Vehicle posture estimation and active control method

Abstract

This invention proposes a method for vehicle posture estimation and active control, comprising: step 1: acquiring vehicle state and parameter information; step 2: fusing information obtained from vehicle kinematic model calculations and sensor integration using an adaptive fusion filtering algorithm to obtain vehicle posture measurement information; step 3: acquiring vehicle posture estimation information; step 4: calculating the feedforward quantity required for vehicle posture control; step 5: determining the optimal control quantity for vehicle posture; and step 6: based on the optimal control quantity for vehicle posture, achieving optimal distribution of active forces of the suspension system according to the principle of optimal tire adhesion utilization. This invention features good real-time decoupling, high control accuracy, and optimal tire adhesion utilization.

Description

Technical Field

[0001] The present invention belongs to the field of vehicle active or semi-active suspension control, and in particular relates to a vehicle body posture estimation and active control method thereof. Background Art

[0002] The suspension system is a key component of the vehicle chassis, playing a vital role in ensuring a smooth ride and handling stability. Compared to passive suspension, active suspension proactively adjusts suspension force, improving vehicle posture and dynamic driving performance. While driving, improper driver control or road bumps can exacerbate vehicle pitch, roll, and vertical motion, further adversely affecting ride comfort and handling stability. Therefore, the key and challenging aspect of vehicle posture control technology is how to quickly and accurately adjust the vehicle's posture to the ideal range using active suspension.

[0003] To ensure the accuracy of active adjustment of the vehicle body posture, it is first necessary to obtain the vehicle's motion state in real time and accurately to provide support for active control of the vehicle body posture; secondly, based on the vehicle body posture estimation results, it is necessary to calculate the forces of each active suspension through optimal control means, so as to adjust the vehicle body posture to a reasonable range and improve ride comfort and handling stability.

[0004] Therefore, the accuracy of vehicle posture estimation and vehicle posture control directly affect the vehicle's riding and driving performance.

[0005] Patent No. CN117284035A discloses a vehicle body posture compensation active control method and system. This method calculates the optimal vertical suspension force based on a model prediction method and uses a fuzzy controller to compensate for each active suspension force. However, this method does not consider the impact of sensor noise on vehicle body posture information acquisition. Furthermore, during the vehicle posture compensation process, it does not consider the coupling relationship between the vehicle body's vertical, pitch, and roll motions, nor the impact of the active suspension force on tire adhesion utilization.

[0006] Patent No. CN105974917A discloses a high-level vehicle posture control method based on pitch and roll force compensation. This method calculates the initial value of the semi-active suspension control force using a skyhook damping algorithm, obtains pitch and roll compensation forces from the vehicle's lateral and longitudinal accelerations, and finally achieves vehicle posture control by adjusting the magnetorheological suspension damping force using current. However, this method does not control the vertical movement of the vehicle body, and since it is based on a semi-active suspension, the control range is small and the vehicle posture adjustment capability is limited.

[0007] In summary, current vehicle posture control methods have the problem of low vehicle posture signal accuracy due to sensor noise. In addition, most methods only consider vertical, pitch, and roll motions separately or individually, without comprehensive consideration of the coupling relationship between the three, thus failing to ensure optimal control of the actual vehicle posture. Summary of the Invention

[0008] The present invention aims to provide a vehicle body posture estimation and active control method that has the characteristics of good real-time decoupling, high control accuracy, and optimal tire adhesion utilization. The technical solution adopted is as follows:

[0009] A vehicle body posture estimation and active control method includes the following steps:

[0010] Step 1: Obtain vehicle status information and vehicle parameter information;

[0011] Step 2: Obtain vehicle body posture measurement information:

[0012] Based on vehicle status information and vehicle parameter information, the adaptive fusion filtering algorithm is used to fuse the signals obtained from the vehicle kinematic model calculation and sensor integration to output vehicle posture measurement information, including vehicle pitch angle, vehicle roll angle, and vehicle vertical displacement;

[0013] Step 3: Get vehicle body posture estimation information:

[0014] Based on the vehicle posture measurement information obtained in step 2 and the vertical seven-degree-of-freedom vehicle dynamics model that considers the influence of lateral acceleration and longitudinal acceleration, the vehicle posture is estimated using a Kalman filter algorithm, and the vehicle posture estimation information is output, including the vehicle pitch angle, the vehicle pitch angular velocity, the vehicle roll angle, the vehicle roll angular velocity, the vehicle vertical displacement, and the vehicle vertical velocity;

[0015] Step 4: Calculate the feedforward amount required for vehicle posture control:

[0016] Based on the vehicle posture estimation information obtained in step 3 and the fuzzy PID control algorithm, the feedforward quantity required for vehicle posture control is calculated, including the additional vertical force, the additional pitching moment, and the additional roll moment of the vehicle body;

[0017] Step 5: Calculate the optimal control value of the vehicle body posture

[0018] Based on a vertical seven-degree-of-freedom vehicle dynamics model that considers the effects of lateral and longitudinal accelerations, and using the feedforward values ​​obtained in step 4 as constraints, a model predictive control algorithm is used to output the optimal control variables for the vehicle body posture, including the optimal additional vertical force, the optimal additional pitching moment, and the optimal additional roll moment.

[0019] Step 6: Based on the optimal control value of the vehicle body posture, select the tire adhesion utilization rate as the control target and solve the four optimal active suspension forces.

[0020] Preferably, the vehicle status information in step 1 includes:

[0021] The vehicle's longitudinal acceleration, lateral acceleration, vertical acceleration, pitch angular velocity, and roll angular velocity are measured by the inertial measurement unit (IMU), and the dynamic deflection of the four suspensions is measured by the suspension height sensor.

[0022] Among them, the IMU is installed in the center of the chassis. The geometric relationship between the center of the chassis and the center of mass of the vehicle is used to convert the vehicle center of mass position status information. The four suspension height sensors are installed at the connection between the body and the upper end of the suspension.

[0023] The vehicle parameter information includes:

[0024] The vehicle's sprung and unsprung mass; the vehicle's mass is determined at the factory; the sprung mass is the mass of the vehicle body, and the unsprung mass is the mass of the suspension and tires;

[0025] Vehicle pitch moment of inertia, vehicle roll moment of inertia;

[0026] The position of the vehicle's center of mass, including the height of the vehicle's center of mass, the distance from the front and rear axles to the vehicle's center of mass, and the distance from the left and right wheels to the vehicle's center of mass;

[0027] Vehicle track width, suspension stiffness, suspension damping and tire stiffness.

[0028] Preferably, step 2 specifically includes the following steps:

[0029] Step 2-1: Calculate the vehicle body posture based on the vertical seven-degree-of-freedom vehicle kinematic model, vehicle parameter information, and suspension dynamic deflection information;

[0030] Step 2-2: Estimate the vehicle body posture through integration based on the vehicle body pitch angular velocity, vehicle body roll angular velocity, and vehicle body vertical acceleration information;

[0031] Step 2-3: Determine the adaptive law of the adaptive fusion filtering algorithm based on the vertical acceleration information;

[0032] The specific formula of the adaptive law of the fusion algorithm is:

[0033]

[0034] Among them, ω H is the cutoff frequency ω c The upper bound of

[0035] ω Lis the cutoff frequency ω c The lower bound of

[0036] Δ z is the vertical acceleration a z Deviation from the acceleration due to gravity g;

[0037] Δ dz is Δ z The derivative of a dz for a z The derivative of

[0038] Step 2-4: The vehicle body posture obtained in step 2-1 and step 2-2 is fused through the adaptive fusion filter algorithm and the adaptive law to obtain the vehicle body posture measurement information;

[0039] The specific formula of the adaptive fusion filtering algorithm is:

[0040]

[0041] Among them, ω c is the cutoff frequency of the high / low frequency filter;

[0042] θ1, φ1, and z s1 are the body pitch angle, body roll angle and body vertical displacement calculated based on the integration of the body pitch angular velocity, body roll angular velocity and body vertical acceleration respectively;

[0043] θ2, φ2, and z s2 They are the body pitch angle, body roll angle and body vertical displacement calculated based on the vehicle kinematic model.

[0044] Preferably, the vertical seven-degree-of-freedom vehicle kinematic model in step 2-1 is:

[0045]

[0046] Among them, l w The wheelbase of the vehicle is the distance between the two front wheel centers, or the distance between the two rear wheel centers;

[0047] l f is the distance from the front axle to the center of mass of the vehicle;

[0048] l r is the distance from the rear axle to the center of mass of the vehicle;

[0049] z sfl ,z sfr ,z srl ,z srr are the vertical displacements of the four endpoints of the suspension on the vehicle body; the endpoints are the connections between the upper end of the suspension and the vehicle body;

[0050] fl, fr, rl, rr represent the left front wheel, right front wheel, left rear wheel and right rear wheel respectively;

[0051] s represents the sprung position; z represents the z-direction displacement;

[0052] θ represents the vehicle body pitch angle;

[0053] Indicates the body roll angle;

[0054] The calculation formula for the displacement of the upper end point of the suspension is:

[0055]

[0056] Among them, z ufl ,z ufr ,z url ,z urr The displacements of the unsprung masses corresponding to the left front wheel, right front wheel, left rear wheel, and right rear wheel, respectively;

[0057] u means unsprung;

[0058] h fl ,h fr ,h rl ,h rr The suspension dynamic deflections corresponding to the left front wheel, right front wheel, left rear wheel and right rear wheel are measured by the suspension height sensor.

[0059] Preferably, step 3 specifically includes the following steps:

[0060] Step 3-1: Consider the influence of vehicle lateral acceleration and longitudinal acceleration to establish the system state space equation:

[0061] The vertical seven-degree-of-freedom vehicle dynamics model is as follows:

[0062]

[0063] Where x = [z s θφz ufl z ufr z url z urr ] T is the state vector;

[0064] M is the mass matrix; C is the damping matrix; K is the stiffness matrix; F is the force matrix;

[0065] z s Indicates the vertical displacement of the vehicle body, that is, the vertical displacement of the upper end point of the suspension;

[0066] The expressions of each coefficient matrix are:

[0067] M=diag{m s ,I y ,I x ,m ufl ,m ufr ,m url ,m urr}

[0068]

[0069]

[0070] C 22 =diag{c fl ,c fr ,c rl ,c rr}

[0071]

[0072]

[0073] K 22 =diag{k fl ,k fr ,k rl ,k rr}

[0074]

[0075] Among them, m s ,m ufl ,m ufr ,m url ,m urr The sprung mass and the four unsprung masses are listed in order;

[0076] I x ,I y are the pitch moment of inertia and the roll moment of inertia, respectively;

[0077] x represents the x-axis; y represents the y-axis;

[0078] c fl ,c fr ,c rl ,c rr is the damping of each suspension;

[0079] k fl ,k fr ,k rl ,k rr is the stiffness of each suspension;

[0080] k utfl ,k utfr ,k utrl ,k utrr is the tire stiffness of each wheel;

[0081] ut means tire;

[0082] F afl ,F afr ,F arl ,F arr It is the main force of each suspension, generated by the active suspension actuator;

[0083] a means active;

[0084] F z ,M θ ,M φ They are the additional vertical force, additional pitching moment and additional rolling moment generated by the main force of the suspension.

[0085] F zfl ,F zfr ,F zrl ,F zrr is the vertical force of the wheel;

[0086] z ufl ,z ufr ,z url ,z urr is the unsprung mass displacement corresponding to each wheel;

[0087] z rfl ,z rfr ,z rrl ,z rrr is the road displacement of each wheel;

[0088] a x Indicates the longitudinal acceleration of the vehicle body;

[0089] a y Indicates the lateral acceleration of the vehicle body;

[0090] h represents the height of the vehicle's center of mass from the ground;

[0091] Step 3-2: Based on the state vector in step 3-1, list the state transfer matrix and observation matrix; solve the Kalman filter gain based on the state transfer matrix and observation matrix, and complete the vehicle posture estimation by combining the vehicle posture measurement information;

[0092] x k+1 =A k x k +B k u k +f k +Wk ,y k =H k x k +V k

[0093] Among them, x k =[z s,k θ k φ k z ufl,k z ufr,k z url,k z urr,k z s,k+1 θ k+1 φ k+1 z ufl,k+1 z ufr,k+1 z url,k+ 1z urr,k+1 ] T is the state vector;

[0094] u k =[F z,k M θ,k M φ,k ] T is the control vector

[0095] y k =[θ k+1 φ k+1 h fl,k h fr,k h rl,k h rr,k ] T is the observation vector;

[0096] Among them, h fl ,h fr ,h rl ,h rr is the suspension dynamic deflection;

[0097] W k is the process noise, V k To measure noise;

[0098] k represents the sampling time;

[0099] A k represents the state transfer matrix, B k represents the state control matrix;

[0100] H k represents the observation matrix, f k Represents the external input;

[0101] B k =[O 7x3 I3 B13 ] T ,

[0102]

[0103] Preferably, in step 4, a vertical fuzzy PID controller, a pitch fuzzy PID controller and a roll fuzzy PID controller are used to control the vehicle body respectively;

[0104] For the vertical fuzzy PID controller, select the target vertical displacement z d The error Δz between the actual vertical displacement z and the error change rate As input; its output is additional vertical force; d represents the target;

[0105] For the pitch fuzzy PID controller, select the target pitch angle θ d The error Δθ between the actual pitch angle θ and the error change rate As input; its output is additional pitching moment;

[0106] For the roll fuzzy PID controller, select the target roll angle φ d The error Δφ between the actual roll angle φ and the error change rate As input; its output is additional rolling moment;

[0107] The two variables in each of the above controllers are transformed into domain, fuzzy reasoning and defuzzification respectively, and finally the control parameters Δk are obtained respectively. P,z ,Δk I,z ,Δk D,z , Δk P,θ ,Δk I,θ ,Δk D,θ and

[0108] Δk P,φ ,Δk I,φ ,Δk D,φ , and the control parameter k in the original PID controller P0,z ,k I0,z ,k D0,z 、k P0,θ ,k I0,θ ,k D0,θ and k P0,φ ,k I0,φ ,k D0,φ Together they form the new parameter k P,z ,k I,z ,k D,z 、k P,θ ,k I,θ ,k D,θ and k P,φ,k I,φ ,k D,φ This enables the controller parameters to have adaptive adjustment capabilities, thereby improving the control effect of the system. The expressions of each PID parameter are as follows:

[0109]

[0110] The fuzzy set in the fuzzy control process is expressed as {negative large, negative medium, negative small, zero, positive small, positive medium, positive large}, abbreviated as: {NB, NM, NS, Z, PS, PM, PB};

[0111] The membership function and fuzzy rules are determined through simulation experiments and expressed through fuzzy surfaces. Finally, the additional vertical force, additional rolling moment and additional pitching moment actually required are calculated.

[0112] Preferably, the vehicle body posture model predictive control algorithm in step 5 specifically includes the following steps:

[0113] Step 5-1: Based on the vertical seven-degree-of-freedom vehicle dynamics model, list the state space equations of the system;

[0114]

[0115] Step 3-2 is the Kalman filter algorithm, which needs to consider noise. The equation here is in the state estimation category; the equation here does not need to consider noise and is in the control category, because the same system may manifest differently in different systems.

[0116] in, is the state variable of the system;

[0117] y(t)=[θφh fl h fr h rl h rr ] T is the output variable of the system, including pitch angle, roll angle and dynamic deflection of each suspension;

[0118] u(t)=[F z M θ M φ ] T are the control variables of the system, including the additional vertical force of the vehicle body, the additional pitching moment of the vehicle body, and the additional rolling moment of the vehicle body;

[0119] f(t) is the external input of the system f(t) = f k ;

[0120] t represents the current moment;

[0121] A, B, C, and D are the state space state matrix, input control matrix, output matrix, and output control matrix, respectively;

[0122] Step 5-2: Discretize the equation in step 5-1 to obtain the discrete system state equation:

[0123]

[0124] A d =A·T samp +I,B d =B·T samp ,C d =C,D d =D

[0125] Among them, A d Represents the discretized state space state matrix;

[0126] B d represents the discretized input control matrix;

[0127] C d Represents the discretized output matrix;

[0128] D d represents the discretized output control matrix;

[0129] T samp Indicates the sampling time;

[0130] x(k) represents the discretized state variable;

[0131] u(k) represents the discretized control variable;

[0132] y(k) represents the discretized output variable;

[0133] f(k) represents the external input after f(t) is discretized;

[0134] Step 5-3: Select the vehicle body pitch angle, body roll angle, and suspension dynamic deflection as the control targets of the model predictive controller, construct the objective function, and establish the constraints:

[0135]

[0136] Where J is the objective function value;

[0137] N P is the prediction step length;

[0138] N C To control the step length;

[0139] y(k+j|k) is the predicted value of the control output;

[0140] y ref (k+j|k) is the control output reference value, including the reference values ​​of pitch angle, roll angle and dynamic deflection of each suspension, that is, Figure 2 Target vehicle body posture in ;

[0141] u(k+j|k) is the control variable, including the additional vertical force of the vehicle body, the additional rolling moment of the vehicle body, and the additional pitching moment of the vehicle body;

[0142] k+j|k means predicting the value at time k+j based on the information at time k, j = 1, 2, ..., N P ;

[0143] u(k+j|k) represents the control input at time k+j; that is, the value of the control variable after discretization; j = 1, 2, ..., N C ;

[0144] u(u(k+j|k)) and y(y(k+j|k)) are vectors. u includes three quantities and y includes six quantities. These vectors correspond one to one in matrix operations.

[0145] Q is the system output weight coefficient matrix;

[0146] R is the system control weight coefficient matrix;

[0147] The constraints of the control quantity are:

[0148]

[0149] Among them, F z,fuzzy ,M θ,fuzzy ,M φ,fuzzy are the additional vertical force, additional pitching moment and additional rolling moment of the vehicle body determined by the fuzzy PID controller in step 4 respectively;

[0150] ΔF z ,ΔM θ ,ΔM φ The increments of the vehicle body additional vertical force, vehicle body additional pitching moment and vehicle body additional rolling moment;

[0151] The constraints controlling the increment are:

[0152]

[0153] Where ΔF z,min ,ΔF z,max ,ΔM θ,min ,ΔM θ,max ,ΔM φ,min ,ΔM φ,max are the minimum and maximum values ​​of each control variable, respectively, which are determined by the vehicle driving mode;

[0154] The output constraint can be expressed as:

[0155]

[0156] Among them, θ min and θ max is the minimum and maximum value of the vehicle body pitch angle;

[0157] φ min and φ max is the minimum and maximum value of the vehicle body roll angle;

[0158] h i,min and h i,max The maximum compression and maximum tension of the suspension dynamic deflection;

[0159] Step 5-4: Solve and obtain the optimal control quantity for each step

[0160] Preferably, step 6 specifically includes the following steps:

[0161] Step 6-1: Select tire adhesion utilization as the control target and transform the active suspension torque distribution problem into the following nonlinear programming problem:

[0162]

[0163] stAx=b

[0164] lb≤x≤ub

[0165]

[0166] Among them, F x,i ,F y,i ,F z,i are the longitudinal force, lateral force and vertical force of each wheel; i represents the wheel;

[0167] F x,i Calculated from the tire rotation equation;

[0168] F z,i The equations to be solved are listed based on the vertical seven-degree-of-freedom vehicle model, which include the corresponding active suspension forces;

[0169] F y,i Solve the equations listed by the magic formula, which contains Fz ,i ;

[0170] x=[F afl ,F afr ,F arl ,F arr ] Tis the force of the four active suspensions;

[0171] lb and ub are the minimum and maximum values ​​of the active suspension force, respectively;

[0172] Step 6-2: Solve the above nonlinear programming problem and finally obtain four optimal active suspension forces.

[0173] Compared with the prior art, the advantages of the present invention are:

[0174] 1. High Precision: The vehicle posture estimation method based on adaptive Kalman filtering proposed in this invention can simultaneously utilize the vehicle's longitudinal and lateral acceleration information, vertical seven-degree-of-freedom vehicle model information, inertial measurement unit (IMU) and suspension height sensor measurement information. Compared with existing technologies, the vehicle posture estimation accuracy is higher.

[0175] 2. Taking into account decoupling real-time performance, safety and comfort: The vehicle body posture active control method proposed in the present invention uses the additional forces and torques in all directions calculated based on the fuzzy PID controller as the feedforward of the MPC controller, thereby improving the decoupling real-time performance of the algorithm; at the same time, the method takes into account the vertical, pitch and roll movements of the vehicle body, and adjusts the tire adhesion utilization rate by changing the active force of the suspension, thereby ensuring the comfort and safety of vehicle driving. BRIEF DESCRIPTION OF THE DRAWINGS

[0176] Figure 1 Schematic diagram of the vehicle body posture estimation and active control method in the present invention;

[0177] Figure 2 This is a diagram of the vehicle body posture active control algorithm architecture in an embodiment of the present invention;

[0178] Figure 3 Schematic diagram of a vertical seven-degree-of-freedom vehicle dynamics model in an embodiment of the present invention;

[0179] Figure 4 2 is a comparison diagram of the simulation effect of vehicle body roll angle control in an embodiment of the present invention. DETAILED DESCRIPTION

[0180] The following is a more detailed description of the vehicle body posture estimation and active control method of the present invention, with reference to schematic diagrams. These schematic diagrams illustrate preferred embodiments of the present invention. It should be understood that those skilled in the art may modify the present invention described herein while still achieving the beneficial effects of the present invention. Therefore, the following description should be understood as generally known to those skilled in the art and is not intended to limit the present invention.

[0181] A vehicle body posture estimation and active control method, the process is as follows Figure 1 As shown, including:

[0182] Step 1: Get vehicle status and parameter information;

[0183] Step 2: Based on the adaptive fusion filtering algorithm, the signals calculated by the vehicle kinematic model and the sensor integration are fused to obtain the vehicle body posture measurement information;

[0184] Step 3: Based on the vertical seven-degree-of-freedom vehicle dynamics model that considers the influence of lateral and longitudinal acceleration and the vehicle posture measurement information obtained in step 2, the Kalman filter algorithm is used to estimate the vehicle posture;

[0185] Step 4: Adjust the vehicle suspension damping force based on the skyhook damping control algorithm, and calculate the additional vertical force, additional pitching moment, and additional rolling moment required by the vehicle based on the vertical fuzzy PID control algorithm, the pitch fuzzy PID control algorithm, and the roll fuzzy PID control algorithm.

[0186] Step 5: Based on the additional forces and moments obtained in step 4 as feedforward quantities, the vehicle posture model predictive control algorithm is used to obtain the additional vertical forces, pitching moments, and rolling moments actually required by the vehicle.

[0187] Step 6: Based on the additional force and torque calculated in step 5 and the optimal principle of tire adhesion utilization, the optimal distribution of the active force of the suspension system is achieved.

[0188] The control algorithm architecture diagram of steps 5 to 6 is as follows Figure 2 shown.

[0189] In step 3-2, the vehicle body posture is estimated based on the Kalman filter algorithm. The Kalman filter estimation process is as follows:

[0190] (1) Time update process:

[0191]

[0192] (2) Status update process:

[0193]

[0194] Among them, P k is the covariance matrix, K k is the Kalman gain, H k represents the observation matrix, Q is the process noise covariance matrix, and R is the measurement noise covariance matrix. In this example, Q = 0.01*eye(14), R = 0.01*eye(6), and are the estimated values ​​of the state vector at the previous moment and the current moment respectively, is the prior value of the state vector at this moment, is the prior value of the covariance matrix at this moment, y kis the output vector measurement value. eye is the unit diagonal matrix.

[0195] The vehicle body posture estimation results are obtained through the Kalman filter algorithm, including the vehicle body pitch angle, vehicle body pitch angular velocity, vehicle body roll angle, vehicle body roll angular velocity, vehicle body vertical displacement and vertical velocity, etc.

[0196] The specific process of step 4 is:

[0197] The specific ceiling damping algorithm is:

[0198] For any wheel suspension, the dynamic model of a quarter vehicle suspension under skyhook damping control is:

[0199]

[0200] Where M is one quarter of the sprung mass, m i is the unsprung mass, z s,i is the displacement of the upper end point of each suspension, z u,i is the displacement of the lower end point of each suspension, c sky,i is the ideal ceiling damping coefficient, is the calibration value;

[0201] z r,i is the displacement of the contact point between each wheel end and the road;

[0202] k i is the stiffness of each suspension, k t,i is the stiffness of each tire;

[0203] i=fl,fr,rl,rr represent the left front wheel, left rear wheel, right front wheel and right rear wheel respectively.

[0204] The dynamic model of a quarter vehicle suspension under actual adjustable damping is:

[0205]

[0206] Among them, c i is the actual variable damping coefficient of each suspension. The above two equations are equivalent, and we can get

[0207]

[0208] Then the damping of each suspension can be expressed as:

[0209]

[0210] Among them, c i is the actual damping of each suspension, C min and C max Represent the upper and lower limits of variable damping respectively. Finally, the damping size is adjusted by the input current i.

[0211] The specific fuzzy PID control algorithm for vehicle body vertical, pitch and roll is as follows:

[0212] For the vertical fuzzy PID controller, select the target vertical displacement z d The error Δz between the actual vertical displacement z and the error change rate As input quantity;

[0213] For the pitch fuzzy PID controller, select the target pitch angle θ d The error Δθ between the actual pitch angle θ and the error change rate As input quantity;

[0214] For the roll fuzzy PID controller, select the target roll angle φ d The error Δφ between the actual roll angle φ and the error change rate As input quantity;

[0215] The two variables in each of the above controllers are transformed into domain, fuzzy reasoning and defuzzification respectively, and finally the control parameters Δk are obtained respectively. P,z ,Δk I,z ,Δk D,z , Δk P,θ ,Δk I,θ ,Δk D,θ and Δk P,φ ,Δk I,φ ,Δk D,φ .

[0216] Compared with the control parameter k in the original PID controller P0,z ,k I0,z ,k D0,z 、k P0,θ ,k I0,θ ,k D0,θ and k P0,φ ,k I0,φ ,k D0,φ Together they form the new parameter k P,z ,k I,z ,k D,z 、k P,θ ,k I,θ ,k D,θ and k P,φ ,k I,φ ,k D,φ .

[0217] This enables the controller parameters to have adaptive adjustment capabilities, thereby improving the control effect of the system. The expressions of each PID parameter are as follows:

[0218]

[0219] The fuzzy set in the fuzzy control process is expressed as {negative large, negative medium, negative small, zero, positive small, positive medium, positive large}, abbreviated as: {NB, NM, NS, Z, PS, PM, PB}.

[0220] Where P represents the proportional unit of the PID controller;

[0221] D represents the differential unit of the PID controller;

[0222] I represents the integral unit of the PID controller;

[0223] The membership function and fuzzy rules are determined through simulation experiments and expressed through fuzzy surfaces. Finally, the additional vertical force, additional rolling moment and additional pitching moment actually required are calculated.

[0224] In step 5, the vehicle body pitch angle, body roll angle, and suspension dynamic deflection are selected as the control targets of the model predictive controller to improve the vehicle body posture while avoiding excessive changes in the additional forces in each direction. The objective function is constructed by combining the discrete system state equation.

[0225] When the vehicle is in comfort driving mode, due to the lower suspension stiffness and damping, the incremental constraint range of the vehicle body's additional vertical force, additional pitching moment, and additional rolling moment is reduced, thereby improving ride comfort.

[0226] When the vehicle is in sports driving mode, due to the greater suspension stiffness and damping, the incremental constraint range of the vehicle body's additional vertical force, additional pitching moment and additional roll moment is increased, thereby improving handling stability.

[0227] In step 6, the optimal distribution algorithm of the active force of the suspension system is specifically as follows:

[0228] First, through the vertical seven-degree-of-freedom vehicle model (such as Figure 3 Calculate the vehicle longitudinal force as shown:

[0229]

[0230] Then, the tire lateral force is calculated using the magic formula:

[0231]

[0232] Among them, B y is the tire stiffness factor for lateral force, C y is the tire curve shape factor, D y is the peak factor, E y is the curvature factor, α fl ,α fr ,α rl ,α rr are the four-wheel slip rates respectively, and the calculation formula is as follows:

[0233]

[0234] in, is the yaw rate, and δ is the front wheel turning angle.

[0235] v x Indicates the longitudinal speed of the vehicle;

[0236] v y Indicates the lateral speed of the vehicle;

[0237] Calculate the tire longitudinal force using the tire rotation equation:

[0238]

[0239] Among them, T dfl ,T dfr ,T drl ,T drr It is the four-wheel drive torque, obtained through the drive motor;

[0240] T bfl ,T bfr ,T brl ,T brr is the four-wheel braking torque, obtained through the brake controller;

[0241] R w is the wheel radius, I w is the wheel moment of inertia;

[0242] dri means drive;

[0243] is the wheel angular acceleration, which is obtained by differentiating the wheel speed signal.

[0244] Selecting tire adhesion utilization (the ratio of longitudinal force and lateral force to total adhesion) as the control target, the active suspension torque distribution problem can be transformed into the following nonlinear programming problem:

[0245]

[0246] stAx=b

[0247] lb≤x≤ub

[0248]

[0249] Among them, the quantity to be solved is x=[F afl ,F afr ,F arl ,F arr ] Tare the four active suspension forces, lb and ub are the minimum and maximum active suspension forces, respectively.

[0250] like Figure 2 As shown, the optimal value of the active suspension force is finally solved The above algorithm was preliminarily verified under the condition of 90km / h constant speed double lane change on a dry and straight road (road adhesion coefficient μ≈0.85). The vehicle body posture control (roll angle) results are shown as follows: Figure 4 As shown in the figure, compared with not applying control, the roll angle of the vehicle body is significantly reduced after applying the above control method, indicating that the control effect of the above method is better.

[0251] The above description is merely a preferred embodiment of the present invention and does not limit the present invention in any way. Any person skilled in the art who, without departing from the scope of the present invention, makes any equivalent substitution, modification, or other changes to the technical solution and technical content disclosed in the present invention shall be deemed to be within the scope of the present invention and still fall within the scope of protection of the present invention.

Claims

1. A vehicle body posture estimation and active control method, characterized in that: The following steps are involved: Step 1: Obtain vehicle status information and vehicle parameter information; Step 2: Obtain vehicle body posture measurement information: Based on vehicle status information and vehicle parameter information, the adaptive fusion filtering algorithm is used to fuse the signals obtained from the vehicle kinematic model calculation and sensor integration to output vehicle posture measurement information, including vehicle pitch angle, vehicle roll angle, and vehicle vertical displacement; Step 3: Get vehicle body posture estimation information: Based on the vehicle posture measurement information obtained in step 2 and the vertical seven-degree-of-freedom vehicle dynamics model that considers the influence of lateral acceleration and longitudinal acceleration, the vehicle posture is estimated using a Kalman filter algorithm, and the vehicle posture estimation information is output, including the vehicle pitch angle, the vehicle pitch angular velocity, the vehicle roll angle, the vehicle roll angular velocity, the vehicle vertical displacement, and the vehicle vertical velocity; Step 4: Calculate the feedforward amount required for vehicle posture control: Based on the vehicle posture estimation information obtained in step 3 and the fuzzy PID control algorithm, the feedforward quantity required for vehicle posture control is calculated, including the additional vertical force, the additional pitching moment, and the additional roll moment of the vehicle body; Step 5: Calculate the optimal control value of the vehicle body posture Based on a vertical seven-degree-of-freedom vehicle dynamics model that considers the effects of lateral and longitudinal accelerations, and using the feedforward values ​​obtained in step 4 as constraints, a model predictive control algorithm is used to output the optimal control variables for the vehicle body posture, including the optimal additional vertical force, the optimal additional pitching moment, and the optimal additional roll moment. Step 6: Based on the optimal control value of the vehicle body posture, select the tire adhesion utilization rate as the control target and solve the four optimal active suspension forces.

2. The vehicle body posture estimation and active control method according to claim 1, characterized in that: The vehicle status information in step 1 includes: The vehicle's longitudinal acceleration, lateral acceleration, vertical acceleration, pitch angular velocity, and roll angular velocity are measured by the inertial measurement unit (IMU), and the dynamic deflection of the four suspensions is measured by the suspension height sensor. Among them, the IMU is installed in the center of the chassis. The geometric relationship between the center of the chassis and the center of mass of the vehicle is used to convert the vehicle center of mass position status information. The four suspension height sensors are installed at the connection between the body and the upper end of the suspension. The vehicle parameter information includes: The vehicle's sprung and unsprung masses; Vehicle pitch moment of inertia, vehicle roll moment of inertia; The position of the vehicle's center of mass, including the height of the vehicle's center of mass, the distance from the front and rear axles to the vehicle's center of mass, and the distance from the left and right wheels to the vehicle's center of mass; Vehicle track width, suspension stiffness, suspension damping and tire stiffness.

3. The vehicle body posture estimation and active control method according to claim 2, characterized in that: Step 2 specifically includes the following steps: Step 2-1: Calculate the vehicle body posture based on the vertical seven-degree-of-freedom vehicle kinematic model, vehicle parameter information, and suspension dynamic deflection information; Step 2-2: Estimate the vehicle body posture through integration based on the vehicle body pitch angular velocity, vehicle body roll angular velocity, and vehicle body vertical acceleration information; Step 2-3: Determine the adaptive law of the adaptive fusion filtering algorithm based on the vertical acceleration information; The specific formula of the adaptive law of the fusion algorithm is: Among them, ω H is the cutoff frequency ω c The upper bound of ω L is the cutoff frequency ω c The lower bound of Δ z is the vertical acceleration a z Deviation from the acceleration due to gravity g; Δ dz is Δ z The derivative of a dz for a z The derivative of Step 2-4: The vehicle body postures obtained in steps 2-1 and 2-2 are fused through the adaptive fusion filter algorithm and the adaptive law to obtain the vehicle body posture measurement information; The specific formula of the adaptive fusion filtering algorithm is: Among them, ω c is the cutoff frequency of the high / low frequency filter; θ1, φ1, and z s1 are the body pitch angle, body roll angle and body vertical displacement calculated based on the integration of the body pitch angular velocity, body roll angular velocity and body vertical acceleration respectively; θ2, φ2, and z s2 They are the body pitch angle, body roll angle and body vertical displacement calculated based on the vehicle kinematic model.

4. The vehicle body posture estimation and active control method according to claim 3, characterized in that: The vertical seven-degree-of-freedom vehicle kinematic model in step 2-1 is: Among them, l w The wheelbase of the vehicle is the distance between the two front wheel centers, or the distance between the two rear wheel centers; l f is the distance from the front axle to the center of mass of the vehicle; l r is the distance from the rear axle to the center of mass of the vehicle; z sfl ,z sfr ,z srl ,z srr are the vertical displacements of the four endpoints of the suspension on the vehicle body; the endpoints are the connections between the upper end of the suspension and the vehicle body; fl, fr, rl, rr represent the left front wheel, right front wheel, left rear wheel and right rear wheel respectively; s represents the sprung position; z represents the z-direction displacement; θ represents the vehicle body pitch angle; Indicates the body roll angle; The calculation formula for the upper end point displacement of the suspension is: Among them, z ufl ,z ufr ,z url ,z urr The displacements of the unsprung masses corresponding to the left front wheel, right front wheel, left rear wheel, and right rear wheel, respectively; u means unsprung; h fl ,h fr ,h rl ,h rr The suspension dynamic deflections corresponding to the left front wheel, right front wheel, left rear wheel and right rear wheel are measured by the suspension height sensor.

5. The vehicle body posture estimation and active control method according to claim 4, characterized in that: Step 3 specifically includes the following steps: Step 3-1: Consider the influence of vehicle lateral acceleration and longitudinal acceleration to establish the system state space equation: The vertical seven-degree-of-freedom vehicle dynamics model is as follows: Where x = [z s θ φ z ufl z ufr z url z urr ] T is the state vector; M is the mass matrix; C is the damping matrix; K is the stiffness matrix; F is the force matrix; z s Indicates the vertical displacement of the vehicle body, that is, the vertical displacement of the upper end point of the suspension; The expressions of each coefficient matrix are: M=diag{m s ,I y ,I x ,m ufl ,m ufr ,m url ,m urr } Among them, m s ,m ufl ,m ufr ,m url ,m urr The sprung mass and the four unsprung masses are listed in order; I x ,I y are the pitch moment of inertia and the roll moment of inertia, respectively; x represents the x-axis; y represents the y-axis; c fl ,c fr ,c rl ,c rr is the damping of each suspension; k fl ,k fr ,k rl ,k rr is the stiffness of each suspension; k utfl ,k utfr ,k utrl ,k utrr is the tire stiffness of each wheel; ut means tire; F afl ,F afr ,F arl ,F arr It is the main force of each suspension, generated by the active suspension actuator; a means active; F z ,M θ ,M φ They are the additional vertical force, additional pitching moment and additional rolling moment generated by the main force of the suspension. F zfl ,F zfr ,F zrl ,F zrr is the vertical force of the wheel; z ufl ,z ufr ,z url ,z urr is the unsprung mass displacement corresponding to each wheel; z rfl ,z rfr ,z rrl ,z rrr is the road displacement of each wheel; a x Indicates the longitudinal acceleration of the vehicle body; a y Indicates the lateral acceleration of the vehicle body; h represents the height of the vehicle's center of mass from the ground; Step 3-2: Based on the state vector in step 3-1, list the state transfer matrix and observation matrix; solve the Kalman filter gain based on the state transfer matrix and observation matrix, and complete the vehicle posture estimation by combining the vehicle posture measurement information; x k+1 =A k x k +B k u k +f k +W k ,y k =H k x k +V k where, x k = [z s,k θ k φ k z ufl,k z ufr,k z url,k z urr,k z s,k+1 θ k+1 φ k+1 z ufl,k+1 z ufr,k+1 z url,k+1 z urr,k+1 T is the state vector;​ u k =[F z,k M θ,k M φ,k ] T is the control vector y k =[θ k+1 φ k+1 h fl,k h fr,k h rl,k h rr,k ] T is the observation vector; Among them, h fl ,h fr ,h rl ,h rr is the suspension dynamic deflection; W k is the process noise, V k To measure noise; k represents the sampling time; A k represents the state transfer matrix, B k represents the state control matrix; H k represents the observation matrix, f k Represents the external input; 6. The vehicle body posture estimation and active control method according to claim 1, characterized in that: In step 4, the vertical fuzzy PID controller, the pitch fuzzy PID controller and the roll fuzzy PID controller are used to control the vehicle body respectively; For the vertical fuzzy PID controller, select the target vertical displacement z d The error Δz between the actual vertical displacement z and the error change rate As input; its output is additional vertical force; d represents the target; For the pitch fuzzy PID controller, select the target pitch angle θ d The error Δθ between the actual pitch angle θ and the error change rate As input quantity; Its output is the additional pitching moment; For the roll fuzzy PID controller, select the target roll angle φ d The error Δφ between the actual roll angle φ and the error change rate As input; its output is additional rolling moment; The two variables in each of the above controllers are transformed into domain, fuzzy reasoning and defuzzification respectively, and finally the control parameters Δk are obtained respectively. P,z ,Δk I,z ,Δk D,z , Δk P,θ ,Δk I,θ ,Δk D,θ and Δk P,φ ,Δk I,φ ,Δk D,φ , and the control parameter k in the original PID controller P0,z ,k I0,z ,k D0,z 、k P0,θ ,k I0,θ ,k D0,θ and k P0,φ ,k I0,φ ,k D0,φ Together they form the new parameter k P,z ,k I,z ,k D,z 、k P,θ ,k I,θ ,k D,θ and k P,φ ,k I,φ ,k D,φ This enables the controller parameters to have adaptive adjustment capabilities, thereby improving the control effect of the system. The expressions of each PID parameter are as follows: The fuzzy set in the fuzzy control process is expressed as {negative large, negative medium, negative small, zero, positive small, positive medium, positive large}, abbreviated as: {NB, NM, NS, Z, PS, PM, PB}; The membership function and fuzzy rules are determined through simulation experiments and expressed through fuzzy surfaces. Finally, the additional vertical force, additional rolling moment and additional pitching moment actually required are calculated.

7. The vehicle body posture estimation and active control method according to claim 1, characterized in that: The vehicle body posture model predictive control algorithm in step 5 specifically includes the following steps: Step 5-1: Based on the vertical seven-degree-of-freedom vehicle dynamics model, list the state space equations of the system; in, is the state variable of the system; y(t)=[θ φ h fl h fr h rl h rr ] T is the output variable of the system, including pitch angle, roll angle and dynamic deflection of each suspension; u(t)=[F z M θ M φ ] T are the control variables of the system, including the additional vertical force of the vehicle body, the additional pitching moment of the vehicle body, and the additional rolling moment of the vehicle body; f(t) is the external input of the system f(t) = f k ; t represents the current moment; A, B, C, and D are the state space state matrix, input control matrix, output matrix, and output control matrix, respectively; Step 5-2: Discretize the equation in step 5-1 to obtain the discrete system state equation: A d =A·T samp +I,B d =B·T samp ,C d =C,D d =D Among them, A d Represents the discretized state space state matrix; B d represents the discretized input control matrix; C d Represents the discretized output matrix; D d represents the discretized output control matrix; T samp Indicates the sampling time; x(k) represents the discretized state variable; u(k) represents the discretized control variable; y(k) represents the discretized output variable; f(k) represents the external input after f(t) is discretized; Step 5-3: Select the vehicle body pitch angle, body roll angle, and suspension dynamic deflection as the control targets of the model predictive controller, construct the objective function, and establish the constraints: Where J is the objective function value; N P is the prediction step length; N C To control the step length; y(k+j|k) is the predicted value of the control output; y ref (k+j|k) is the control output reference value, including the reference values ​​of the pitch angle, roll angle and each suspension dynamic deflection, that is, the target vehicle body posture in Figure 2; u(k+j|k) is the control variable, including the additional vertical force of the vehicle body, the additional rolling moment of the vehicle body, and the additional pitching moment of the vehicle body; k+j|k means predicting the value at time k+j based on the information at time k, j = 1, 2, ..., N P ; u(k+j|k) represents the control input at time k+j; that is, the value of the control variable after discretization; j = 1, 2, ..., N C ; Q is the system output weight coefficient matrix; R is the system control weight coefficient matrix; The constraints of the control quantity are: Among them, F z,fuzzy ,M θ,fuzzy ,M φ,fuzzy are the additional vertical force, additional pitching moment and additional rolling moment of the vehicle body determined by the fuzzy PID controller in step 4 respectively; ΔF z ,ΔM θ ,ΔM φ The increments of the vehicle body additional vertical force, vehicle body additional pitching moment and vehicle body additional rolling moment; The constraints controlling the increment are: Where ΔF z,min ,ΔF z,max ,ΔM θ,min ,ΔM θ,max ,ΔM φ,min ,ΔM φ,max are the minimum and maximum values ​​of each control variable, respectively, which are determined by the vehicle driving mode; The output constraint can be expressed as: Among them, θ min and θ max is the minimum and maximum value of the vehicle body pitch angle; φ min and φ max is the minimum and maximum value of the vehicle body roll angle; h i,min and h i,max The maximum compression and maximum tension of the suspension dynamic deflection; Step 5-4: Solve and obtain the optimal control quantity for each step 8. The vehicle body posture estimation and active control method according to claim 1, characterized in that: Step 6 specifically includes the following steps: Step 6-1: Select tire adhesion utilization as the control target and transform the active suspension torque distribution problem into the following nonlinear programming problem: stAx=b lb≤x≤ub Among them, F x,i ,F y,i ,F z,i are the longitudinal force, lateral force and vertical force of each wheel; i represents the wheel; F x,i Calculated from the tire rotation equation; F z,i The equations to be solved are listed based on the vertical seven-degree-of-freedom vehicle model, which include the corresponding active suspension forces; F y,i Solve the equations listed by the magic formula, which contains F z,i ; x=[F afl ,F afr ,F arl ,F arr ] T is the force of the four active suspensions; lb and ub are the minimum and maximum values ​​of the active suspension force, respectively; Step 6-2: Solve the above nonlinear programming problem and finally obtain four optimal active suspension forces.

Images