Robust control method for automobile electric control semi-active suspension

Abstract

The invention discloses a robust control method of an automobile electric control semi-active suspension, which is applied to a semi-active suspension system comprising a magneto-rheological shock absorber, an electromagnetic valve type shock absorber and the like, and comprises the following steps: (1) under a uniform discrete state space frame, carrying out robust control on the semi-active suspension; establishing a multi-cell vertex model by taking the current of the semi-active shock absorber or the opening degree of an electromagnetic valve and other control quantities as direct control quantities; and (2) integrating the sampling uncertainty, the amplitude of the actuator and the change rate constraint into a single H-infinity robust control for comprehensive solution through the frequency domain weight and the energy consumption weight. And (3) realizing cooperative control of four corners of the vehicle body and pitching and rolling postures of the whole vehicle by adopting structured feedback gain. Robust control can be achieved on an existing electric control semi-active suspension structure, the damping force of the semi-active shock absorber is precisely and dynamically adjusted within the safety boundary, and therefore the smoothness and operation stability of a vehicle are comprehensively improved.

Description

Technical Field

[0001] This invention relates to the field of electronically controlled semi-active suspension control algorithms, specifically a robust control method for automotive electronically controlled semi-active suspension. Background Technology

[0002] As a crucial component of automobiles, the suspension system significantly impacts a vehicle's overall performance, including ride comfort, handling stability, and safety. With increasing vehicle speeds and diversified operating conditions, the road excitations and load variations encountered by vehicles on real roads exhibit significant randomness and uncertainty. Traditional passive suspensions with fixed parameters struggle to achieve a optimal balance between comfort and handling stability. To improve suspension performance and enhance a vehicle's adaptability to complex operating conditions, semi-active suspensions based on electronically controlled variable damping dampers are increasingly being widely adopted, including various electronically controlled forms such as magnetorheological dampers and solenoid valve dampers.

[0003] As the core of a semi-active suspension system, suspension control algorithms have always been a key focus of research and engineering development both domestically and internationally. Existing semi-active suspension controls mostly employ a "separate-channel control + post-arbitration" architecture, meaning they address different control objectives such as vertical ride comfort, vehicle attitude control, and tire grip by combining corresponding control algorithms (e.g., roof control, ADD control, sliding mode control, H∞ / H2 control, etc.); or they use manual or empirical arbitration by setting priorities and weights for different operating conditions. Some solutions trigger control mode switching under extreme conditions by setting thresholds to meet safety or attitude control requirements. While this type of control structure can achieve certain results under decoupled operating conditions, in complex real-world conditions, it easily leads to the problem of different channels "competing" for control, making it difficult to simultaneously achieve optimal vehicle attitude and ride comfort at low frequencies and optimal tire grip and safety at high frequencies.

[0004] Furthermore, most existing semi-active suspension control strategies assume that the control period and sampling period are fixed constants, ignoring the sampling period jitter and parameter uncertainty caused by changes in task load and bus delays in actual operation of the vehicle's electronic control unit. Simultaneously, semi-active dampers themselves have current or valve position limitations, rate-of-change limitations, and energy consumption constraints. In engineering implementation, these are often addressed through simple saturation, amplitude limiting, or rate-of-change limiting modules performing "post-arbitration" after the control law. This easily introduces additional nonlinearity, leading to decreased robustness and even causing oscillations and performance degradation. Moreover, some existing H∞ or multi-objective optimization control methods typically rely on offline weighted parameter settings to make fixed trade-offs between ride comfort, attitude control, and tire adhesion. Once the weighted parameters are determined, the control tendency is also fixed, making it difficult to flexibly adjust according to changes in operating conditions. Furthermore, they are often designed only for a specific type of actuator or a simplified model (such as a quarter-vehicle model), with insufficient consideration of the multi-angle coupling effects of the entire vehicle.

[0005] In summary, existing semi-active suspension control schemes often lack a unified robust control framework when facing sampling uncertainty and actuator amplitude and rate of change constraints. They are unable to simultaneously and collaboratively suppress low-frequency vehicle posture and high-frequency tire bounce under four-wheel multi-angle coupling conditions, and achieve precise dynamic adjustment of damper damping force. Summary of the Invention

[0006] To overcome the shortcomings of existing technologies, this invention proposes a robust control method for automotive electronically controlled semi-active suspension. This method aims to utilize the structure of existing electronically controlled semi-active suspension systems to incorporate sampling period uncertainty and the control amplitude and rate of change constraints of the electronically controlled semi-active damper into the same robust control framework. This achieves coordinated suppression of low-frequency vehicle posture at the four corners and high-frequency tire bounce, thereby enabling precise dynamic adjustment of the actuator's damping force and comprehensively optimizing vehicle posture stability, ride comfort, and energy consumption.

[0007] To achieve the above objectives, the present invention adopts the following technical solution: The present invention discloses a robust control method for an electronically controlled semi-active automotive suspension, applicable to semi-active automotive suspensions that include magnetorheological dampers and electromagnetic valve dampers. The robust control method is characterized by the following steps: Step 1: Obtain vehicle status information and vehicle parameters, and establish the actuator output damping force of the j-th wheel of the semi-active suspension. With control quantity Relationship; Step 2: Based on the output damping force of the actuator With actual control quantity The relationship is used to establish the control quantity for the actuator controlling the j-th wheel. Spatial state equation; Step 3: Discretize the spatial state equations and construct the multi-cell vertex models at the four corners of the vehicle body to obtain the multi-cell vertex set; Step 4: Based on vehicle parameters and vehicle state information, construct the weighted performance output matrix for the j-th wheel. Low-frequency attitude matrix of the whole vehicle Thus, the overall performance matrix is ​​obtained. ; Step 5: Establish the structured gain matrix K and the overall vehicle state matrix respectively. Thus, the overall state matrix of the vehicle is constructed. With control quantity theory output matrix The structured control law relation; Step 6: Based on the multi-cell vertex model at the four corners of the vehicle body, establish a discrete multi-cell model of the whole vehicle, and solve the discrete multi-cell model of the whole vehicle through H∞ optimization control, thereby obtaining a structured gain matrix that satisfies the robust performance and constraint conditions on the set of multi-cell vertices. ; Based on structured gain matrix The theoretical output matrix of the control quantity is obtained using the structured control law relationship. and to After applying constraints, the actual output matrix of the control quantity is obtained. An actuator used to apply electronically controlled semi-active suspension to a vehicle.

[0008] The robust control method for an electronically controlled semi-active suspension of an automobile, as described in this invention, is characterized in that step 1 is performed as follows: Step 1.1: Obtain vehicle parameters, including: horizontal distance from the front axle to the center of gravity. Horizontal distance from rear axle to center of gravity Front wheel track Rear wheel track The sprung mass of the j-th wheel The unsprung mass of the j-th wheel The suspension spring stiffness of the j-th wheel The stiffness of the j-th wheel The suspension damping of the j-th wheel The maximum suspension travel of the j-th wheel ,in, Indicates the index of the wheel, and These respectively represent left front, right front, left rear, and right rear; Step 1.2: Obtain vehicle status information, including: vertical displacement of the j-th wheel on the road surface. ,speed and its acceleration The vertical displacement of the vehicle body of the j-th wheel The vertical displacement of the j-th wheel The suspension travel of the j-th wheel Vertical displacement of the vehicle's center of gravity Vertical velocity of the vehicle's center of gravity Body center of gravity roll angle Body center of gravity roll rate Vehicle body center of gravity pitch angle Vehicle body center of gravity pitch angle acceleration The control quantity of the actuator of the j-th wheel and its rate of change Thus, the actual output matrix of the control quantity is established. and the matrix of actual output change rate of control quantity ; When the actuator of a car's semi-active suspension is a magnetorheological damper, let The coil current of the magnetorheological damper for the j-th wheel ; When the actuator of a car's semi-active suspension is a solenoid valve type shock absorber, then... The valve opening of the electromagnetic valve type shock absorber for the j-th wheel ; Step 1.3: Establish the output damping force of the actuator of the j-th wheel of the semi-active suspension of the car. With control quantity Relationship: When the actuator of the semi-active suspension of the car is a magnetorheological damper, the damping force output by the magnetorheological damper of the j-th wheel can be obtained using equation (1). : (1) In equation (1), , , These represent the three characteristic parameters of the magnetorheological vibration damper. The parameter representing the adjustment rate of the magnetorheological damper. This represents the operating state of the magnetorheological damper of the j-th wheel, and , It is a constant. This represents the suspension speed of the j-th wheel; When the actuator of the semi-active suspension of the car is an electromagnetic valve type shock absorber, the damping force output by the electromagnetic valve type shock absorber of the j-th wheel can be obtained using equation (2). : (2) In equation (2), This represents the minimum damping force of the electromagnetic valve type vibration damper. This indicates the maximum damping force of the electromagnetic valve type vibration damper.

[0009] Furthermore, step 2 is performed as follows: Step 2.1: Use equation (3) to establish a two-degree-of-freedom suspension model for the j-th wheel: (3) Step 2.2: Construct the state vector of the j-th wheel. and the derivative of the state vector of the j-th wheel Thus, equation (3) can be written as equation (4), where T represents transpose. This represents the suspension speed of the j-th wheel. Let represent the tire travel of the j-th wheel, and , This represents the tire speed of the j-th wheel. This represents the corner suspension acceleration of the j-th wheel. This represents the tire acceleration of the j-th wheel: (4) In equation (4), Let j be the state matrix of the j-th wheel. It is the damping force input matrix of the j-th wheel. It is the road disturbance input matrix for the j-th wheel. Let represent the road disturbance of the j-th wheel, and , Indicates the road disturbance acceleration of the j-th wheel: Step 2.3, Substituting into equation (4), we can then use equation (5) to establish the control quantity for the actuator controlling the j-th wheel. Spatial state equation: (5) In equation (5), Let represent the equivalent state matrix after the j-th wheel is introduced into the actuator model, and , Let represent the control input matrix for the j-th wheel, and , Let represent the constant term matrix of the j-th wheel, and .

[0010] Furthermore, step 3 is performed as follows: Step 3.1: Set the sampling jitter time and actuator gain ,in, Indicates the minimum sampling duration. Indicates the maximum sampling duration. Indicates the minimum actuator gain. Indicates the maximum actuator gain; right The spatial state equations within the wheel are discretized, and the parameter discrete model of the j-th wheel is constructed using equation (6): (6) In equation (6), Let represent the state vector of the j-th wheel of the semi-active suspension at time k. Let represent the state vector of the j-th wheel of the semi-active suspension at time k+1. Indicates that the j-th wheel is in The discrete state transition matrix within, and I represents the identity matrix. Indicates that the j-th wheel is in inner harmony The discrete control input matrix under the given conditions, and , Indicates that the j-th wheel is in The discrete road surface disturbance input matrix within, and , Let represent the road disturbance of the j-th wheel at time k. Indicates that the j-th wheel is in inner harmony The discrete constant term matrix under the given conditions, and ; Step 3.2, in Discrete vertex set ,in, Indicates the first The sampling duration corresponding to each vertex Indicates the first The actuator gain corresponding to each vertex, where N represents the total number of vertices; Construct the first using equation (7) Multicell Vertex The multi-cell vertex model at the four corners of the lower body is used to obtain the multi-cell vertex set. ,in, The matrix representing the corner of the j-th wheel: (7) In equation (7), Indicates the first Each weighting coefficient Indicates the first The j-th wheel corresponding to the multicell vertex is in The discrete state transition matrix within, the j-th wheel in inner harmony The discrete control input matrix and the j-th wheel are shown below. Discrete road surface disturbance input matrix within, j-th wheel in inner harmony The discrete constant term matrix within the given information.

[0011] Furthermore, step 4 is performed as follows: Step 4.1: Define the performance output matrix of the j-th wheel using equation (8). : (8) In equation (8), T represents transpose. This represents the vehicle body acceleration of the j-th wheel. This represents the maximum displacement of the actuator of the j-th wheel. This indicates the energy consumption of a car's semi-active suspension. This represents the rate of change of the control quantity of the actuator. Represents gravitational acceleration; Step 4.2: Use equation (9) to obtain the low-frequency vehicle body bandpass weight. : (9) In equation (9), s represents the Laplace complex frequency domain variable. This represents the low-frequency weighting amplitude coefficient. The center angular frequency represents the low-frequency vehicle body bandpass weight, and , This indicates the center angular frequency of the low-frequency vehicle body bandpass weight. Indicates the low-frequency bandwidth parameter; Step 4.3: Use equation (10) to obtain the high-frequency tire bandpass weight. : (10) In equation (10), Represents the high-frequency weighting amplitude coefficient. The center angular frequency represents the high-frequency tire bandpass weight. The center angular frequency represents the high-frequency tire bandpass weight. Indicates high-frequency bandwidth parameters; Step 4.4: Use equation (11) to obtain the weighted performance output matrix of the j-th wheel. : (11) In equation (11), Represents a diagonal matrix. This indicates the energy consumption weight of the control input of the actuator. Indicates the energy consumption weight of the rate of change of the control quantity of the actuator; Step 4.5: Obtain the low-frequency attitude matrix of the entire vehicle using equation (12). Thus, the overall performance matrix is ​​constructed. : (12) In equation (12), These represent the weights of the vehicle's vertical, pitch, and roll attitudes, respectively.

[0012] Furthermore, step 5 is performed as follows: Step 5.1: Define the structured gain matrix K using equation (13): (13) In equation (13), Represents the vehicle body state gain matrix. This represents the gain matrix for the vehicle corner state feedback. This indicates the construction of a diagonal block splicing matrix. Indicates the modal compensation gain. , , , These represent the local feedback gain matrices for the left front corner, right front corner, left rear corner, and right rear corner of the vehicle body, respectively. Let represent the geometrical distribution matrix, and we have: Step 5.2: Obtain the vehicle body attitude state matrix using equation (14). With the vehicle angle state matrix The relationship is used to establish the overall vehicle state matrix. ,in, express transpose: (14) In equation (14), It is a linear mapping matrix; Step 5.3: Use equation (15) to establish the overall vehicle state matrix. With the control quantity theoretical output matrix Structured control law relation: (15) In equation (15), , This represents the theoretical control quantity of the semi-active shock absorber for the j-th wheel.

[0013] Furthermore, step 6 is performed as follows: Step 6.1: Calculate the vehicle-enhanced road surface disturbance at time k. The j-th wheel is The whole vehicle augmented disturbance matrix ; Will In Merged into the first The augmented perturbation vertex of the j-th wheel corresponding to the multicell vertex. This allows for the construction of a merged set of multiple cell vertices for the entire vehicle. ; Step 6.2: In the merged set of vertices of the entire vehicle multicell Establish a discrete multicell model of the whole vehicle: (16) In equation (16), This represents the state matrix of the entire vehicle at time k. This represents the state matrix of the entire vehicle at time k+1. This represents the theoretical control output matrix of the vehicle at time k. Indicates the whole vehicle is Discrete state transition matrix within, Indicates the whole vehicle is Discrete control input matrix within; Step 6.3: Output matrix with total performance To provide the performance output for robust control, two performance output matrices are constructed. , and in the merged set of multiple cell vertices Establish conditions 1 and 2 respectively; Condition 1 is the output matrix of the closed-loop system to the total performance from road disturbance. The transfer gain satisfies the preset upper bound. And establish using equation (17): (17) In equation (17), T denotes transpose, and I denotes the identity matrix. It is a symmetric positive definite matrix. Let be a matrix, and , It is a scalar, and ; Condition 2 is to satisfy the constraints on the amplitude and rate of change of the control quantity, and is established using equations (18) and (19): Amplitude: (18) Rate of change: (19) In equations (18) and (19), This represents the matrix showing the actual rate of change of the control output, and , This represents the actual output matrix of the control quantity at time k. This represents the actual output matrix of the control quantity at time k-1. This represents the theoretical output rate of change matrix of the control quantity. , This represents the matrix showing the actual rate of change of the control output at time k. This represents the matrix showing the actual rate of change of the control output at time k-1. , Let each represent a matrix of two decision variables, and , I represents the identity matrix. This indicates that there are two weighting coefficients, and θ is the coefficient.

[0014] Step 6.4: Obtain the solution by solving equations (17), (18), and (19). Thus, the structured gain matrix that satisfies the robust performance and constraint conditions can be obtained using equation (20). And the theoretical control output matrix of the actuator is obtained using equation (15). : (20) Step 6.5: Obtain the actual output matrix of the control quantity using equation (21). : (twenty one) In equation (21), This represents the amplitude saturation function, used to limit the actual control quantity of the j-th wheel. satisfy , This represents the minimum control quantity. Indicates the maximum control value. Let the rate of change constraint function be defined such that the theoretical control rate of change matrix satisfies the following condition at any given time. , where r is the threshold.

[0015] The present invention provides an electronic device, including a memory and a processor, characterized in that the memory is used to store a program supporting the processor in executing the robust control method, and the processor is configured to execute the program stored in the memory.

[0016] The present invention provides a computer-readable storage medium storing a computer program, characterized in that the computer program, when executed by a processor, performs the steps of the robust control method.

[0017] Compared with existing technologies, the beneficial effects of this invention are reflected in: 1. This invention overcomes the problem of existing technologies requiring separate controller design or switching for different sampling periods by adopting the H∞ robust control method under the discrete robust control framework and representing the sampling period uncertainty in the form of a multi-cell model, thereby improving the consistency of the control strategy and the simplicity of engineering implementation.

[0018] 2. By introducing structured feedback gain constraints, this invention synchronously synthesizes the four-corner control of the vehicle body and the vertical, pitch, and roll compensation of the entire vehicle in the same solution, thereby achieving the coordinated suppression of low-frequency vehicle body attitude and high-frequency tire bounce. This solves the problem that existing architectures are unable to balance ride comfort and attitude stability under complex coupled conditions.

[0019] 3. This invention embeds the amplitude and rate of change of the semi-active damper control quantity into the H∞ synthesis solution during the control law design stage, thereby achieving a pre-guarantee of actuator constraints, improving the robust stability and performance of the closed-loop system, and reducing the adverse effects of saturation nonlinearity on the control effect.

[0020] 4. This invention is compatible with magnetorheological dampers and electromagnetic valve dampers. It does not require an accurate inverse model of the force-current and force-valve position relationship of the damper, nor does it require an online complex learner or time delay compensation structure. It is easy to implement on existing vehicle electronic control units with a low computational burden and has good engineering portability and practical value. Attached Figure Description

[0021] Figure 1 This is a schematic diagram of the principle and flow of the robust control method of the present invention; Figure 2 This is a schematic diagram of the overall vehicle layout and parameters of the present invention; Figure 3 This is a schematic diagram of the principle and flow of the discrete modeling layer of this invention; Figure 4 This is a schematic diagram of the robust control layer principle of the present invention. Detailed Implementation

[0022] In this embodiment, as Figure 1 As shown, a robust control method for an electronically controlled semi-active suspension is applied to a semi-active suspension system containing an electronically controlled variable damping damper. The method preferably uses a magnetorheological damper as the actuator. Without changing the structure of the existing electronically controlled semi-active suspension system, it achieves robust control over sampling uncertainty, actuator current constraints, and suspension safety boundaries by utilizing existing sensors and ECU calibration capabilities.

[0023] The control method includes: within a unified discrete state-space framework, using the current of the magnetorheological damper as the direct control variable, incorporating sampling uncertainties into the H∞ robust control synthesis solution through dual-peak frequency band weights for the vehicle body and tires, and energy consumption weights, setting a safety threshold, and combining time windows and hysteresis mechanisms to achieve safety monitoring and control mode degradation and recovery, and employing structured feedback gain to achieve coordinated control of the vehicle's four corners and overall pitch and roll attitudes. Specifically, the control method proceeds in the following steps: Step 1: Obtain vehicle status information and vehicle parameters, such as Figure 2The diagram shows the overall vehicle layout and parameters, and establishes the actuator output damping force of the j-th wheel of the semi-active suspension. With control quantity Relationship: Step 1.1: Obtain vehicle parameters, including: horizontal distance from the front axle to the center of gravity. Horizontal distance from rear axle to center of gravity Front wheel track Rear wheel track The sprung mass of the j-th wheel The unsprung mass of the j-th wheel The suspension spring stiffness of the j-th wheel The stiffness of the j-th wheel The suspension damping of the j-th wheel The maximum suspension travel of the j-th wheel ,in, Indicates the index of the wheel, and , respectively representing left front, right front, left back, and right back.

[0024] Step 1.2: Obtain vehicle status information, including: the vertical displacement, velocity, and acceleration of the j-th wheel on the road surface. , , The vertical displacement of the vehicle body of the j-th wheel The vertical displacement of the j-th wheel The suspension travel of the j-th wheel Vertical displacement of the vehicle's center of gravity Vertical velocity of the vehicle's center of gravity Body center of gravity roll angle Body center of gravity roll rate Vehicle body center of gravity pitch angle Vehicle body center of gravity pitch angle acceleration And the actual control quantity of the magnetorheological damper of the j-th wheel. and the rate of change of actual controlled quantity Establish the actual output matrix of the control quantity and the matrix of actual output change rate of control quantity The dimensions are 4×1 and 4×1 respectively, where the car corner index is... Steps 1.1 and 1.2 are used to provide unified parameters and state inputs for subsequent dynamic modeling, discretization, and robust control.

[0025] Step 1.3: Establish the output damping force of the actuator of the j-th wheel of the semi-active suspension of the car. With control quantity This relationship ensures that the controller uses control quantities such as current or valve position as input, achieving a control closed loop that is feasible for engineering applications. This embodiment employs a hyperbolic tangent magnetorheological model to represent the control quantity... Using the damper mechanics formula in equation (1) as the input, the damping force output by the magnetorheological damper of the j-th wheel is obtained. : (1) In equation (1), , , These represent the three characteristic parameters of the magnetorheological vibration damper. The parameter representing the adjustment rate of the magnetorheological damper. This represents the operating state of the magnetorheological damper of the j-th wheel, and , It is a constant. This represents the suspension speed of the j-th wheel.

[0026] Step 2, as follows Figure 3 The schematic diagram shown illustrates the principle flow of the discrete modeling layer, based on the output damping force of the actuator. With control quantity The relationship will control the damping force output by the actuator of the j-th wheel of the semi-active suspension of the car. Converted into control quantity for magnetorheological vibration dampers It is expressed using a spatial state equation, laying the foundation for subsequent discretization and sampling uncertainty modeling.

[0027] Step 2.1: Use equation (2) to establish a two-degree-of-freedom suspension model for the j-th wheel: (2) Step 2.2: Construct the state vector of the j-th wheel. and the derivative of the state vector of the j-th wheel Thus, equation (3) can be written as equation (4), where T represents transpose. This represents the suspension speed of the j-th wheel. Let represent the tire travel of the j-th wheel, and , This represents the tire speed of the j-th wheel. This represents the corner suspension acceleration of the j-th wheel. Let represent the tire acceleration of the j-th wheel.

[0028] (3) In equation (3), Let j be the state matrix of the j-th wheel, with dimensions 4×4. This is the damping force input matrix for the j-th wheel, with dimensions 4×1. This is the road surface disturbance input matrix for the j-th wheel, with dimensions 4×1. Let represent the road disturbance of the j-th wheel, and , Let represent the road surface disturbance acceleration of the j-th wheel.

[0029] Step 2.3, Substituting into equation (3), we can then use equation (4) to establish the control quantity for the actuator controlling the j-th wheel. The spatial state equation determines the output damping force of the actuator controlling the j-th wheel of the semi-active suspension of the car. This is converted into a control quantity for controlling the magnetorheological damper of the j-th wheel. : (4) In equation (4), This represents the equivalent state matrix of the j-th wheel after the actuator model is introduced. Its dimensions are 4×4. Let represent the control input matrix for the j-th wheel, and Its dimensions are 4×1. Let represent the constant term matrix of the j-th wheel, and Its dimensions are 4×1.

[0030] Step 3: Discretize the spatial state equation within the sampling jitter interval and the actuator gain interval, and construct a multi-cell vertex model at the four corners of the vehicle body to obtain the multi-cell vertex set. This step integrates sampling uncertainty and actuator gain into the same robust control framework, enabling a single structured controller to cover all operating conditions within the interval. Step 3.1: Set the sampling jitter time and actuator gain ,in, Indicates the minimum sampling duration. Indicates the maximum sampling duration. Indicates the minimum actuator gain. Indicates the maximum actuator gain; for The spatial state equations within the wheel are discretized, and the parameter discrete model of the j-th wheel is constructed using equation (5): (5) In equation (5), Let represent the state vector of the j-th wheel of the semi-active suspension at time k. Let represent the state vector of the j-th wheel of the semi-active suspension at time k+1. Indicates that the j-th wheel is in The discrete state transition matrix within, and The dimension is 4×4, and I represents the identity matrix. Indicates that the j-th wheel is in The discrete control input matrix within, and The dimension is 4×1. Indicates that the j-th wheel is in The discrete road surface disturbance input matrix has a dimension of 4×1. Let represent the road disturbance of the j-th wheel at time k. Indicates that the j-th wheel is in The discrete constant term matrix within it has a dimension of 4×1.

[0031] Step 3.2, in Discrete vertex set ,in, This represents the sampling duration corresponding to the l-th vertex. This represents the actuator gain corresponding to the l-th vertex, and N represents the total number of vertices; Construct the first using equation (6) Multicell Vertex The multi-cell vertex model at the four corners of the lower body is used to obtain the multi-cell vertex set. ,in, The matrix representing the corner of the j-th wheel: (6) In equation (6), Indicates the first Each weighting coefficient Indicates the first The j-th wheel corresponding to the multicell vertex is in The discrete state transition matrix within, the j-th wheel in Discrete control input matrix within, j-th wheel in Discrete road surface disturbance input matrix within, j-th wheel in The discrete constant term matrices within the matrix have dimensions of 4×4, 4×1, 4×1, and 4×1, respectively.

[0032] Step 4: Based on vehicle parameters and vehicle state information, construct the weighted performance output matrix for the j-th wheel. Low-frequency attitude matrix of the whole vehicle Thus, the overall performance matrix is ​​obtained. This results in overall performance indicators that combine smoothness, handling stability, and energy efficiency. Step 4.1: Define the performance output matrix of the j-th wheel using equation (7). Its dimensions are 5×1: (7) In equation (7), T represents transpose. This represents the vehicle body acceleration of the j-th wheel. This represents the maximum displacement of the actuator of the j-th wheel. This indicates the energy consumption of a car's semi-active suspension. This represents the rate of change of the control quantity of the actuator. It represents the acceleration due to gravity.

[0033] Step 4.2: Use equation (8) to obtain the low-frequency vehicle body bandpass weight. : (8) In equation (8), s represents the Laplace complex frequency domain variable. This represents the low-frequency weighting amplitude coefficient. The center angular frequency represents the low-frequency vehicle body bandpass weight, and , This indicates the center angular frequency of the low-frequency vehicle body bandpass weight. Indicates the low-frequency bandwidth parameter; Step 4.3: Use equation (9) to obtain the high-frequency tire bandpass weight. : (9) In equation (9), Represents the high-frequency weighting amplitude coefficient. The center angular frequency represents the high-frequency tire bandpass weight. The center angular frequency represents the high-frequency tire bandpass weight. Indicates high-frequency bandwidth parameters; and By assigning different frequency weights to vehicle posture (low frequency) and road feel (high frequency), H∞ control can automatically make a more reasonable trade-off during optimization, rather than relying on threshold switching or simple superposition.

[0034] Step 4.4: Use equation (10) to obtain the weighted performance output matrix of the j-th wheel, which has a dimension of 5×1: (1) In equation (10), This represents the construction of a 5×5 diagonal matrix, where I is the identity matrix. The control variable energy consumption weight is represented, and , The energy consumption weight represents the rate of change of the control input of the actuator, and , This represents the weighting of the adjustment parameter, used to adjust the intensity of the penalty applied to the rate of change of the control variable. , The unit delay operator for discrete systems.

[0035] Step 4.5: Discretize the continuous weighting element using equation (11) at the sampling period to construct a low-frequency attitude matrix with a dimension of 3×1, thereby constructing the overall performance matrix. Its dimension is 23×1, and T represents transpose: (11) In equation (11), These represent the weights for the vehicle's vertical, pitch, and roll attitudes, respectively. This indicates the low-frequency attitude state input.

[0036] Step 5: Establish the structured gain matrix K and the overall vehicle state matrix respectively. Thus, the overall state matrix of the vehicle is constructed. With the control quantity theoretical output matrix The structured control law relation, such as Figure 4 The schematic diagram of the robust control layer principle shown below implements system control: Step 5.1: Define the structured gain matrix K using equation (12), which has a dimension of 4×22, resulting in fewer controller parameters, better calibrability, and easier implementation on the ECU. (12) In equation (12), This represents the vehicle body state gain matrix, which has dimensions of 4×6. This represents the vehicle corner state feedback gain matrix, which has a dimension of 4×16. This indicates the construction of a diagonal block splicing matrix. Indicates the modal compensation gain. , , , These represent the local feedback gain matrices for the front left, front right, rear left, and rear right corners of the vehicle body, respectively, each with a dimension of 1×6. Let represent the geometrical distribution matrix, and we have: Step 5.2: Obtain the vehicle body attitude state matrix using equation (14). With the vehicle angle state matrix The relationship is used to establish the overall vehicle state matrix. Its dimensions are 16×1, where, express transpose: (14) In equation (14), The linear mapping matrix makes the vehicle body attitude state matrix... It can be calculated from the vehicle angle state matrix, which has a dimension of 6×16.

[0037] Step 5.3: Use equation (15) to establish the overall vehicle state matrix χ and the theoretical output matrix of the control quantity. The structured control law relation, where The total dimension is 22×1, and the total dimension is 4×1. This represents the theoretical control quantity of the semi-active shock absorber for the j-th wheel: (15) Step 6: Based on the multi-cell vertex model at the four corners of the vehicle body, establish a discrete multi-cell model of the whole vehicle, and solve the discrete multi-cell model of the whole vehicle through H∞ optimization control, thereby obtaining a structured gain matrix that satisfies the robust performance and constraint conditions on the set of multi-cell vertices. The theoretical output matrix of the control quantity is obtained using the control law relationship in step 5. The actual output matrix of the control quantity is obtained by applying constraints. An actuator used to apply electronically controlled semi-active suspension to a vehicle.

[0038] Step 6.1: To incorporate the influence of constant terms into robust integrated control, calculate the vehicle-enhanced road disturbance at time k according to equation (5). The dimension is 5×1, and the j-th wheel is in The whole vehicle augmented disturbance matrix Its dimensions are 16×5; Will In The augmented perturbation vertex merged into the j-th wheel This allows for the construction of a merged set of multiple cell vertices for the entire vehicle. .

[0039] Step 6.2: In the merged set of vertices of the entire vehicle multicell Establish a discrete multicell model of the whole vehicle: (16) In equation (16), This represents the state matrix of the entire vehicle at time k. This represents the state matrix of the entire vehicle at time k+1. Let represent the theoretical control output matrix of the vehicle at time k, with dimensions 4×1. Indicates the whole vehicle is The discrete state transition matrix within the matrix has a dimension of 16×16. Indicates the whole vehicle is The discrete control input matrix within the matrix has a dimension of 16×4.

[0040] Step 6.3: Output matrix based on total performance. For the performance output of robust control, construct a matrix to generate the total performance output. Performance output matrix , Their dimensions are 23×16 and 23×4 respectively, and they are in the merged multicell vertex set. Establish the following matrices to satisfy condition 1: the closed-loop system outputs a matrix to the total performance from road surface disturbance. The transfer gain satisfies the preset upper bound. This ensures that the robust performance remains controllable under uncertain sampling; and condition 2: satisfy the constraints of the magnitude and rate of change of the control quantity as shown in equations (18) and (19), and incorporate the constraints of the magnitude and rate of change of the control quantity as part of the LMI conditions into the control solution, rather than cutting them out afterward.

[0041] Using equation (17), condition 1 is established: the existence of a symmetric positive definite matrix is ​​expressed in equivalent LMI form. The dimension is 16×16, matrix Its dimension is 4×16, and the scalar γ > 0, such that for all merged multicell vertex sets... The following conditions apply: (17) In equation (17), T represents transpose and I represents identity matrix.

[0042] Condition 2 is established using equations (18) and (19): Amplitude: (18) Rate of change: (19) In equations (18) and (19), This represents the matrix showing the actual rate of change of the control output, and its dimension is 4×1. , This represents the actual output matrix of the control quantity at time k. This represents the actual output matrix of the control quantity at time k-1, and its dimension is 4×1. This represents the theoretical output rate of change matrix of the control quantity. , This represents the matrix showing the actual rate of change of the control output at time k. The actual output rate of change matrix of the control quantity at time k-1 has a dimension of 4×1. Step 6.4: Obtain the results by solving equations (17), (18), and (19). Thus, by using equation (20), the structured gain matrix that simultaneously satisfies the robust performance and constraint conditions at all vertices in a single solution is obtained. The theoretical control output matrix of the actuator is obtained using equation (15). : (20) Step 6.5: Obtain the actual output matrix of the control quantity using equation (21). This, together with the constraints in step 6.3, forms a "control design + safety margin constraint," ensuring that the control quantity never exceeds the limit. (twenty one) In equation (21), This represents the amplitude saturation function, used to limit the actual control quantity of the j-th wheel to satisfy... , This represents the minimum control quantity. Indicates the maximum control value. Let the rate of change constraint function be defined such that the theoretical control rate of change matrix satisfies the following condition at any given time. , where r is the threshold. represents the matrix showing the actual rate of change of the control output at time k-1. , Let represent two 4×4 decision variable matrices, and , I represents the identity matrix. This indicates that there are two weighting coefficients, and θ is the coefficient.

[0043] Each control cycle in this embodiment includes "sampling - constructing the overall vehicle state matrix χ - calculation". —Cut out "Input to actuator" – through the above steps, "safety constraints" are embedded into the control method. Current amplitude and current rate of change constraints are introduced in the H∞ synthesis stage, so that the system can still remain within the safety boundary under complex operating conditions and sampling anomalies.

[0044] Finally, the above steps complete one suspension control cycle. In actual conditions, the vehicle dynamic state variables are collected in real time during each control cycle to provide an appropriate output damping force.

[0045] Meanwhile, in this embodiment, a magnetorheological damper is used as the actuator. Without changing the control law structure, the magnetorheological damper can also be replaced with an electromagnetic valve type damper or other electronically controlled variable damping dampers. This only requires adjusting the mechanical model of the damper. Replace it with the corresponding damping force—control current or look up the table.

[0046] Meanwhile, the parameter range given in this embodiment (such as...) (etc.) can be adjusted according to different vehicle models and shock absorber characteristics, without affecting the basic concept and scope of protection of the present invention.

[0047] In summary, this invention utilizes the structure of existing electronically controlled semi-active suspension systems, incorporating sampling period uncertainty, the control quantity of the electronically controlled semi-active damper, and its rate of change constraints into a single robust control framework. This achieves coordinated suppression of low-frequency vehicle posture and high-frequency tire bounce, thereby enabling precise dynamic adjustment of the actuator's damping force within safe boundaries, comprehensively optimizing vehicle posture stability, ride comfort, and energy efficiency. This invention provides a control algorithm for an electronically controlled semi-active suspension system and also expands the development of control methods for complex real-world operating conditions and sampling uncertainties.

[0048] In this embodiment, an electronic device includes a memory and a processor. The memory stores a program that supports the processor in executing the methods described above, and the processor is configured to execute the program stored in the memory.

[0049] In this embodiment, a computer-readable storage medium stores a computer program, which is executed by a processor to perform the steps of the above method.

Claims

1. A robust control method for an automobile electrically controlled semi-active suspension, which is applied to an automobile semi-active suspension including a magneto-rheological damper and a solenoid valve type damper, characterized in that, The robust control method is performed according to the following steps: Step 1, obtain vehicle state information and vehicle parameters, and establish the actuator output damping force of the jth wheel of the automobile semi-active suspension relationship with the control quantity ​ Step 2, the output damping force of the actuator in relation to the actual control variable , a spatial state equation for the control variable of the actuator for controlling the jth wheel is established; Step 3, discretizing the space state equation, constructing a polytope vertex model of the four corners of the vehicle body, thereby obtaining a polytope vertex set; Step 4, based on vehicle parameters and vehicle state information, construct a weighted performance output matrix of the jth wheel with the low frequency attitude matrix of the whole vehicle , thus obtaining the total performance matrix ; Step 5, structured gain matrix K and the total state matrix of the vehicle are established respectively , and the total state matrix of the vehicle is constructed , and the structured control law relationship of the control variable theoretical output matrix is established Step 6, based on the four-corner polytope model of the vehicle body, a discrete polytope model of the whole vehicle is established, and the discrete polytope model of the whole vehicle is solved through H∞ optimization control, so as to obtain a structured gain matrix on the polytope vertex set which satisfies the robust performance and constraint conditions ; Based on structured gain matrix , the control variable theoretical output matrix is obtained by using the structured control law relationship equation , and the control variable actual output matrix is obtained after constraint , which is used to apply the actuator of the electrically controlled semi-active suspension of the automobile.

2. The robust control method of an automobile electrically controlled semi-active suspension according to claim 1, wherein Step 1 is performed according to the following steps: Step 1.

1. Obtain vehicle parameters, including: horizontal distance of front axle to center of mass horizontal distance of rear axle to center of mass front wheel track rear wheel track unsprung mass of jth wheel spring rate of jth wheel stiffness of jth wheel suspension damping of jth wheel suspension maximum travel of jth wheel where, denotes the index of the wheel, and denotes the front left, front right, rear left, and rear right, respectively; Step 1.2, obtaining vehicle state information, including: road vertical displacement of the jth wheel , speed and acceleration of the jth wheel, body vertical displacement of the jth wheel , vertical displacement of the jth wheel , suspension stroke of the jth wheel , body mass center vertical displacement , body mass center vertical speed , body mass center roll angle , body mass center roll angular speed , body mass center pitch angle , body mass center pitch angular acceleration , control amount of the actuator of the jth wheel and its rate of change ; thus establishing the control amount actual output matrix and the control amount actual output rate of change matrix ; When the actuator of the automobile semi-active suspension is a magneto-rheological damper, let the coil current of the magneto-rheological damper of the jth wheel ; When the actuator of the automobile semi-active suspension is an electromagnetic valve shock absorber, the valve opening of the electromagnetic valve shock absorber of the jth wheel is the valve opening of the electromagnetic valve shock absorber of the jth wheel ; Step 1.3, establishing an output damping force of an actuator of the jth wheel of the vehicle semi-active suspension in relation to the control variable : When the actuator of the automobile semi-active suspension is a magneto-rheological damper, the damping force of the magneto-rheological damper output of the jth wheel is obtained by using formula (1) : (1) In formula (1), , , respectively represent three characteristic parameters of the magneto-rheological damper, represents a parameter of the magneto-rheological damper adjustment rate, represents the working state of the magneto-rheological damper of the jth wheel, and , is a constant, represents the suspension speed of the jth wheel; When the actuator of the automobile semi-active suspension is the electromagnetic valve type shock absorber, the damping force of the electromagnetic valve type shock absorber output of the jth wheel is obtained by using formula (2) : (2) In formula (2), represents the minimum damping force of the electromagnetic valve type shock absorber, represents the maximum damping force of the electromagnetic valve type shock absorber.

3. The robust control method of an automobile electrically controlled semi-active suspension according to claim 2, wherein Step 2 is performed according to the following steps: Step 2.1, using formula (3) to establish a two-degree-of-freedom suspension model of the jth wheel: (3) Step 2.

2. Constructing the state vector of the jth wheel and the derivative of the state vector of the jth wheel so that equation (3) is written as equation (4), where T denotes transpose, denotes the suspension velocity of the jth wheel, denotes the tire travel of the jth wheel, and , denotes the tire velocity of the jth wheel, denotes the steering suspension acceleration of the jth wheel, denotes the tire acceleration of the jth wheel: (4) In formula (4), is a state matrix of the jth wheel, is a damping force input matrix of the jth wheel, is a road disturbance input matrix of the jth wheel, denotes a road disturbance of the jth wheel, and , denotes a road disturbance acceleration of the jth wheel: (4a) (4b) (4c) Step 2.3, the into equation (4) to establish the control quantity for controlling the actuator of the jth wheel by equation (5) spatial state equation: (5) In formula (5), denotes the equivalent state matrix of the jth wheel after introducing the actuator model, and , denotes the control amount input matrix of the jth wheel, and , denotes the constant term matrix of the jth wheel, and .

4. The robust control method of an automobile electrically controlled semi-active suspension according to claim 3, wherein Step 3 is performed according to the following steps: Step 3.1, Set sampling jitter time and actuator gain wherein, denotes a minimum sampling duration, denotes a maximum sampling duration, denotes a minimum actuator gain, denotes a maximum actuator gain; right The spatial state equations within the wheel are discretized, and the parameter discrete model of the j-th wheel is constructed using equation (6): (6) in formula (6), denotes the state vector of the jth wheel of the semi-active suspension of the car at the kth time instant, denotes the state vector of the jth wheel of the semi-active suspension of the car at the k+1th time instant, denotes the discrete state transition matrix of the jth wheel within and , I denotes the identity matrix, denotes the discrete control input matrix of the jth wheel within and , and , denotes the discrete road disturbance input matrix of the jth wheel within and , and denotes the road disturbance of the jth wheel at the kth time instant, denotes the discrete constant term matrix of the jth wheel within and , and ; Step 3.2, in taking a discrete set of vertices wherein, denotes the sampling duration corresponding to the denotes the actuator gain corresponding to the denotes the actuator gain corresponding to the denotes the actuator gain corresponding to the The first polytope vertex is constructed by using formula (7) The polytope vertex model of the lower body four corners is obtained The polytope vertex set is obtained Wherein, The steering angle matrix of the jth wheel is represented (7) In formula (7), denotes the weight coefficient, denotes the discrete state transition matrix of the jth wheel within , the discrete control input matrix of the jth wheel within and , the discrete road disturbance input matrix of the jth wheel within , the discrete constant term matrix of the jth wheel within and .

5. The robust control method of an automobile electrically controlled semi-active suspension according to claim 4, wherein Step 4 is performed according to the following steps: Step 4.1, define the performance output matrix of the jth wheel using formula (8) : (8) In formula (8), T denotes transposition, denotes the body acceleration of the jth wheel, denotes the maximum displacement of the actuator of the jth wheel, denotes the energy consumption of the car semi-active suspension, denotes the control amount change rate of the actuator, denotes the gravitational acceleration; Step 4.2, obtaining low frequency body band pass weights using formula (9) : (9) In formula (9), s represents a Laplace complex frequency variable, represents a low frequency weight amplitude coefficient, represents a center angular frequency of a low frequency body band pass weight, and , represents a center angular frequency of a low frequency body band pass weight, represents a low frequency bandwidth parameter; Step 4.

3. Obtain high frequency tire band pass weights using formula (10) : (10) In formula (10), denotes a high frequency weight amplitude coefficient, denotes a center angular frequency of the high frequency tire band pass weight, denotes a center angular frequency of the high frequency tire band pass weight, denotes a high frequency bandwidth parameter; Step 4.4, obtain the weighted performance output matrix of the jth wheel using formula (11) : (11) In formula (11), denotes a diagonal matrix, denotes an energy consumption weight of a control amount of an actuator, denotes an energy consumption weight of a control amount change rate of an actuator; Step 4.5, obtain the low frequency attitude matrix of the whole vehicle using formula (12) , thereby constructing the total performance matrix : (12) In formula (12), respectively represent the weights of the vertical, pitch and roll attitudes of the vehicle body.

6. The robust control method of an automobile electrically controlled semi-active suspension according to claim 1, wherein Step 5 is performed according to the following steps: Step 5.1, using formula (13) to define a structured gain matrix K: (13) in formula (13), denotes the body state feedback gain matrix, denotes the body angle state feedback gain matrix, denotes the construction of a diagonal block splicing matrix, denotes the modal compensation gain, 、 、 、 denote the local feedback gain matrix of the body left front corner, right front corner, left rear corner, right rear corner, respectively, denotes the geometric distribution matrix, and has: (13a) Step 5.2, the body attitude state matrix is obtained using formula (14) and the relationship of the vehicle angle state matrix , thus the total state matrix of the vehicle is established wherein, denotes the transpose of ​ (14) In formula (14), is a linear mapping matrix; Step 5.3, Establish the overall vehicle state matrix with the equation of (15) The structured control law relationship with the control quantity theoretical output matrix : (15) In formula (15), , represents the theoretical control amount of the semi-active damper of the jth wheel.

7. The robust control method of an automobile electrically controlled semi-active suspension according to claim 1, wherein Step 6 is performed according to the following steps: Step 6.1, Calculate the augmented road disturbance of the whole vehicle at time k , the augmented disturbance matrix of the whole vehicle in ;​ Merge the in the augmented disturbance vertex of the jth wheel corresponding to the first polyhedral vertex pair , so as to construct the merged whole vehicle polyhedral vertex set ; Step 6.2, in the combined vehicle polytope vertex set The vehicle discrete polytope model is built above: (16) In formula (16), denotes the state matrix of the whole vehicle at time k, denotes the state matrix of the whole vehicle at time k+1, denotes the theoretical control output matrix of the whole vehicle at time k, denotes the discrete state transition matrix of the whole vehicle within denotes the discrete control input matrix of the whole vehicle within denotes the discrete state transition matrix of the whole vehicle within denotes the discrete control input matrix of the whole vehicle within Step 6.3, with total performance output matrix For the performance output of robust control, two performance output matrices are constructed , and conditions 1 and 2 are established on the merged set of polytope vertices respectively; The condition 1 is that the transfer gain of the closed loop system to the road disturbance to the total performance output matrix satisfies a preset upper bound and is established by using formula (17): (17) In formula (17), T denotes transposition, I denotes a unit matrix, is a symmetric positive definite matrix, is a matrix, and , is a scalar, and ; The condition 2 is to satisfy the amplitude and rate constraints of the control quantity, and is established by using formula (18) and formula (19): Amplitude: (18) Rate of change: (19) in formula (18) and formula (19), denotes a control actual output variation rate matrix, and , denotes a control actual output matrix at k time, denotes a control actual output matrix at k-1 time, denotes a control theoretical output variation rate matrix, , denotes a control actual output variation rate matrix at k time, denotes a control actual output variation rate matrix at k-1 time, , denote respectively two decision variable matrices, and , , I denotes a unit matrix, denote respectively two weight coefficients, and θ is a coefficient; Step 6.4, obtain the structured gain matrix satisfying the robust performance and the constraint condition by solving equation (17), equation (18), equation (19) , so as to obtain the structured gain matrix satisfying the robust performance and the constraint condition by equation (20) , and obtain the actuator theoretical control output matrix by equation (15) : (20) Step 6.

5. Obtain the control quantity actual output matrix using formula (21) : (21) In formula (21), denotes an amplitude saturation function for limiting the actual control amount of the jth wheel satisfies , denotes a minimum control amount, denotes a maximum control amount, denotes a rate limit function such that the rate of change of the theoretical control amount matrix satisfies at any time, and r is a threshold value.

8. An electronic device comprising a memory and a processor, characterized in that The memory is used to store a program supporting the processor to execute the robust control method in any one of claims 1-7, and the processor is configured to execute the program stored in the memory.

9. A computer-readable storage medium having stored thereon a computer program, characterized in that The computer program is executed by the processor to perform the steps of the robust control method in any one of claims 1-7.

Images