A Distributed Drive-by-Wire Vehicle Motion Control Method Considering Parameter Robustness

Abstract

The present invention is applicable to the field of vehicle motion control technology, and provides a distributed wire-controlled vehicle motion control method considering parameter robustness, comprising the following steps: establishing a nominal model for system state prediction based on a two-degree-of-freedom dynamics model of a distributed wire-controlled vehicle; parameterizing model mismatch as bounded disturbance, constructing a system deviation model and a deviation suppression cost function, and solving a robust state feedback matrix using the LMI method; based on the tube‑based MPC theory, approximately solving the nominal model minimum robust invariant set, reconstructing state and control quantity constraints, solving the front and rear wheel angles and the expected additional yaw moment through rolling optimization; and distributing wheel drive torque to respond to the expected additional yaw moment. The control method provided by the present invention improves the parameter robustness of the coordinated motion control of distributed wire-controlled vehicles, and can effectively improve the vehicle handling stability under model mismatch conditions.

Description

Technical Field

[0001] The present invention belongs to the technical field of vehicle motion control, and in particular relates to a distributed wire-controlled vehicle motion control method considering parameter robustness. Background Art

[0002] Electric vehicles can be divided into two categories according to chassis structure and control mode: centralized and distributed wire control. Distributed wire control vehicles have independent controllable execution systems, which can achieve independent and precise control of vehicle steering and motor drive, expand the adaptability of the vehicle chassis, and create ideal conditions for improving vehicle motion control under extreme conditions. Compared with the traditional centralized structure, distributed wire control vehicles with overdrive system characteristics can improve stability through four-wheel steering control (4WS) and direct yaw moment control (DYC). Ensuring stability under extreme conditions is the primary goal of vehicle collaborative motion controller design.

[0003] The distributed chassis structure and control method not only compensate for the limitations of the centralized vehicle single execution system and expand the vehicle's safe driving area, but also increase the difficulty of building the control model: the model mismatch caused by its parameter uncertainty is more serious, which becomes a major challenge for motion control. In response to this challenge, the corresponding robust control strategy will set the error interference of the controlled system to the maximum value. Although it enhances the system's anti-interference ability, it is often accompanied by high conservatism. In addition, there is currently a lack of effective means to quantify and constrain the impact of robust feedback control on the system state.

[0004] Existing parameter robust controllers are often based on H∞ Law, μ The system error state feedback matrix is ​​designed by the synthesis method, and the difference between the simplified nominal model predicted state and the actual state of the system is multiplied to obtain the additional control quantity; the additional control quantity is added to the output of the inner membrane controller based on the nominal model to obtain the robust control output of the system. However, the above method has the following shortcomings: H∞ Law, μ The design of the system error state feedback matrix by the synthesis method will set the system error to a large fixed value by default. The feedback matrix remains unchanged during the entire control process and does not have the ability to adapt to the environment. After obtaining the additional control quantity, the above method does not combine it with the nominal model controller to correct the traditional internal model controller, but instead has the risk of causing system instability. Therefore, a distributed wire-controlled vehicle motion control method considering parameter robustness is proposed. Summary of the invention

[0005] The object of the present invention is to provide a distributed drive-by-wire vehicle motion control method taking parameter robustness into consideration, aiming to solve the problems raised in the above-mentioned background technology.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] A distributed control-by-wire vehicle motion control method considering parameter robustness comprises the following steps:

[0008] Step 1: Driving intention identification: Calculate and correct the expected yaw rate that reflects the driver's intention in real time;

[0009] Step 2: Nominal model construction: Considering the overdriving characteristics of the distributed wire-controlled vehicle, based on the two-degree-of-freedom dynamics model of the vehicle and performing linear simplification, a continuous state space equation that reflects the state evolution of the system is constructed; the continuous state space equation is discretized according to the controller operation cycle to obtain a nominal model for predicting the system state;

[0010] Step 3: Solve the robust state feedback matrix, including:

[0011] The model mismatch phenomenon caused by the uncertainty of key parameters in the process of establishing the nominal model is structured as a bounded disturbance and introduced into the nominal model, and a system deviation model is constructed to predict the error between the nominal model and the actual system state.

[0012] Construct a reactive robust control objective, that is, a deviation suppression cost function that suppresses the error between the nominal model predicted state and the actual system state and ensures a smooth change in the robust control quantity, and determine the minimum value of the deviation suppression cost function as the control objective of the robust state feedback matrix;

[0013] The Lyapunov function is introduced as the upper bound of the deviation suppression cost function;

[0014] Based on Schur's complement theorem, the constrained cost function optimization problem is transformed into a standard LMI problem to solve the robust state feedback matrix.

[0015] Step 4: Optimize the problem construction and solution, including:

[0016] Based on the tube-based MPC theory, The approximate method is used to solve the minimum robust invariant set of the nominal model; the robust state feedback matrix and the minimum robust invariant set are combined to tighten the constraint set of the nominal model state and control quantities, and the model predictive control optimization problem for tracking the desired state quantity is constructed. The optimal control quantity, i.e., the front and rear wheel steering angles and the desired additional yaw moment, is solved through rolling optimization;

[0017] Step 5: Driving torque distribution: distribute the driving torque of each wheel according to the expected additional yaw torque and longitudinal force.

[0018] Furthermore, in step 1, the expected yaw angular velocity is calculated as follows:

[0019] ;

[0020] In the formula, To control the yaw rate; Yaw rate correction boundary; is the weight; is a symbolic function;

[0021] Control yaw rate The calculation method is as follows:

[0022] ;

[0023] In the formula, is the vehicle speed; is the steering wheel angle input; is the steering gear ratio; , and They are the front and rear wheelbases; is the ratio of the front and rear wheel turning angles during steady-state steering, is the stability coefficient, which are expressed as follows:

[0024] ;

[0025] In the formula, and are the equivalent cornering stiffness of the front and rear wheels respectively; is the vehicle mass;

[0026] Yaw rate correction boundary The calculation method is as follows:

[0027] ;

[0028] In the formula, is the lateral acceleration; is the rate of change of lateral acceleration; is the rate of change of the steering wheel angle; , and They are the correction coefficients for lateral acceleration, steering and vehicle speed respectively;

[0029] Weight The calculation method is as follows:

[0030] .

[0031] Furthermore, in step 2, the nominal model state quantity is defined ,in are the sideslip angle and yaw rate of the center of mass respectively; the nominal model control quantity ,in are the rear wheel steering angle and the additional yaw moment respectively; the state, control and additional coefficient matrices of the system nominal model are defined respectively , and ;

[0032] by The discrete step lengths are discretized to obtain the discrete nominal model:

[0033] ;

[0034] In the formula, is the nominal model state quantity, where for The sideslip angle of the center of mass at time for The yaw rate at the moment; is the nominal model control quantity, where for The rear wheel turning angle at the moment, for The expected additional yaw moment at time; is the virtual control quantity, where for Front wheel turning angle at the moment; They are the coefficient matrices of nominal model state quantity, nominal model control quantity, and virtual control quantity respectively: , , ,in is discrete time, is the identity matrix.

[0035] Furthermore, the specific steps of step 3 are as follows:

[0036] The model mismatch caused by the uncertainty of key parameters in the nominal model establishment process is constructed as a bounded disturbance Introducing the nominal model, is a bounded interference set, and the system deviation model is constructed as:

[0037] ;

[0038] In the formula, is the coefficient matrix, ; , is the actual state quantity, is the nominal model state quantity; is the coefficient matrix; , is the robust state feedback matrix; ;

[0039] Robust state feedback matrix While suppressing system state deviation, ensure the control quantity Smooth changes require the following control objectives to be met:

[0040] ;

[0041] In the formula, for The state quantity of the system deviation model at time; for The system deviation model control quantity at the moment; and They are the state quantities of the system deviation model and system deviation model control quantity The weight of

[0042] Cost Function The upper bound of is the Lyapunov function ,set up The upper bound is , so the feedback control objective is transformed into the following constrained optimization problem:

[0043] ;

[0044] In the formula, and Systematic deviation model Moment and Lyapunov function of time; is the stability coefficient matrix; for The system deviation model control quantity at the moment;

[0045] The robust state feedback matrix The solution is transformed into a standard LMI problem, and the Schur complement theorem is introduced to deal with the related inequality constraints.

[0046] Furthermore, the specific steps of step 4 are as follows:

[0047] Based on tube-based MPC theory, define the minimum robust invariant set sequence as follows:

[0048] ;

[0049] In the formula, is the Minkowski sum operation between sets; ; r represents the matrix The rth power; t is a matrix The maximum value of the power;

[0050] use Approximate method to solve the minimum robust invariant set : For a given Obtain , So that the following is true:

[0051] ;

[0052] In the formula, is the set of positive real numbers; is a set of positive integers; Substitute the system nominal model and system deviation model to obtain the robust constraints of the nominal model state quantity and control quantity , , construct the optimization problem of coordinated motion control of distributed wire-controlled vehicles:

[0053] ;

[0054] In the formula, is the cost function; is the prediction step length; is the reference state quantity; To control the step length; is the control quantity to be optimized; is the state quantity weight; is the control amount weight; is the relaxation factor; is the initial state of the system deviation model; is the initial state of the system; is the initial state of the system nominal model; is a bounded state set; It is the Pontryagin difference operation between sets; is a bounded control set;

[0055] Solve the optimization problem to obtain the optimal control quantity of the nominal model and the initial state of system deviation , so the optimal control input of the actual system is:

[0056] .

[0057] Furthermore, in step 5, the driving torque distribution formula of each wheel is as follows:

[0058] ;

[0059] In the formula, Indicates the driving torque of the left front wheel; Indicates the driving torque of the right front wheel; Indicates the driving torque of the left rear wheel; Indicates the driving torque of the right rear wheel; Indicates the driving force of the vehicle; Indicates wheelbase; Indicates wheel radius; , , and They represent the load transfer coefficients of the left front, right front, left rear and right rear wheels respectively and are defined as follows:

[0060] ;

[0061] In the formula, Indicates the left front wheel load; Indicates the right front wheel load; Indicates the left rear wheel load; Indicates the right rear wheel load.

[0062] Compared with the prior art, the present invention has the following beneficial effects:

[0063] The method provided by the present invention constructs the model mismatch as a bounded disturbance compensation nominal model, which can adaptively perform robust control according to the vehicle motion state and has environmental adaptability; after obtaining the robust state feedback matrix, the state quantity and control quantity constraints of the nominal model are corrected to ensure the stability of the controlled vehicle system. The present invention improves the parameter robustness of the distributed wire-controlled vehicle cooperative motion control and can effectively improve the vehicle handling stability under model mismatch conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] Figure 1 Flow chart of the method of the present invention.

[0065] Figure 2 Graphs showing experimental results of the conventional control method and the method of the present invention in Example 1 regarding yaw rate and yaw rate error; wherein (a) is the yaw rate, and (b) is the yaw rate error.

[0066] Figure 3 Graphs showing experimental results of the conventional control method and the method of the present invention in Example 1 regarding the sideslip angle at the center of mass and the vehicle speed; wherein (a) is the sideslip angle at the center of mass, and (b) is the vehicle speed.

[0067] Figure 4 Graphs showing experimental results of the conventional control method and the method of the present invention in relation to the rear wheel steering angle and the additional yaw moment in Example 1; wherein (a) is the rear wheel steering angle, and (b) is the additional yaw moment. DETAILED DESCRIPTION

[0068] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0069] like Figure 1 As shown, a distributed wire-controlled vehicle motion control method considering parameter robustness is provided in one embodiment of the present invention, comprising the following steps:

[0070] Step 1: Driving intention identification: Calculate and correct the expected yaw rate that reflects the driver's intention in real time. The driver's intention can be characterized by the steady-state yaw rate response under a certain vehicle speed and steering wheel angle input; it is corrected as the expected yaw rate according to the change in the vehicle's lateral acceleration to improve the stability of the vehicle.

[0071] The expected yaw rate is calculated as follows:

[0072] ;

[0073] In the formula, To control the yaw rate; Yaw rate correction boundary; is the weight; is a symbolic function;

[0074] Control yaw rate The calculation method is as follows:

[0075] ;

[0076] In the formula, is the vehicle speed; is the steering wheel angle input; is the steering gear ratio; , and They are the front and rear wheelbases; is the ratio of the front and rear wheel turning angles during steady-state steering, is the stability coefficient, which are expressed as follows:

[0077] ;

[0078] In the formula, and are the equivalent cornering stiffness of the front and rear wheels respectively; is the vehicle mass;

[0079] Yaw rate correction boundary The calculation method is as follows:

[0080] ;

[0081] In the formula, is the lateral acceleration; is the rate of change of lateral acceleration; is the rate of change of the steering wheel angle; , and They are the correction coefficients for lateral acceleration, steering and vehicle speed respectively;

[0082] Weight The calculation method is as follows:

[0083] .

[0084] Step 2: Nominal model construction: Considering the over-driving characteristics of distributed wire-controlled vehicles, based on the vehicle's two-degree-of-freedom dynamics model and combined with engineering practice, linear simplification is performed to construct a continuous state-space equation that reflects the state evolution of the system; the continuous state-space equation is discretized according to the controller operation cycle to obtain a nominal model for predicting the system state.

[0085] The kinetic model is expressed as follows:

[0086] ;

[0087] In the formula, is the derivative of the sideslip angle at the center of mass; is the yaw rate derivative; and are the equivalent cornering stiffness of the front and rear wheels respectively; is the vehicle mass; is the vehicle speed; and They are the front and rear wheelbases; is the moment of inertia of the rotating vehicle around the z-axis; is the sideslip angle of the center of mass; is the yaw angular velocity; is the rear wheel turning angle; is the additional yaw moment; is the front wheel turning angle;

[0088] Nominal model state , nominal model control quantity , define the state, control and additional coefficient matrices of the system nominal model as , and ;

[0089] by The discrete step lengths are discretized to obtain the discrete nominal model:

[0090] ;

[0091] In the formula, is the nominal model state quantity, where for The sideslip angle of the center of mass at time for The yaw rate at the moment; is the nominal model control quantity, where for The rear wheel turning angle at the moment, for The expected additional yaw moment at time; is the virtual control quantity, where for Front wheel turning angle at the moment; They are the coefficient matrices of nominal model state quantity, nominal model control quantity, and virtual control quantity respectively: , , ,in is discrete time, is the identity matrix.

[0092] Step 3: Solve the robust state feedback matrix, including:

[0093] The key parameters in the process of establishing the nominal model ( ) The model mismatch phenomenon caused by uncertainty is structured into a bounded disturbance by introducing a nominal model, and a system deviation model is constructed to predict the error between the nominal model and the actual system state;

[0094] Construct a reactive robust control objective, that is, a deviation suppression cost function that suppresses the error between the nominal model predicted state and the actual system state and ensures a smooth change in the robust control quantity, and determine the minimum value of the deviation suppression cost function as the control objective of the robust state feedback matrix;

[0095] The Lyapunov function is introduced as the upper bound of the deviation suppression cost function, that is, the robust state feedback matrix should be solved to minimize the Lyapunov function value and satisfy the constraint that the Lyapunov function is always used as the upper bound of the deviation suppression cost function;

[0096] Based on Schur's complement theorem, the constrained cost function optimization problem is transformed into a standard LMI problem to solve the robust state feedback matrix.

[0097] The specific steps of step 3 are as follows:

[0098] The model mismatch caused by the uncertainty of key parameters in the nominal model establishment process is constructed as a bounded disturbance Introducing the nominal model, is a bounded interference set, and the system deviation model is constructed as:

[0099] ;

[0100] In the formula, is the coefficient matrix, ; , is the actual state quantity, is the nominal model state quantity; is the coefficient matrix; , is the robust state feedback matrix; ;

[0101] Robust state feedback matrix While suppressing system state deviation, ensure the control quantity Smooth changes require the following control objectives to be met:

[0102] ;

[0103] In the formula, for The state quantity of the system deviation model at time; for The system deviation model control quantity at the moment; and They are the state quantities of the system deviation model and system deviation model control quantity The weight of

[0104] Cost Function The upper bound of is the Lyapunov function ,set up The upper bound is , so the feedback control objective is transformed into the following constrained optimization problem:

[0105] ;

[0106] In the formula, and Systematic deviation model Moment and Lyapunov function of time; is the stability coefficient matrix; for The system deviation model control quantity at the moment;

[0107] The robust state feedback matrix The solution is transformed into a standard LMI problem, and the Schur complement theorem is introduced to deal with the related inequality constraints.

[0108] Step 4: Optimize the problem construction and solution, including:

[0109] Based on the tube-based MPC theory, The approximate method is used to solve the minimum robust invariant set of the nominal model. The robust state feedback matrix and the minimum robust invariant set are combined to tighten the constraint set of the state and control quantities of the nominal model, and a model predictive control optimization problem for tracking the desired state quantities is constructed. The optimal control quantities, namely the front and rear wheel steering angles and the desired additional yaw moment, are solved through rolling optimization.

[0110] The specific steps of step 4 are as follows:

[0111] Based on tube-based MPC theory, define the minimum robust invariant set sequence as follows:

[0112] ;

[0113] In the formula, is the Minkowski sum operation between sets; ; r represents the matrix The rth power; t is a matrix The maximum value of the power;

[0114] use Approximate method to solve the minimum robust invariant set : For a given Obtain , So that the following is true:

[0115] ;

[0116] In the formula, is the set of positive real numbers; is a set of positive integers; Substitute the system nominal model and system deviation model to obtain the robust constraints of the nominal model state quantity and control quantity , , construct the optimization problem of coordinated motion control of distributed wire-controlled vehicles:

[0117] ;

[0118] In the formula, is the cost function; is the prediction step length; is the reference state quantity; To control the step length; is the control quantity to be optimized; is the state quantity weight; is the control amount weight; is the relaxation factor; is the initial state of the system deviation model; is the initial state of the system; is the initial state of the system nominal model; is a bounded state set; It is the Pontryagin difference operation between sets; is a bounded control set;

[0119] Solve the optimization problem to obtain the optimal control quantity of the nominal model and the initial state of system deviation , so the optimal control input of the actual system is:

[0120] .

[0121] Step 5: Drive torque distribution: In order to respond to the expected additional yaw torque and meet the total longitudinal force requirements, the drive torque of each wheel is distributed. When the vehicle starts and accelerates on a low-adhesion road, the wheels also need to be anti-slip controlled to ensure the output of the drive torque.

[0122] In step 5, the driving torque distribution formula of each wheel is as follows:

[0123] ;

[0124] In the formula, Indicates the driving torque of the left front wheel; Indicates the driving torque of the right front wheel; Indicates the driving torque of the left rear wheel; Indicates the driving torque of the right rear wheel; Indicates the driving force of the vehicle; Indicates wheelbase; Indicates wheel radius; , , and They represent the load transfer coefficients of the left front, right front, left rear and right rear wheels respectively and are defined as follows:

[0125] ;

[0126] In the formula, Indicates the left front wheel load; Indicates the right front wheel load; Indicates the left rear wheel load; Indicates the right rear wheel load.

[0127] The specific implementation of the present invention is described in detail below in conjunction with specific embodiments.

[0128] Example 1, low attachment meandering test model mismatch condition;

[0129] This embodiment selects a low-adhesion snake test with a vehicle speed of 50 km / h and an adhesion coefficient of 0.2 to compare and verify the robustness of the algorithm parameters. It is set to 1.2 times the actual value to simulate severe model misfit caused by inaccurate parameter estimation. Figure 2-4 The comparison of vehicle state and control variable changes between the traditional collaborative control method of the hardware-in-the-loop test and the method of the present invention (taking parameter robustness into consideration) under this working condition is demonstrated.

[0130] In this experiment, from Figure 2 , 3 It can be seen that the controlled vehicle started to turn to participate in the test at 3.5s. At this time, both methods can track the expected value more accurately. The traditional control method is difficult to effectively suppress the error caused by model mismatch. After 5.5s, the vehicle state oscillates around the expected value. The state prediction error caused by the mismatch is not effectively suppressed, and the impact on stability control continues to increase until it becomes unstable at 16s. The method of the present invention can track the expected value more stably and smoothly. Figure 2 It can be seen from (b) that the maximum tracking error of the present invention is 3.10( deg / s ), the maximum tracking error of the traditional control method in the non-instability stage is only 9.87 ( deg / s ) 31.4%; Figure 3 In (a), it can be seen that the peak value of the center of mass side slip angle of the present invention in the non-instability section is 0.25 ( deg / s ) compared to the traditional control method of 1.31 ( deg / s ) is also smaller, only 19% of it. The control amount changes of the two methods in the experiment are as follows Figure 4 As shown, since the model response error of the traditional control method can never be effectively suppressed, the control amount changes drastically, eventually leading to vehicle instability; while the control amount of the method of the present invention changes more slowly.

[0131] The above are only preferred embodiments of the present invention. It should be pointed out that, for those skilled in the art, several modifications and improvements can be made without departing from the concept of the present invention. These should also be regarded as the protection scope of the present invention. These will not affect the effect of the implementation of the present invention and the practicality of the patent.

Claims

1. A distributed control-by-wire vehicle motion control method considering parameter robustness, characterized in that: The following steps are involved: Step 1: Driving intention identification: Calculate and correct the expected yaw rate that reflects the driver's intention in real time; Step 2: Nominal model construction: Considering the overdriving characteristics of the distributed wire-controlled vehicle, based on the two-degree-of-freedom dynamics model of the vehicle and performing linear simplification, a continuous state space equation that reflects the state evolution of the system is constructed; the continuous state space equation is discretized according to the controller operation cycle to obtain a nominal model for predicting the system state; Step 3: Solve the robust state feedback matrix, including: The model mismatch phenomenon caused by the uncertainty of key parameters in the process of establishing the nominal model is structured as a bounded disturbance and introduced into the nominal model, and a system deviation model is constructed to predict the error between the nominal model and the actual system state. Construct a reactive robust control objective, that is, a deviation suppression cost function that suppresses the error between the nominal model predicted state and the actual system state and ensures a smooth change in the robust control quantity, and determine the minimum value of the deviation suppression cost function as the control objective of the robust state feedback matrix; The Lyapunov function is introduced as the upper bound of the deviation suppression cost function; Based on Schur complement theorem, the constrained cost function optimization problem is transformed into a standard LMI problem to solve the robust state feedback matrix. Step 4: Optimize the problem construction and solution, including: Based on the tube-based MPC theory, the ε approximation method is used to solve the minimum robust invariant set of the nominal model. The robust state feedback matrix and the minimum robust invariant set are combined to tighten the constraint set of the state and control quantities of the nominal model, and the model predictive control optimization problem for tracking the expected state quantity is constructed. The optimal control quantity, i.e., the front and rear wheel steering angles and the expected additional yaw moment, is solved through rolling optimization. Step 5: Driving torque distribution: Distribute the driving torque of each wheel according to the expected additional yaw torque and longitudinal force; The specific steps of step 3 are as follows: The model mismatch caused by the uncertainty of key parameters in the nominal model establishment process is constructed as a bounded disturbance ω k ∈W introduces the nominal model, W is the bounded interference set, and the system deviation model is constructed as: In the formula, A d is the coefficient matrix, A d =T s A c +I; x k is the actual state quantity, is the nominal model state quantity; B 1,d is the coefficient matrix; K is the robust state feedback matrix; The robust state feedback matrix K ensures the control quantity while suppressing the system state deviation Smooth changes require the following control objectives to be met: In the formula, is the state quantity of the system deviation model at time k+i; is the control quantity of the system deviation model at time k+i; Q d and R d They are the state quantities of the system deviation model and system deviation model control quantity The weight of Cost Function The upper bound of is the Lyapunov function set up The upper bound is λ, so the feedback control objective is transformed into the following constrained optimization problem: In the formula, and are the Lyapunov functions of the system deviation model at time k and time k+1 respectively; P d is the stability coefficient matrix; is the control quantity of the system deviation model at time k; The solution of the robust state feedback matrix K is transformed into a standard LMI problem, and the Schur complement theorem is introduced to deal with related inequality constraints; In step 5, the driving torque distribution formula of each wheel is as follows: Where, T xfl Indicates the driving torque of the left front wheel; T xfr Indicates the driving torque of the right front wheel; T xrl Indicates the driving torque of the left rear wheel; T xrr Indicates the driving torque of the right rear wheel; F t Indicates the driving force of the vehicle; w d Indicates wheelbase; r w Indicates wheel radius; c fl 、c fr 、c rl With c rr They represent the load transfer coefficients of the left front, right front, left rear and right rear wheels respectively and are defined as follows: In the formula, F zfl Indicates the load on the left front wheel; F zfr Indicates the right front wheel load; F zrl Indicates the left rear wheel load; F zrr Indicates the right rear wheel load.

2. The distributed drive-by-wire vehicle motion control method considering parameter robustness according to claim 1, characterized in that: In step 1, the expected yaw rate is calculated as follows: c des =sign(γ h )(oh γ c h |+(1-ω γ )c c |); In the formula, γ h is the control yaw rate; c Yaw rate correction boundary; ω γ is the weight; sign is the sign function; Control yaw rate γ h The calculation method is as follows: In the formula, v x is the vehicle speed; δ sw is the steering wheel angle input; i sw is the steering gear ratio; L = l f +l r , l f and l r Respectively, front and rear wheelbase; K δ is the ratio of the front and rear wheel turning angles during steady-state steering, K s is the stability coefficient, which are expressed as follows: In the formula, k f and k r are the equivalent cornering stiffness of the front and rear wheels respectively; m is the vehicle mass; Yaw rate correction boundary γ c The calculation method is as follows: In the formula, a y is the lateral acceleration; is the rate of change of lateral acceleration; is the rate of change of the steering wheel angle; k1, k2 and k3 are the correction coefficients of lateral acceleration, steering and vehicle speed respectively; Weight ω γ The calculation method is as follows:

3. The distributed drive-by-wire vehicle motion control method considering parameter robustness according to claim 1, characterized in that: In step 2, the nominal model state quantity is defined Where β and γ are the sideslip angle and yaw rate of the center of mass respectively; the nominal model control quantity where δ r ,M zd are the rear wheel steering angle and the additional yaw moment respectively; the state, control and additional coefficient matrices of the nominal model of the system are defined as A c , B 1,c With B 2,c ; Discretization is performed with a discrete step length of t = 0.02s to obtain the discrete nominal model: In the formula, is the nominal model state quantity, where β k is the sideslip angle of the center of mass at time k, γ k is the yaw angular velocity at time k; is the nominal model control quantity, where is the rear wheel turning angle at time k, is the expected additional yaw moment at time k; is the virtual control quantity, where is the front wheel turning angle at time k; A d ,B 1,d ,B 2,d A is the coefficient matrix of nominal model state quantity, nominal model control quantity and virtual control quantity respectively: d =T s A c +I, B 1,d =T s B 1,c , B 2,d =T s B 2,c , where T s is the discrete time, and I is the identity matrix.

4. The distributed drive-by-wire vehicle motion control method considering parameter robustness according to claim 1, characterized in that: The specific steps of step 4 are as follows: Based on tube-based MPC theory, we define the minimum robust invariant set sequence Z t as follows: In the formula, is the Minkowski sum operation between sets; r represents the matrix The rth power; t is a matrix The maximum value of the power; Using ε approximation method to solve the minimum robust invariant set Z * :For a given ε, find t∈N + , α∈[0,1) makes the following equation true: In the formula, R + is a set of positive real numbers; N + is a set of positive integers; Z * Substitute the system nominal model and system deviation model to obtain the robust constraints X of the nominal model state quantity and control quantity r , U r , construct the optimization problem of coordinated motion control of distributed wire-controlled vehicles: Where J is the cost function; N p is the prediction step length; is the reference state quantity; N c To control the step length; is the control quantity to be optimized; Q is the state quantity weight; R is the control quantity weight; ρ is the relaxation factor; is the initial state of the system deviation model; x0 is the initial state of the system; is the initial state of the system nominal model; X is the bounded state set; is the Pontryagin difference operation between sets; U is a bounded control set; Solve the optimization problem to obtain the optimal control quantity of the nominal model and the initial state of system deviation Therefore, the optimal control input of the actual system is:

Images