Distributed wheel hub driven electric vehicle lateral and vertical coordinated control method and related equipment

Abstract

The invention discloses a lateral and vertical coordinated control method for a distributed wheel hub driven electric vehicle and related equipment. The method establishes a 14-degree-of-freedom model of a whole vehicle taking into account the unbalanced magnetic force of a wheel hub motor, so as to accurately represent the state response of the vehicle under different inputs; a three-dimensional piecewise affine tire model performs piecewise linearization on the lateral force of the tire under different sideslip angles and vertical loads, and can describe the nonlinear characteristics of the tire under extreme working conditions; an AFS controller establishes an active front wheel steering controller based on hybrid model predictive control, and obtains the additional turning angle of the vehicle with reference to the sideslip angle and yaw angular velocity of the center of mass obtained from a bicycle model, thereby calculating the additional turning angle of the vehicle to ensure vehicle stability; an ASS controller takes roll stability, ride comfort, i.e., tire dynamic load, vehicle body acceleration, and wheel hub motor stator and rotor eccentricity as suspension design control targets, and designs an active suspension controller with multi-constraint input and multi-constraint output based on dual model predictive control, so as to accurately describe tire characteristics.

Description

Technical Field

[0001] The present invention belongs to the technical field of vehicle control, and in particular relates to a distributed wheel hub driven electric vehicle lateral and vertical coordinated control method and related equipment. Background Art

[0002] Distributed Drive Electric Vehicle (DDEV) has the characteristic of independent drive control of four wheels, which enables it to have better dynamic control and energy transfer path, and facilitates the implementation of more complex and reliable active safety control technology.

[0003] As a new type of drive for electric vehicles, wheel hub motor drive simplifies the vehicle structure, improves transmission efficiency and driving stability, but also increases the unsprung mass of the vehicle, affecting the lateral-vertical dynamic performance. Existing studies have shown that the vertical component of the unbalanced magnetic force generated by the relative eccentricity of the wheel hub motor stator and rotor is coupled with the suspension system, affecting the vehicle's smoothness and comfort. At the same time, the eccentricity of the wheel hub motor stator and rotor is positively correlated with the unbalanced magnetic force, which leads to increased bearing wear and reduced motor life. In addition, the unbalanced magnetic force of the wheel hub motor also directly acts on the tire, affecting the tire's road holding ability, and thus affecting the vehicle's handling stability. When the vehicle is in a steering condition, the body roll causes the wheel hub motor stator and rotor eccentricity to increase, increasing the unbalanced magnetic force, thereby deteriorating the lateral dynamic performance. However, existing lateral dynamics research has not considered this issue. When the vehicle is in a steering condition, the body will roll and cause axle load transfer, and there is an obvious coupling relationship between the vehicle's lateral-vertical dynamics. However, when designing lateral-vertical integrated controllers, most researchers often use a linear tire model for the lateral dynamics model. The designed controller is difficult to cope with the unstable state when the vehicle tire enters the nonlinear region or even the saturated region. It is even possible that under extreme conditions, the vehicle enters a more dangerous state due to the decision error of the controller. In addition, the lateral force of the tire is not only related to the sideslip angle, but also to the vertical load of the tire and the road adhesion coefficient. When the vehicle is turning, the body rolls, which will cause a large difference in the vertical load of the inner and outer tires.

[0004] It can be seen that the existing hub motor drive control method only uses a linear model of the tire lateral force under a single vertical load, which is difficult to accurately describe the nonlinear characteristics of the tire, resulting in poor control performance under extreme working conditions. Summary of the invention

[0005] The present invention provides a distributed wheel hub driven electric vehicle lateral and vertical coordinated control method and related equipment to solve the technical problems of low control accuracy and poor control effect of existing distributed drive electric vehicle stability control under extreme working conditions.

[0006] In order to achieve the above object, the present invention adopts the following technical solution:

[0007] A distributed wheel hub driven electric vehicle lateral and vertical coordinated control method, comprising:

[0008] Construct a 14-DOF vehicle model based on the unbalanced magnetic force model of the hub motor;

[0009] Based on the 14-DOF vehicle model and the optimized magic tire correction formula, a three-dimensional piecewise affine tire model of lateral force-sideslip angle-vertical load is constructed;

[0010] Based on the three-dimensional piecewise affine tire model, an active front wheel steering controller based on hybrid model predictive control is built; at the same time, with roll stability, ride comfort and wheel hub motor stator and rotor eccentricity as the design control objectives, an active suspension controller based on multi-constraint input and multi-constraint output is built;

[0011] Based on the front wheel angle, The phase plane and lateral load transfer rate are used to construct a coordinated control strategy for the active front wheel steering controller and the active suspension controller to achieve coordinated control of the lateral and vertical directions of the distributed wheel hub drive electric vehicle.

[0012] Furthermore, the unbalanced magnetic force model of the hub motor is obtained by separating the stator and rotor of the hub motor and considering the tire and the rotor as an integral whole; the 14-DOF vehicle model includes longitudinal, lateral, yaw, pitch, roll and vertical motion of the sprung mass, stator mass and rotor and tire mass; wherein,

[0013] The differential equations of motion in the longitudinal and lateral directions are given by:

[0014] ma x =F x1 cos(δ)+F x2 cos(δ)+F x3 +F x4

[0015] ma y =F y1 cos(δ)+F y2 cos(δ)+F y3 +F y4

[0016] Among them, m is the total mass of the vehicle, δ is the front wheel turning angle, a x is the longitudinal acceleration, and the expression is a y is the lateral acceleration, and the expression is v x is the longitudinal speed, v yis the lateral speed, F xi is the wheel longitudinal force, F yi is the lateral force of the wheel, i=1,2,3,4, representing the left front, right front, left rear, and right rear tires respectively;

[0017] The differential equations of motion for yaw, pitch, and roll are given by:

[0018]

[0019] Where ω is the yaw rate, a and b are the distances from the center of mass to the front and rear axes, and B f , B r are the front and rear axle track widths, θ is the vehicle body pitch angle, h g is the height from the center of mass to the ground, h s is the height from the roll center to the ground, g is the acceleration due to gravity, F si is the suspension force, m b is the sprung mass, φ is the body roll angle, I x , I y , I z are the moments of inertia around the vehicle coordinate axes X, Y, and Z, respectively;

[0020] The vertical motion of the sprung mass, stator mass, and tire and rotor mass can be expressed as:

[0021]

[0022] in, m si is the stator mass, m ri is the mass of the tire and rotor, K si is the suspension stiffness coefficient, C si is the suspension damping coefficient, K mi is the stiffness coefficient between the stator and rotor of the motor, K ti is the tire stiffness coefficient, Z i1 is the vertical displacement of the tire + motor rotor, Z i2 is the vertical displacement of the motor stator, Z i3 is the vertical displacement of the sprung mass, Z 0i is the roughness of the four-wheel road surface, F i is the main force of the suspension, F Ei It is the vertical unbalanced excitation of the hub motor.

[0023] Furthermore, the optimization steps of the magic tire correction formula are as follows:

[0024] Introduce the road adhesion coefficient into the magic tire correction formula, the specific formula is as follows:

[0025] Among them, F yis the wheel lateral force, μ is the road adhesion coefficient, α is the tire slip angle, and B y Stiffness factor, C y is the shape factor, D y is the peak factor, E y is the curvature factor;

[0026] Based on wheel dynamics analysis, the tire slip angle is expressed as:

[0027]

[0028] The vertical tire load is expressed as:

[0029]

[0030] Among them, F z is the vertical load on the tire.

[0031] Furthermore, the construction process of the three-dimensional piecewise affine tire model of lateral force-sideslip angle-vertical load is as follows:

[0032] The lateral force-sideslip angle-vertical load data obtained by the optimized magic tire correction formula is used as the original data set Ω, and the improved K-plane clustering is used to cluster the original data Ω. The steps are as follows:

[0033] Step 1: Divide the original data set into uniform n×m sub-data sets Ω along the x and y axes. j , (j=1…n×m), the least squares estimation method is used to fit each subset of data sets Ω j The plane parameters are expressed as follows:

[0034]

[0035] in, Plane parameters for each sub-dataset;

[0036] Assuming there are p initial clustering planes, randomly select a plane as the first initial plane, recorded as:

[0037] M 1 ={a 1 ,b 1 ,c 1 ,d 1};

[0038] Step 2: The determined initial plane {M 1 ,M 2 ,…,M q The average value of the parameters is recorded as:

[0039]

[0040] The remaining plane parameters and The difference can be expressed as:

[0041]

[0042] Where, q is the number of initial planes that have been determined, and q≤p;

[0043] Select the maximum difference, that is, maxD i (Ω j ,M i ) as the next initial plane;

[0044] Step 3: Repeat step 2 until p initial planes {M 1 ,M 2 ,…,M p}, to complete the initialization of K-plane clustering;

[0045] Step 4: Use p planes to cluster the tire data points. The corresponding data subset is {J 1 ,J 2 ,…,J p}, the data classification rules are as follows:

[0046]

[0047] Among them, Dis i-j Represents a point (x i ,y i ,z i ) to the jth plane, [a j ,b j ,c j ,d j ] is the parameter of the jth plane, and the data classification rule is used to classify the data point (x i ,y i ,z i ) is classified to the plane J that is closest to it j ;

[0048] Step 5: Identify and reclassify the boundary outliers. i ,y i ,z i ) to a certain plane is less than a certain threshold and the number of points is greater than or equal to 2, it is judged as a boundary outlier point, and the following conditions are met:

[0049] A=[Dis i-1 ,Dis i-2 ,…,Dis i-p ]

[0050] ifnum(A(i)≤ε boundary )≥2

[0051] Then point(i)is boundary outlier

[0052] Where A is the point (x i ,y i ,z i ) to each plane, ε boundary is the boundary outlier threshold;

[0053] After identifying the boundary outliers, the outliers are reclassified using the minimum distance from the outliers to the centers of p data subsets to obtain a three-dimensional segmented affine tire sub-model data set.

[0054] Step 6: According to the data subset corresponding to each plane, use LSM to re-identify its plane parameters and repeat steps 4-5 to obtain a three-dimensional piecewise affine model through multiple cycles of K-plane clustering.

[0055] Step 7: Identify and reclassify the cutting anomaly points, and convert the data set with cutting anomaly points into Project the data onto the x, y or z axis and transform it into one-dimensional data. Use the k-means clustering method to perform cluster analysis, and identify outlier points by the proportion of the number of sub-data sets after clustering:

[0056]

[0057] in, is the sub-dataset obtained by k-means clustering, e=1,2,...,p, j is the number of categories of k-means clustering, which should be determined according to the data set The distribution of size(*) indicates the number of data in the data set, ξ cut is the threshold for cutting outliers;

[0058] After the outlier points are identified, the distance between the outlier points and the centers of p data subsets is minimized to perform secondary classification of the outlier points and obtain the final tire sub-model data. And use LSM to re-identify its plane parameters;

[0059] Step 8, using a support vector machine to estimate the coefficient matrix of the interface of the three-dimensional piecewise affine tire submodel;

[0060] Finally, the three-dimensional piecewise affine tire model is expressed as:

[0061] F y =θ i-1 α+θi-2 F z +θ i-3 ,i=(1,2,…,p)

[0062] Among them, θ i-1 ,θ i-2 and θ i-3 are the identified three-dimensional piecewise affine tire model parameters respectively.

[0063] Furthermore, based on the three-dimensional piecewise affine tire model, the specific steps of building an active front wheel steering controller based on hybrid model predictive control include:

[0064] Based on the 14-DOF vehicle model, the lateral dynamics model is simplified to obtain the error state space equation.

[0065] The error state space equation is transformed by introducing auxiliary discrete variables and combining IF-THEN-ELSE rules;

[0066] The quadratic programming problem is solved based on the three-dimensional piecewise affine tire model and the transformed error state space equation to obtain the additional steering angle of control to complete the construction of the active front wheel steering controller based on hybrid model predictive control.

[0067] Furthermore, the specific steps of building an active suspension controller based on multi-constraint input and multi-constraint output include:

[0068] The 14-DOF vehicle model is simplified to obtain a half-vehicle suspension model;

[0069] Based on the half-vehicle suspension model, the roll stability, ride comfort and the eccentricity of the hub motor stator and rotor are taken as the design control objectives. The quadratic programming problem is solved to obtain the objective functions and coordination parameters corresponding to the roll stability, ride comfort and the eccentricity of the hub motor stator and rotor, so as to complete the construction of an active suspension controller based on multi-constraint input and multi-constraint output.

[0070] Furthermore, based on the front wheel angle, phase plane and lateral load transfer rate, a coordinated control strategy of the active front wheel steering controller and the active suspension controller is constructed, including:

[0071] (1) When the vehicle is traveling in a straight line, the active front wheel steering controller does not work, and the dynamic suspension controller works to optimize ride comfort, that is, p 1 =0, p 2 =1;

[0072] (2) When the vehicle is in a steering condition, both the active front wheel steering controller and the dynamic suspension controller are working. Here, p 1 =1, p 2 Combining LTR indicators and The design is based on the phase plane stability boundary, and its expression is:

[0073]

[0074] Among them, LTR * =0.5 and They are the stability threshold and rollover threshold respectively.

[0075] A distributed wheel hub driven electric vehicle lateral and vertical coordinated control system, used to implement the steps of the above-mentioned distributed wheel hub driven electric vehicle lateral and vertical coordinated control method, comprising:

[0076] The vehicle model building module is used to build a 14-DOF vehicle model based on the unbalanced magnetic force model of the hub motor;

[0077] The three-dimensional piecewise affine tire model building module is used to build a three-dimensional piecewise affine tire model of lateral force-sideslip angle-vertical load based on a 14-DOF vehicle model combined with an optimized magic tire correction formula;

[0078] The controller building module is used to build an active front wheel steering controller based on a hybrid model predictive control based on a three-dimensional piecewise affine tire model; at the same time, with roll stability, ride comfort and wheel hub motor stator and rotor eccentricity as design control objectives, an active suspension controller based on multi-constraint input and multi-constraint output is built;

[0079] Cooperative control module for The phase plane and lateral load transfer rate are used to construct a coordinated control strategy for the active front wheel steering controller and the active suspension controller to achieve coordinated control of the lateral and vertical directions of the distributed wheel hub drive electric vehicle.

[0080] A device comprising:

[0081] Memory for storing computer programs;

[0082] The processor is used to implement the steps of the above-mentioned distributed wheel hub driven electric vehicle lateral and vertical coordinated control method when executing the computer program.

[0083] A computer-readable storage medium stores a computer program, which, when executed by a processor, is used to implement the steps of the above-mentioned distributed wheel hub driven electric vehicle lateral and vertical coordinated control method.

[0084] Compared with the prior art, the present invention has the following beneficial effects:

[0085] The present invention provides a method for coordinated lateral and vertical control of a distributed hub-driven electric vehicle. The method establishes a 14-degree-of-freedom model of the whole vehicle taking into account the unbalanced magnetic force of the hub motor, which can accurately represent the state response of the vehicle under different inputs. The three-dimensional piecewise affine tire model piecewise linearizes the lateral force of the tire under different sideslip angles and vertical loads, which can describe the nonlinear characteristics of the tire under extreme working conditions. The AFS controller establishes an active front wheel steering controller based on hybrid model predictive control, and obtains the additional turning angle of the vehicle with reference to the sideslip angle and yaw angular velocity of the center of mass from the bicycle model to ensure vehicle stability. The ASS controller uses roll stability, ride comfort, i.e., tire dynamic load, vehicle body acceleration, and hub motor stator and rotor eccentricity as suspension design control targets, and designs an active suspension controller with multi-constraint input and multi-constraint output based on dual model predictive control. The coordinated control strategy is based on the front wheel angle, The two controllers are designed based on the phase plane and the lateral load transfer rate, and they complement each other to improve the dynamic performance of the vehicle. At the same time, a coordination strategy between the two control systems is designed to achieve lateral and vertical coordinated control of the vehicle and improve the stability and comfort of the vehicle control.

[0086] Preferably, in the present invention, the fourteen-degree-of-freedom vehicle model covers the longitudinal, lateral, yaw, pitch, roll and vertical motion of each mass of the vehicle, and can comprehensively analyze the dynamic response of the vehicle under complex working conditions, providing strong support for the design of the controller.

[0087] Preferably, in the present invention, by optimizing the magic tire correction formula and combining it with wheel dynamics analysis, the lateral force under different tire slip angles and vertical loads can be calculated more accurately, thereby improving the accuracy and reliability of the tire model. BRIEF DESCRIPTION OF THE DRAWINGS

[0088] Figure 1 A flow chart of a distributed wheel hub driven electric vehicle lateral and vertical coordinated control method provided by an embodiment of the present invention;

[0089] Figure 2 A diagram of a lateral dynamics model provided by an embodiment of the present invention;

[0090] Figure 3 A vertical dynamics model diagram provided for an embodiment of the present invention;

[0091] Figure 4 A flow chart of three-dimensional segmented affine tire model identification provided by an embodiment of the present invention;

[0092] Figure 5 A three-dimensional segmented affine tire model diagram provided by an embodiment of the present invention;

[0093] Figure 6AFS / ASS integrated control block diagram provided by an embodiment of the present invention;

[0094] Figure 7 A flow chart of a distributed wheel hub driven electric vehicle lateral and vertical coordinated control method provided by the present invention;

[0095] Figure 8 A schematic structural diagram of a distributed wheel hub driven electric vehicle lateral and vertical coordinated control system provided by the present invention. DETAILED DESCRIPTION

[0096] The present invention provides a distributed wheel hub driven electric vehicle lateral and vertical coordinated control method, such as Figure 7 As shown, the following steps are included:

[0097] S1: Based on the unbalanced magnetic model of the hub motor, a 14-DOF vehicle model is constructed; S2: Based on the 14-DOF vehicle model combined with the optimized magic tire correction formula, a three-dimensional segmented affine tire model of lateral force-side slip angle-vertical load is constructed; S3: Based on the three-dimensional segmented affine tire model, an active front wheel steering controller based on hybrid model predictive control is constructed; at the same time, with roll stability, ride comfort and hub motor stator and rotor eccentricity as the design control objectives, an active suspension controller based on multi-constraint input and multi-constraint output is constructed; S4: Based on the front wheel angle, The phase plane and lateral load transfer rate are used to construct a coordinated control strategy for the active front wheel steering controller and the active suspension controller to achieve coordinated control of the lateral and vertical directions of the distributed wheel hub drive electric vehicle.

[0098] like Figure 8 As shown, the present invention also provides a distributed wheel hub driven electric vehicle lateral and vertical collaborative control system, including: a vehicle model building module, used to build a 14-degree-of-freedom vehicle model based on the unbalanced magnetic model of the wheel hub motor; a three-dimensional segmented affine tire model building module, used to build a three-dimensional segmented affine tire model of lateral force-side slip angle-vertical load based on the 14-degree-of-freedom vehicle model combined with an optimized magic tire correction formula; a controller building module, used to build an active front wheel steering controller based on hybrid model predictive control based on the three-dimensional segmented affine tire model; at the same time, with roll stability, ride comfort and wheel hub motor stator and rotor eccentricity as design control objectives, an active suspension controller based on multi-constraint input and multi-constraint output is built; a collaborative control module, used to build an active suspension controller based on front wheel angle, The phase plane and lateral load transfer rate are used to construct a coordinated control strategy for the active front wheel steering controller and the active suspension controller to achieve coordinated control of the lateral and vertical directions of the distributed wheel hub drive electric vehicle.

[0099] The present invention also provides a device, comprising: a memory for storing a computer program; a processor for implementing the steps of the distributed wheel hub driven electric vehicle lateral and vertical coordinated control method when executing the computer program.

[0100] When the processor executes the computer program, the steps of the lateral and vertical coordinated control of the above-mentioned distributed hub-driven electric vehicle are implemented, for example: a 14-degree-of-freedom vehicle model is constructed based on the unbalanced magnetic model of the hub motor; a three-dimensional segmented affine tire model of lateral force-side slip angle-vertical load is constructed based on the 14-degree-of-freedom vehicle model combined with the optimized magic tire correction formula; based on the three-dimensional segmented affine tire model, an active front wheel steering controller based on hybrid model predictive control is constructed; at the same time, an active suspension controller based on multi-constraint input and multi-constraint output is constructed with roll stability, ride comfort and hub motor stator and rotor eccentricity as design control objectives; based on the front wheel angle, The phase plane and lateral load transfer rate are used to construct a coordinated control strategy for the active front wheel steering controller and the active suspension controller to achieve coordinated control of the lateral and vertical directions of the distributed wheel hub drive electric vehicle.

[0101] Alternatively, the processor implements the functions of each module in the above system when executing the computer program, for example: a vehicle model construction module, which is used to construct a 14-degree-of-freedom vehicle model based on the unbalanced magnetic model of the hub motor; a three-dimensional segmented affine tire model construction module, which is used to construct a three-dimensional segmented affine tire model of lateral force-side slip angle-vertical load based on the 14-degree-of-freedom vehicle model combined with the optimized magic tire correction formula; a controller construction module, which is used to build an active front wheel steering controller based on hybrid model predictive control based on the three-dimensional segmented affine tire model; at the same time, with roll stability, ride comfort and hub motor stator and rotor eccentricity as design control objectives, an active suspension controller based on multi-constraint input and multi-constraint output is constructed; a collaborative control module is used to build an active suspension controller based on the front wheel angle, The phase plane and lateral load transfer rate are used to construct a coordinated control strategy for the active front wheel steering controller and the active suspension controller to achieve coordinated control of the lateral and vertical directions of the distributed wheel hub drive electric vehicle.

[0102] Exemplarily, the computer program may be divided into one or more modules / units, which are stored in the memory and executed by the processor to complete the present invention. The one or more modules / units may be a series of computer program instruction segments capable of completing preset functions, and the instruction segments are used to describe the execution process of the computer program in the distributed wheel hub driven electric vehicle lateral and vertical coordinated control device. For example, the computer program can be divided into a whole vehicle model building module, a three-dimensional segmented affine tire model building module, a controller building module and a collaborative control module; the specific functions of each module are as follows: a whole vehicle model building module, which is used to build a fourteen-degree-of-freedom whole vehicle model based on the unbalanced magnetic model of the hub motor; a three-dimensional segmented affine tire model building module, which is used to build a three-dimensional segmented affine tire model of lateral force-side slip angle-vertical load based on the fourteen-degree-of-freedom whole vehicle model combined with the optimized magic tire correction formula; a controller building module, which is used to build an active front wheel steering controller based on hybrid model predictive control based on the three-dimensional segmented affine tire model; at the same time, with roll stability, ride comfort and hub motor stator and rotor eccentricity as design control objectives, an active suspension controller based on multi-constraint input and multi-constraint output is built; a collaborative control module is used to build an active suspension controller based on front wheel angle, The phase plane and lateral load transfer rate are used to construct a coordinated control strategy for the active front wheel steering controller and the active suspension controller to achieve coordinated control of the lateral and vertical directions of the distributed wheel hub drive electric vehicle.

[0103] The distributed wheel hub driven electric vehicle lateral and vertical collaborative control device can be a computing device such as a desktop computer, a notebook, a PDA, and a cloud server. The distributed wheel hub driven electric vehicle lateral and vertical collaborative control device may include, but is not limited to, a processor and a memory. Those skilled in the art will understand that the above is an example of a distributed wheel hub driven electric vehicle lateral and vertical collaborative control device, and does not constitute a limitation on the distributed wheel hub driven electric vehicle lateral and vertical collaborative control device, and may include more components than the above, or a combination of certain components, or different components. For example, the distributed wheel hub driven electric vehicle lateral and vertical collaborative control device may also include input and output devices, network access devices, buses, etc.

[0104] The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may be any conventional processor, etc. The processor is the control center of the distributed wheel hub driven electric vehicle lateral and vertical coordinated control, and uses various interfaces and lines to connect the various parts of the entire distributed wheel hub driven electric vehicle lateral and vertical coordinated control device.

[0105] The memory can be used to store the computer program and / or module, and the processor realizes various functions of the distributed wheel hub driven electric vehicle lateral and vertical collaborative control device by running or executing the computer program and / or module stored in the memory, and calling the data stored in the memory.

[0106] The memory may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function (such as a sound playback function, an image playback function, etc.), etc.; the data storage area may store data created according to the use of the mobile phone (such as audio data, a phone book, etc.), etc. In addition, the memory may include a high-speed random access memory, and may also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (SmartMediaCard, SMC), a secure digital (SecureDigital, SD) card, a flash card (FlashCard), at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0107] The present invention also provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the steps of the distributed wheel hub driven electric vehicle lateral and vertical coordinated control method are implemented.

[0108] If the modules / units integrated in the distributed wheel hub driven electric vehicle lateral and vertical coordinated control system are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium.

[0109] Based on this understanding, the present invention implements all or part of the processes in the above-mentioned distributed wheel hub driven electric vehicle lateral and vertical coordinated control method, and can also be completed by instructing related hardware through a computer program, and the computer program can be stored in a computer-readable storage medium, and when the computer program is executed by a processor, the steps of the above-mentioned distributed wheel hub driven electric vehicle lateral and vertical coordinated control method can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or preset intermediate form, etc.

[0110] The computer-readable storage medium may include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal and software distribution medium, etc.

[0111] It should be noted that the content contained in the computer-readable storage medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable storage media do not include electrical carrier signals and telecommunication signals.

[0112] The present invention will be further described below in conjunction with the embodiments and drawings:

[0113] Example

[0114] The present invention provides a method for coordinated lateral and vertical control of a distributed hub-driven electric vehicle. The method can solve the technical problems in the prior art of the influence of unbalanced magnetic force on suspension and tires and the low modeling accuracy and control accuracy of a lateral controller caused by linear tires.

[0115] Among them, the following terms are explained as follows: AFS stands for active front steering system; ASS stands for active suspension system; PWA stands for piecewise affine; hMPC stands for hybrid model predictive control; DMPC stands for dual model predictive control.

[0116] like Figure 1 As shown, this embodiment provides a distributed wheel hub driven electric vehicle lateral and vertical coordinated control method, which includes four aspects, specifically including the following steps:

[0117] In a first aspect, the present invention establishes a 14-DOF model of the whole vehicle taking into account the unbalanced magnetic force of the hub motor.

[0118] Step 1, separate the stator and rotor of the wheel hub motor, regard the tire and rotor as one, and the 14-DOF vehicle model includes the vertical motion of longitudinal, lateral, yaw, pitch, roll, sprung mass, stator mass and rotor (tire) mass;

[0119] The differential equations of motion in the longitudinal and lateral directions are given by:

[0120] ma x =F x1 cos(δ)+F x2 cos(δ)+F x3 +F x4

[0121] ma y =F y1 cos(δ)+F y2 cos(δ)+F y3 +F y4

[0122] Among them, m is the total mass of the vehicle, δ is the front wheel turning angle, a x is the longitudinal acceleration, and the expression is a y is the lateral acceleration, and the expression is v x is the longitudinal speed, v y is the lateral speed, F xi is the wheel longitudinal force, F yi is the wheel lateral force, i=1,2,3,4, respectively representing the left front, right front, left rear, and right rear tires. The following formulas are similar.

[0123] The differential equations of motion for yaw, pitch, and roll are given by:

[0124]

[0125] Where ω is the yaw rate, a and b are the distances from the center of mass to the front and rear axes, and B f , B r are the front and rear axle track widths, θ is the vehicle body pitch angle, h g is the height from the center of mass to the ground, h s is the height from the roll center to the ground, g is the acceleration due to gravity, F si is the suspension force, m b is the sprung mass, φ is the body roll angle, I x , I y , I z are the moments of inertia around the vehicle coordinate axes X, Y, and Z, respectively.

[0126] The vertical motion of the sprung mass, stator mass, and tire + rotor mass can be expressed as:

[0127]

[0128] in, m si is the stator mass, m ri is the mass of tire + rotor, K si is the suspension stiffness coefficient, C si is the suspension damping coefficient, K mi is the stiffness coefficient between the stator and rotor of the motor, K ti is the tire stiffness coefficient, Z i1 is the vertical displacement of the tire + motor rotor, Z i2 is the vertical displacement of the motor stator, Z i3 is the vertical displacement of the sprung mass, Z 0i is the roughness of the four-wheel road surface, F i is the main force of the suspension, F Ei It is the vertical unbalanced excitation of the hub motor.

[0129] Step 2, establish the magic tire correction formula that introduces the road adhesion coefficient, as follows:

[0130]

[0131] Among them, F y is the wheel lateral force, μ is the road adhesion coefficient, α is the tire slip angle, and B y Stiffness factor, C y is the shape factor, D y is the peak factor, E y is the curvature factor.

[0132] Based on wheel dynamics analysis, the tire slip angle can be approximately expressed as:

[0133]

[0134] The vertical load of the tire can be approximately expressed as:

[0135]

[0136] Among them, F z is the vertical load on the tire.

[0137] The 14-DOF vehicle model of this embodiment is as follows: Figure 2 and Figure 3 shown.

[0138] The relevant parameters of the vehicle 14-DOF model of this embodiment are shown in Table 1:

[0139] Table 1 Vehicle 14-DOF model parameters

[0140]

[0141]

[0142] In the second aspect, the present invention designs an identification method for a three-dimensional piecewise affine tire model of lateral force-sideslip angle-vertical load.

[0143] The lateral force-sideslip angle-vertical load data obtained by the magic tire correction formula is used as the original data set Ω. The improved K-plane clustering is used to cluster the original data Ω. The steps are as follows:

[0144] Step 1: Divide the original data set into uniform n×m sub-data sets Ω along the x and y axes. j , (j=1…n×m), the least squares estimation method is used to fit each subset of data sets Ω j The plane parameters are expressed as follows:

[0145]

[0146] in, Plane parameters for each subdataset.

[0147] Assuming there are p initial clustering planes, randomly select a plane as the first initial plane, recorded as:

[0148] M 1 ={a 1 ,b 1 ,c 1 ,d 1};

[0149] Step 2: The determined initial plane {M 1 ,M 2 ,…,M q The average value of the parameters is recorded as:

[0150]

[0151] The remaining plane parameters and The difference can be expressed as:

[0152]

[0153] Wherein, q is the number of initial planes that have been determined, and q≤p.

[0154] Select the maximum difference, that is, maxD i (Ω j ,M i ) as the next initial plane;

[0155] Step 3: Repeat step 2 until p initial planes {M 1 ,M 2 ,…,M p};

[0156] The above steps are to complete the initialization of K-plane clustering.

[0157] Step 4: Use p planes to cluster the tire data points. The corresponding data subset is {J 1 ,J 2 ,…,J p}, the data classification rules are as follows:

[0158]

[0159] Among them, Dis i-j Represents a point (x i ,y i ,z i ) to the jth plane, [a j ,b j ,c j ,d j ] is the parameter of the jth plane. The role of the rule is to transform the data point (x i ,y i ,z i ) is classified to the plane J that is closest to it j .

[0160] Step 5: Identify and reclassify the boundary outliers. i ,y i ,z i ) to a certain plane whose distance is less than a certain threshold is greater than or equal to 2, it is judged as a boundary outlier point, that is, it meets the following conditions:

[0161] A=[Dis i-1 ,Dis i-2 ,...,Dis i-p ]

[0162] ifnum(A(i)≤ε boundary )≥2

[0163] Then point(i)is boundary outlier

[0164] Where A is the point (x i ,y i ,z i ) to each plane, ε boundary is the boundary outlier threshold.

[0165] After identifying the boundary outliers, the minimum distance between the outliers and the centers of p data subsets is used to reclassify the outliers.

[0166] Then, according to the data subset corresponding to each plane, LSM is used to re-identify its plane parameters and repeat the data classification, boundary outlier point identification and classification steps. Through multiple cycles of iteration, a three-dimensional segmented affine tire sub-model data set is obtained.

[0167] Step 6: According to the data subset corresponding to each plane, use LSM to re-identify its plane parameters and repeat steps 4-5, and obtain a three-dimensional piecewise affine model through multiple cycles of K-plane clustering; that is, use LSM to re-identify its plane parameters and repeat data classification, boundary outlier identification and classification steps, and obtain a three-dimensional piecewise affine tire sub-model data set through multiple cycles.

[0168] Step 7: Identify and reclassify the cutting anomaly points, and convert the data set with cutting anomaly points into Project the data onto the x, y or z axis and transform it into one-dimensional data. Use the k-means clustering method to perform cluster analysis, and identify outlier points by the proportion of the number of sub-data sets after clustering:

[0169]

[0170] in, The sub-datasets obtained by clustering, e = 1, 2, ..., p, j is the number of categories of k-means clustering, which should be based on the data set The distribution of size(*) indicates the number of data in the data set, ξ cut is the threshold for cutting outliers.

[0171] After identifying the boundary outliers, the outliers are secondary classified using the minimum distance from the outliers to the centers of p data subsets to obtain the final tire sub-model data set. And use LSM to re-identify its plane parameters.

[0172] Step 8: Use support vector machine to estimate the coefficient matrix of the interface of the piecewise affine tire submodel. The steps are as follows:

[0173] First, find two adjacent data sets that satisfy: m i ,m j The data subsets and The data points of two adjacent subsets are taken as input data and recorded as X = {X 1 ,X2 ,…X N}, each sample of the input data contains multiple features, which constitute the feature space X i ={x 1 ,x 2 ,…x n}, given the learning goal y = {y 1 ,y 2 ,…y N},make The sample data label in the subset is y(i) = 1, indicating the positive class; The sample data label in the subset is y(j) = -1, indicating the negative class. In order to prevent the existence of data that cannot be linearly separable, a slack variable υ is introduced to construct the following convex quadratic programming problem:

[0174]

[0175] 0≤α i ≤υ,i=1,2,…N

[0176] Among them, α i is a Lagrange multiplier, and α i ≥0, κ(*·*) is the kernel function, and N is the number of samples in the training data set. Solving the above equation yields the optimal solution Finally, the hyperplane coefficient matrix is ​​solved as follows:

[0177]

[0178] Finally, the PWA tire model can be expressed as: y =θ i-1 α+θ i-2 F z +θ i-3 ,i=(1,2,…,p)

[0179] Among them, θ i-1 ,θ i-2 and θ i-3 are the identified PWA tire sub-model parameters respectively.

[0180] It should be noted that traditional K-plane clustering clusters data points based on the minimum distance from each data point to the plane. However, when data points are very close to multiple planes, clustering based solely on distance will produce boundary outliers, that is, indivisible data points. In addition, the plane after K-plane clustering has infinite extension, which may cause multiple planes to cut each other in three-dimensional space, thus generating cutting outliers. Therefore, these two types of outliers need to be identified and classified separately.

[0181] The improved magic tire three-dimensional segmented affine process of this embodiment is as follows Figure 4 As shown, the affine result is Figure 5 As shown, it can be seen that the three-dimensional surface is divided into 9 linear planes, the front wheel tire cornering force contains 9 linear control areas, and the rear wheel tire cornering force also contains 9 linear control areas. Therefore, the piecewise linear system has a total of 81 linear control areas.

[0182] In a third aspect, the present invention designs an active front wheel steering controller of hMPC based on the established piecewise affine tire model.

[0183] Step 1: Based on the 14-DOF vehicle model, assuming that the front wheel turning angle is very small and the longitudinal forces of the four wheels are equal, the lateral dynamics model is simplified:

[0184] ∑F Y =F y1 +F y2 +F y3 +F y4

[0185] ∑M z =a(F y1 +F y2 )-b(F y3 +F y4 )

[0186] The desired state of the vehicle is in steady state, that is Taking into account the influence of road adhesion, based on the vehicle linear two-degree-of-freedom dynamic model, the final ideal state parameter ω is obtained d and β d :

[0187]

[0188] Among them, v x is the longitudinal speed, L=a+b is the wheelbase of the vehicle, is the vehicle stability factor, k f and k r are the cornering stiffness of the front and rear tires, respectively.

[0189] Let x = [Δβ, Δω] T is the state vector, y=[Δβ,Δω] T is the output, u=Δδ f is the input quantity, where Δω=ω-ω d , Δβ=β-β d , Δδ f is the additional active front wheel steering angle. The error state space equation is obtained, and its expression is:

[0190]

[0191] Among them, l and r are the PWA tire sub-models of the left and right front wheels, l=(1,2,…,p), r=(1,2,…,p), A lr ,B lr ,f lr ,C 1 ,D 1 ,g lr The coefficient matrix expression of the error state space equation is:

[0192]

[0193] Among them, θ l-i-m and θ r-i-m are the left and right PWA tire model parameters, i=(1,2,…,p), m=(1,2,3), K r is the cornering stiffness of the left and right rear tires.

[0194] Step 2: Introduce auxiliary discrete variables σ n ∈{0,1},n=(1,2,…,p 2 ), combined with the IF-THEN-ELSE rule, the error state space equation is rewritten as:

[0195]

[0196] The above formula indicates that when the system enters the nth control area, it is equivalent to σ n =1, otherwise, σ n = 0. Therefore, the continuous auxiliary variable z is introduced n (k) = [A n x(k)+B n u(k)+f n ]·σ n (k) and satisfy the following Boolean relation:

[0197] σ n (k)=0→z n (k) = 0

[0198] σ n (k)=1→z n (k) = A n x(k)+B n u(k)+f n

[0199] Among them, “→” is a Boolean operator of inclusion relationship.

[0200] Define M and m as continuous auxiliary variables z n The upper and lower bounds of (k), i.e. the maximum and minimum values:

[0201]

[0202] According to the conversion relationship between the logical proposition expression of the multiplication of logical variables and continuous problems and linear inequalities, the mixed integer inequality constraints shown below can be obtained:

[0203] z n (k)≤Mσ n (k)

[0204] z n (k)≥mσ n (k)

[0205] z n (k)≤A n x(k)+B n u(k)+f n -m(1-σ n (k))

[0206] z n (k)≥A n x(k)+B n u(k)+f n -M(1-σ n (k))

[0207] The MLD prediction model can be expressed as follows:

[0208] x(k+1)=A * 1 x(k)+B * 1 u(k)+B * 2 σ(k)+B * 3 z(k)

[0209] y(k)=C 1 * x(k)+D * 1 u(k)+D * 2 σ(k)+D * 3 z(k)

[0210] E * 1 σ(k)+E * 2 z(k)≤E * 3 x(k)+E * 4 u(k)+E * 5

[0211] The expression for future output in the prediction time domain is:

[0212]

[0213] Among them, Y(t), U(t), Δ(t), and Z(t) are the sequences of system output, control variables, auxiliary discrete variables, and auxiliary continuous variables in the prediction time domain, respectively. It is the coefficient matrix obtained recursively based on the solution of the MLD prediction model.

[0214] Step 3, the optimization objective function can be expressed as:

[0215]

[0216] Among them, N p is the prediction time domain, N c To control the time domain.

[0217] Solve the following quadratic programming problem to obtain the additional turning angle to be controlled;

[0218]

[0219] The correlation matrix expression is:

[0220]

[0221] in, and are the weight matrices of system output and control increment in the prediction domain and control domain respectively.

[0222] It should be noted that when designing the AFS controller, in order to prevent the increase in calculation amount and slow controller response due to overly complex hybrid logic, the present invention only uses the PWA tire model for the left and right front wheels and the traditional linear model for the rear wheels.

[0223] Fourthly, the present invention takes roll stability, ride comfort and hub motor stator-rotor eccentricity as design control objectives, and uses DMPC to establish a multi-constraint input and multi-constraint output active suspension controller. And designs a coordinated control strategy for active front wheel steering / active suspension:

[0224] Step 1: Simplify the 14-DOF vertical model of the vehicle to obtain an 8-DOF left and right half vehicle suspension model, whose state space expression is:

[0225]

[0226] y roll =C 2 x+D2 u+Ew

[0227]

[0228] The state variables are The system input is u = [F1, F2], which is the active force output by the active suspension system. The present invention regards the unbalanced magnetic force and the road surface roughness as external interference, so w = [FE1, FE2, Z 01 ,Z 02 ], the system output is y roll =[φ], that is, the body roll angle, That is, the eccentricity of the stator and rotor of the left and right wheel hub motors, the dynamic load of the tires, and the acceleration of the vehicle body.

[0229] Step 2: Here we take the MPC controller derivation of roll stability as an example. The same is true for ride comfort. Set:

[0230]

[0231] The forward Euler method is used to discretize the above state space equation to obtain:

[0232]

[0233] The matrix expressions are:

[0234]

[0235] in, T = 0.01 is the simulation step length, Nx is the number of state quantities, Nu is the number of control quantities, and Nz is the number of interference quantities. Therefore, the output matrix of the system at the future moment can be expressed as:

[0236] Y(t)=Ψξ(t|t)+ΘΔU(t)+ΩW(t|t)

[0237] The coefficient matrices in the formula can be obtained by iterative derivation.

[0238] Step 3, the optimization objective function can be expressed as:

[0239]

[0240] Among them, the matrices Q and R are the weight matrices of system output and control increment respectively, and ε is the relaxation factor to prevent the controller from having no solution.

[0241] Since the active suspension actuator can only generate limited active force, and in order to prevent the suspension from frequently hitting the limit block, the suspension travel should be specified within a safe range. In addition, the eccentricity of the hub motor stator and rotor cannot exceed its structural limit. Finally, considering the vehicle stability performance, the tire dynamic load should also be constrained. Therefore, the following constraints are set for the active suspension system:

[0242] u min ≤u≤u max ,Δu min ≤Δu≤Δu max

[0243] |Z ri -Z si |≤Zh max ,|Z si -Z bi |≤Zs max

[0244]

[0245] Among them, u min =-2000N,u max =2000N,Δu min =-200N, Δu max =-200N, Zh max =0.0008, Zs max =0.01.

[0246] Therefore, the optimization problem can finally be expressed as the following quadratic programming problem:

[0247]

[0248] stlb≤[ΔU,ε]≤ub

[0249] A_cons*[ΔU,ε]≤b_cons

[0250] The correlation matrix expression is:

[0251]

[0252] Among them, M1 is the upper limit of relaxation factor, A_I, Ut, ΔU min , ΔU max , U min , U max is the correlation matrix obtained by iterative derivation.

[0253] Considering the coordination strategy of roll stability index and ride comfort index, the final ASS optimization objective function can be expressed as:

[0254]

[0255] Among them, J 1 and ε J1 are the roll stability objective function and its coordination parameters, J 2 and ε J2 are the ride comfort objective function and its coordination parameters respectively.

[0256] Step 4: In order to coordinate the AFS and ASS systems, the present invention uses The phase plane makes a regional judgment on the vehicle stability. The stable region Ψ is obtained by the double straight line method, and the boundary of the stable region can be expressed as:

[0257]

[0258] in, is the stability boundary coefficient;

[0259] At the same time, the load transfer rate is introduced to evaluate the roll stability of the vehicle, and its expression is as follows:

[0260]

[0261] The above formula indicates that the closer LTR is to -1 and 1, the greater the body roll. When LTR=1 or LTR=-1, the vehicle rolls over.

[0262] The present invention designs the following coordination strategies according to the vehicle driving conditions:

[0263] (1) When the vehicle is traveling in a straight line, AFS does not work, and ASS works to optimize ride comfort, i.e., p 1 =0, p 2 =1;

[0264] (2) When the vehicle is in a turning condition, both AFS and ASS are working, that is, p 1 =1, p 2 Combining LTR indicators and β- The design is based on the phase plane stability boundary, and its expression is:

[0265]

[0266] Among them, LTR * =0.5 and The formula shows that when the vehicle state is in the stable area and the LTR is lower than the rollover threshold, the active suspension partially intervenes; when the vehicle state is outside the stable area or the LTR is greater than the rollover threshold, the active suspension fully intervenes.

[0267] The controller designed by the present invention includes an AFS controller, an ASS controller and a coordination strategy between the two. The AFS controller establishes an hMPC controller based on a three-dimensional piecewise affine tire model, and obtains the reference center of mass sideslip angle and yaw angular velocity from a two-degree-of-freedom bicycle model, and then calculates the additional turning angle of the vehicle to ensure vehicle stability; the ASS controller uses roll stability, ride comfort, i.e. tire dynamic load, vehicle body acceleration, and hub motor stator and rotor eccentricity as suspension design control targets, and designs an active suspension controller with multi-constraint input and multi-constraint output based on DMPC. The coordinated control strategy is based on The phase plane and lateral load transfer rate are designed, and the two controllers complement each other to improve the dynamic performance of the vehicle.

[0268] It can be seen that the distributed wheel hub driven electric vehicle lateral and vertical coordinated control method provided by this embodiment has the following advantages: (Ⅰ) A distributed integrated lateral and vertical control method is established, which effectively solves the coupling between different chassis electronic control systems, improves the vehicle's handling stability, and reduces the impact of unbalanced magnetic forces on the suspension and tires, which is specifically reflected in improving the vehicle's roll stability, ride comfort, and wheel hub motor performance. (Ⅱ) A method for identifying the three-dimensional piecewise affine tire model of lateral force-sideslip angle-vertical load is proposed. It effectively solves the problem that the linear tire model is difficult to describe the nonlinear characteristics of the tire under extreme working conditions, and improves the control ability of the lateral controller. (Ⅲ) The coordination strategy of the distributed integrated lateral and vertical control method designed by the present invention comprehensively considers the needs of the vehicle under different working conditions, combined with The phase plane and lateral load transfer rate coordinate the control of AFS / ASS, which can improve the dynamic performance of the vehicle and enhance its anti-rollover capability to a certain extent.

[0269] In summary, the present invention proposes a method for lateral and vertical coordinated control of distributed hub-driven electric vehicles, which includes a vehicle model module, a three-dimensional piecewise affine tire model identification, an AFS controller, an ASS controller, and a coordinated control strategy. A fourteen-degree-of-freedom vehicle model considering the unbalanced electromagnetic force of the hub motor is established; secondly, in order to improve the modeling accuracy of the lateral controller, a piecewise affine (PWA) method is adopted to establish a three-dimensional piecewise affine tire model based on lateral force-side slip angle-vertical load, and a hybrid logic dynamic model is established, and an active front wheel steering controller based on hybrid model predictive control (hMPC) is designed; considering the influence of the unbalanced magnetic force generated by the hub motor on the suspension system and tires, an active suspension controller based on dual model predictive control (DMPC) is designed, and a controller based on front wheel angle, AFS / ASS coordination strategy for phase plane and lateral load transfer rate. It solves the technical problems of the influence of unbalanced magnetic force on suspension and tires and low modeling accuracy and control accuracy of lateral controller caused by linear tires in the prior art, and improves the vehicle's handling stability, roll stability and ride comfort.

[0270] The above embodiment is only one of the implementation methods that can realize the technical solution of the present invention. The scope of protection claimed by the present invention is not limited only to this embodiment, but also includes changes, replacements and other implementation methods that can be easily thought of by any technician familiar with the technical field within the technical scope disclosed by the present invention.

Claims

1. A distributed wheel hub driven electric vehicle lateral and vertical coordinated control method, characterized in that: include: Construct a 14-DOF vehicle model based on the unbalanced magnetic force model of the hub motor; Based on the 14-DOF vehicle model and the optimized magic tire correction formula, a three-dimensional piecewise affine tire model of lateral force-sideslip angle-vertical load is constructed; Based on the three-dimensional piecewise affine tire model, an active front wheel steering controller based on hybrid model predictive control is built; at the same time, with roll stability, ride comfort and wheel hub motor stator and rotor eccentricity as the design control objectives, an active suspension controller based on multi-constraint input and multi-constraint output is built; Based on the front wheel angle, Phase plane and lateral load transfer rate, build coordinated control strategy of active front wheel steering controller and active suspension controller to achieve coordinated control of distributed wheel hub drive electric vehicles in lateral and vertical directions; The unbalanced magnetic model of the hub motor is obtained by separating the stator and rotor of the hub motor and considering the tire and the rotor as an integral whole; the 14-DOF vehicle model includes longitudinal, lateral, yaw, pitch, roll, and vertical motion of the sprung mass, stator mass, rotor and tire mass; wherein, The differential equations of motion in the longitudinal and lateral directions are given by: in, is the total mass of the vehicle, is the front wheel turning angle, is the longitudinal acceleration, and the expression is , is the lateral acceleration, and the expression is , is the longitudinal speed, is the lateral speed, is the wheel longitudinal force, is the wheel lateral force, , respectively represent the left front tire, right front tire, left rear tire, and right rear tire; is the yaw angular velocity; The differential equations of motion for yaw, pitch, and roll are given by: in, , are the distances from the center of mass to the front and rear axes, , Respectively, the front and rear axle track. is the vehicle body pitch angle, is the height from the center of mass to the ground, is the height from the roll center to the ground, is the acceleration due to gravity, is the suspension force, is the sprung mass, is the body roll angle, , , Around the vehicle coordinate axis , , The moment of inertia of the shaft; The vertical motion of the sprung mass, stator mass, and tire and rotor mass can be expressed as: in, , is the stator mass, is the mass of the tire and rotor, is the suspension stiffness coefficient, is the suspension damping coefficient, is the stiffness coefficient between the stator and rotor of the motor, is the tire stiffness coefficient, is the vertical displacement between the tire and the motor rotor, is the vertical displacement of the motor stator, is the vertical displacement of the sprung mass, The roughness of the road surface for four wheels. The main force of the suspension, It is the vertical unbalanced excitation for the hub motor; The optimization steps of the magic tire correction formula are as follows: Introduce the road adhesion coefficient into the magic tire correction formula, the specific formula is as follows: in, is the wheel lateral force, is the road adhesion coefficient, is the tire slip angle, Stiffness factor, is the shape factor, is the peak factor, is the curvature factor; Based on wheel dynamics analysis, the tire slip angle is expressed as: The vertical tire load is expressed as: in, is the vertical load of the tire; L is the wheelbase of the vehicle; The construction process of the three-dimensional piecewise affine tire model of lateral force-sideslip angle-vertical load is as follows: The lateral force-sideslip angle-vertical load data obtained by the optimized magic tire correction formula is used as the original data set , using improved K-plane clustering to cluster the original data To perform clustering, the steps are as follows: Step 1, follow The axis divides the original dataset into even Molecular Datasets , using the least squares estimation method to fit each subset of data sets The plane parameters are expressed as follows: in, Plane parameters for each sub-dataset; Assume that the initial clustering plane has A plane is randomly selected as the first initial plane, recorded as: ； Step 2: Determine the initial plane The average value of the parameters is recorded as: The remaining plane parameters and The difference can be expressed as: in, is the number of initial planes that have been determined, and ; Select the one with the largest difference As the next initial plane; Step 3, repeat step 2 until you select Initial plane , to complete the initialization of K-plane clustering; Step 4, use The tire data points are clustered using planes, and the corresponding data subset is , the data classification rules are as follows: in, Indicate point To The distance between the planes, For the The parameters of the plane, the data classification rule is used to classify the data points Classify to the nearest plane ; Step 5: Identify and reclassify the boundary outliers. When the number of points whose distance to a plane is less than a certain threshold is greater than or equal to 2, it is judged as a boundary outlier point and meets the following conditions: in, For point The distance to each plane, is the boundary outlier threshold; After identifying the boundary outliers, use the outliers to The distance between the centers of the data subsets is minimized to reclassify the outliers. According to the data subsets corresponding to each plane, LSM is used to re-identify the plane parameters and repeat the data classification, boundary outlier identification and classification steps. Through multiple cycles of iteration, a three-dimensional segmented affine tire sub-model data set is obtained. ; Step 6: According to the data subset corresponding to each plane, use LSM to re-identify its plane parameters and repeat steps 4-5 to obtain a three-dimensional piecewise affine model through multiple cycles of K-plane clustering. Step 7: Identify and reclassify the cutting anomaly points, and convert the data set with cutting anomaly points into Project the data onto the x, y or z axis and transform it into one-dimensional data. Use the k-means clustering method to perform cluster analysis, and identify outlier points by the proportion of the number of sub-data sets after clustering: in, is the sub-dataset obtained by k-means clustering, , is the number of categories for k-means clustering, which should depend on the data set depends on the distribution of Indicates the number of data in the data set. is the threshold for cutting outliers; After identifying the cutting outlier points, use the outlier points to The distance between the centers of the data subsets is minimized to perform secondary classification of outliers and obtain the final tire sub-model data. and use LSM to re-identify its plane parameters; Step 8, using a support vector machine to estimate the coefficient matrix of the interface of the three-dimensional piecewise affine tire submodel; Finally, the three-dimensional piecewise affine tire model is expressed as: in, are the identified three-dimensional piecewise affine tire model parameters respectively.

2. A distributed wheel hub driven electric vehicle lateral and vertical coordinated control method according to claim 1, characterized in that: Based on the three-dimensional piecewise affine tire model, the specific steps of building an active front wheel steering controller based on hybrid model predictive control include: Based on the 14-DOF vehicle model, the lateral dynamics model is simplified to obtain the error state space equation. The error state space equation is transformed by introducing auxiliary discrete variables and combining IF-THEN-ELSE rules; Based on the three-dimensional piecewise affine tire model and the transformed error state space equation, the quadratic programming problem is solved to obtain the additional steering angle of control to complete the construction of the active front wheel steering controller based on hybrid model predictive control.

3. The method for lateral and vertical coordinated control of a distributed hub-driven electric vehicle according to claim 1, characterized in that: The specific steps to build an active suspension controller based on multi-constraint input and multi-constraint output include: The 14-DOF vehicle model is simplified to obtain a half-vehicle suspension model; Based on the half-vehicle suspension model, the roll stability, ride comfort and the eccentricity of the hub motor stator and rotor are taken as the design control objectives. The quadratic programming problem is solved to obtain the objective functions and coordination parameters corresponding to the roll stability and ride comfort, so as to complete the construction of an active suspension controller based on multi-constraint input and multi-constraint output.

4. A distributed wheel hub driven electric vehicle lateral and vertical coordinated control system, used to implement the steps of the distributed wheel hub driven electric vehicle lateral and vertical coordinated control method according to any one of claims 1 to 3, characterized in that: include: The vehicle model building module is used to build a 14-DOF vehicle model based on the unbalanced magnetic force model of the hub motor; The three-dimensional piecewise affine tire model building module is used to build a three-dimensional piecewise affine tire model of lateral force-sideslip angle-vertical load based on a 14-DOF vehicle model combined with an optimized magic tire correction formula; The controller building module is used to build an active front wheel steering controller based on a hybrid model predictive control based on a three-dimensional piecewise affine tire model; at the same time, with roll stability, ride comfort and wheel hub motor stator and rotor eccentricity as design control objectives, an active suspension controller based on multi-constraint input and multi-constraint output is built; Cooperative control module for The phase plane and lateral load transfer rate are used to construct a coordinated control strategy for the active front wheel steering controller and the active suspension controller to achieve coordinated control of the lateral and vertical directions of the distributed wheel hub drive electric vehicle.

5. A device, characterized in that: include: Memory for storing computer programs; A processor is used to implement the steps of the distributed wheel hub driven electric vehicle lateral and vertical coordinated control method as described in any one of claims 1 to 3 when executing the computer program.

6. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, it is used to implement the steps of the distributed wheel hub driven electric vehicle lateral and vertical coordinated control method as described in any one of claims 1-3.

Images