiments described herein are intended to help the reader understand the principles of the invention, and should be understood that the scope of protection of the invention is not limited to such specific statements and embodiments. Those skilled in the art can make various other specific modifications and combinations based on the technical teachings disclosed in this invention without departing from the spirit of the invention, and these modifications and combinations are still within the scope of protection of this invention.

Claims

1. A robust control and state observation method for intelligent suspension based on LMI, characterized in that, Includes the following steps: S1. The control system is split into a non-matched disturbance subsystem and a matched disturbance subsystem; S2. Based on the weight matrix of the state control, determine the state feedback control of the LMID algorithm; based on the weight matrix of the matched disturbance subsystem and the disturbance estimation, determine the disturbance compensation of the disturbance observer part in the LMID algorithm, and implement the LMID algorithm using the LMIob observer. S3. Use the sprung acceleration based on the IMU in the suspension system as the input to the LMIob observer to obtain the state estimate of the control system.

2. The method for robust control and state observation of intelligent suspension based on LMI according to claim 1, characterized in that, In S1, the expression for separating the unmatched perturbation subsystem and the matched perturbation subsystem is: ； in, Represents the state of the non-matched disturbance subsystem The derivative, Indicates the state of the matched disturbance subsystem The derivative, This represents the steady system matrix of the unmatched perturbation subsystem. This represents the steady system matrix of the matched perturbation subsystem. Represents the system's state variables. This represents the steady-state input matrix of the unmatched perturbation subsystem. This represents the steady-state input matrix of the matched perturbation subsystem. Indicates the system's control input, This represents the known terms of the non-matched perturbation subsystem. This represents the known terms of the matched perturbation subsystem. This indicates a non-matching perturbation. This indicates a matching perturbation.

3. The method for robust intelligent suspension control and state observation based on LMI according to claim 1, characterized in that, In S2, the disturbance estimation The expression is: ； in, This represents the steady-state feedback gain of the perturbation observer part in the LMID algorithm. Represents the system's state variables. The system matrix represents the steady-state system matrix. Represents the system's constant input matrix. This represents the steady feedback gain of the feedback controller section in the LMID algorithm.

4. The method for robust control and state observation of intelligent suspension based on LMI according to claim 1, characterized in that, In S2, the weight matrix for state control The expression is: ； in, A simplified representation of a diagonal matrix. The elements in the matrix are the diagonal elements. Represents state variables The weights of the components in the first dimension. Represents state variables No. The weights of the components in each dimension. Represents the real number field. Represents state variables Dimensions.

5. The method for robust control and state observation of intelligent suspension based on LMI according to claim 1, characterized in that, In S2, the weight matrix for perturbation estimation The expression is: ； in, Indicates disturbance The weights of the components in the first dimension. Indicates disturbance No. The weights of the components in each dimension. A simplified representation of a diagonal matrix. The elements in the matrix are the diagonal elements.

6. The method for robust control and state observation of intelligent suspension based on LMI according to claim 1, characterized in that, In S3, the spring acceleration The expression is: ； in, This represents the triaxial acceleration at the inertial measurement unit. This represents the triaxial angular acceleration at the IMU. This indicates the radius vector pointing from the IMU to the center of mass of the chassis. This represents the three-axis angular velocity at the IMU.

7. The method for robust control and state observation of intelligent suspension based on LMI according to claim 1, characterized in that, In S3, the expression for the state estimation of the control system is: ； in, Represents the state variables of the system The estimated value The derivative, Represents the state variables of the system The estimated value, Represents the state variables of the system The derivative, The system matrix represents the steady-state system matrix. Represents the system's constant input matrix. Indicates the system's control input, Represents the known terms of the system. This indicates external disturbances to the system.

Images