Wheel type vehicle design parameter and control parameter collaborative optimization method and related device

Abstract

The invention discloses a wheeled vehicle design parameter and control parameter collaborative optimization method and a related device, and relates to the technical field of wheeled vehicle design and development. A dynamical model of a wheeled vehicle is firstly established, and then a collaborative optimization model of the wheeled vehicle is established based on the dynamical model; optimization variables of the collaborative optimization model comprise state parameters, design parameters and control parameters, the collaborative optimization model comprises a cost function and constraint conditions, the constraint conditions comprise dynamic constraints, variable value upper and lower bound constraints, path constraints and boundary constraints, the dynamic constraints are constraints established based on the dynamic model, and the path constraints are constraints established based on the dynamic model. And finally, carrying out optimization solution on the collaborative optimization model to obtain optimal design parameters and optimal control parameters of the wheeled vehicle. According to the invention, collaborative optimization is carried out on the design parameters and the control parameters of the wheeled vehicle based on the dynamical model, so that the optimal performance of the vehicle is realized.

Description

Technical Field

[0001] The present application relates to the technical field of wheeled armored vehicle design and development, and in particular to a method for collaboratively optimizing design parameters and control parameters of an 8×8 wheeled armored vehicle based on a dynamic model and related devices. Background Art

[0002] Among all-electric land warfare platforms, wheeled vehicles have higher driving speeds than tracked vehicles and faster maneuverability on flat, hardened roads, which makes wheeled vehicles more suitable for tasks of rapid maneuver, rapid deployment, and rapid response. Moreover, their structure is relatively simple, easier to maintain and repair, and reduces the cost and complexity of maintenance. The structure is light and relatively easy to transport by air, which may be an important consideration when military forces need to be quickly projected. As an important source of power for all-electric land warfare platforms, the performance of the electric drive system directly determines the level of performance that can be achieved by the vehicle. Among various types of electric drive systems, the wheel hub motor drive system has many advantages, which simplifies the mechanical transmission device and realizes the direct transmission of driving force to the wheel; on the other hand, each electric wheel can be driven and braked independently, which increases the redundancy of the system and provides more degrees of freedom for dynamic control. Each wheel hub can be controlled individually, thereby achieving faster and more accurate turning and vehicle stability. It also improves the modularization and universal design level of vehicles. Vehicles with different tactical functions can use the same chassis system. The electric wheels are universal and interchangeable, which can effectively reduce the cost of early development and later maintenance, and also reduce the logistics supply pressure on the battlefield. In summary, distributed wheel hub electric drive armored vehicles are one of the important development directions of the new generation of military vehicles. It is necessary to develop new equipment for wheeled armored vehicles.

[0003] At present, the development of wheeled armored vehicles is still mostly based on experience-based design and development methods. The design parameters are calculated according to the requirements, and the parameters are optimized based on the static model. After the optimization is completed, the control is carried out according to the specific working conditions, resulting in small optimization space, poor dynamic operation effect and other problems. In addition, each subsystem is designed separately, which limits the overall performance of the vehicle. The existing optimization method based on evolutionary algorithms can consider the vehicle dynamic model, but because the gradient information is not applied, the parameter optimization efficiency is low. The existing optimization algorithms based on dynamic models using dynamic programming or indirect methods are extremely sensitive to the scale of problem solving and the initial solution, and it is difficult to achieve rapid optimization of multiple parameters. In addition, the vehicle design parameters are usually coupled with the control parameters to affect the overall performance of the vehicle, but the existing research only optimizes the design parameters or control parameters, which greatly reduces the global optimization space and limits the space for improving vehicle performance. The driving conditions of four-wheeled vehicles are relatively simple, the optimization problem is solved quickly, and the static optimization design parameters are more practical. The driving conditions of 8×8 wheeled vehicles are complex, and the performance is related to both the design parameters and the control parameters. The coupling of the two affects the vehicle's passability and maneuverability. It is difficult to achieve optimal performance by separating design and control. Summary of the invention

[0004] The purpose of this application is to provide a method and device for collaboratively optimizing design parameters and control parameters of a wheeled vehicle, which can collaboratively optimize the design parameters and control parameters to achieve the optimal performance of the wheeled vehicle.

[0005] To achieve the above objectives, this application provides the following solutions:

[0006] In a first aspect, the present application provides a method for collaboratively optimizing design parameters and control parameters of a wheeled vehicle, comprising:

[0007] Establish a dynamic model of wheeled vehicles;

[0008] A collaborative optimization model of the wheeled vehicle is established based on the dynamic model; the optimization variables of the collaborative optimization model include state parameters, design parameters and control parameters; the collaborative optimization model includes a cost function and constraints, the constraints include dynamic constraints, upper and lower bounds of variable values, path constraints and boundary constraints, and the dynamic constraints are constraints established based on the dynamic model;

[0009] The collaborative optimization model is optimized and solved to obtain the optimal design parameters and optimal control parameters of the wheeled vehicle.

[0010] In a second aspect, the present application provides a computer device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-mentioned method for collaborative optimization of design parameters and control parameters of a wheeled vehicle.

[0011] In a third aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-mentioned method for collaborative optimization of design parameters and control parameters of a wheeled vehicle.

[0012] In a fourth aspect, the present application provides a computer program product, including a computer program, which, when executed by a processor, implements the above-mentioned method for collaborative optimization of design parameters and control parameters of a wheeled vehicle.

[0013] According to the specific embodiments provided in this application, this application discloses the following technical effects:

[0014] The present application provides a method and related device for collaborative optimization of design parameters and control parameters of a wheeled vehicle. A dynamic model of the wheeled vehicle is first established, and then a collaborative optimization model of the wheeled vehicle is established based on the dynamic model. The optimization variables of the collaborative optimization model include state parameters, design parameters and control parameters. Finally, the collaborative optimization model is optimized and solved to obtain the optimal design parameters and optimal control parameters of the wheeled vehicle. The present application collaboratively optimizes the design parameters and control parameters of the wheeled vehicle based on the dynamic model, which can achieve the optimal performance of the vehicle. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0016] Figure 1 A schematic flow chart of a method for collaboratively optimizing design parameters and control parameters of a wheeled vehicle provided in Example 1 of the present application;

[0017] Figure 2 A schematic diagram of the principle of a method for collaboratively optimizing design parameters and control parameters of a wheeled vehicle provided in Example 1 of the present application;

[0018] Figure 3 A schematic diagram of a wheeled vehicle dynamics model framework provided in Example 1 of the present application;

[0019] Figure 4 A schematic diagram of solving the Jacobian matrix provided in Example 1 of the present application. DETAILED DESCRIPTION

[0020] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0021] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below with reference to the accompanying drawings and specific implementation methods.

[0022] Example 1

[0023] like Figure 1 and Figure 2 As shown, this embodiment provides a method for collaboratively optimizing design parameters and control parameters of a wheeled vehicle, taking an 8×8 wheeled armored vehicle as an example, including:

[0024] S1: Establish a dynamic model of wheeled armored vehicles.

[0025] S2: A collaborative optimization model of a wheeled armored vehicle is established based on the dynamic model; the optimization variables of the collaborative optimization model include state parameters, design parameters and control parameters; the collaborative optimization model includes a cost function and constraints, and the constraints include dynamic constraints, upper and lower bounds of variable values, path constraints and boundary constraints, and the dynamic constraints are constraints established based on the dynamic model.

[0026] S3: Optimizing and solving the collaborative optimization model to obtain optimal design parameters and optimal control parameters of the wheeled armored vehicle.

[0027] By implementing the above steps S1 to S3, it is possible to simultaneously optimize the design parameters and control parameters of the wheeled armored vehicle based on its dynamic model. The optimal design parameters can be used for vehicle design, and the optimal control parameters can be used as a reference for subsequent online control strategy formulation, which solves the problem that existing research only optimizes design parameters or control parameters, greatly reducing the global optimization space and limiting the space for improving vehicle performance.

[0028] The electric transmission structure of wheeled armored vehicles is different from that of tracked armored vehicles. A prominent feature of wheeled armored vehicles is that they have many axles and models. According to the number of axles, they can be divided into 6×6, 8×8, 10×10 and other models. As a major member of military vehicles, 8×8 wheeled armored vehicles are mainly responsible for military reconnaissance, personnel transportation, logistics support and other tactical functions. The versatility and adaptability of 8×8 wheeled armored vehicles in combat make them an important part of modern armies and can play a key role in various tasks and battlefield environments. Therefore, it is necessary to develop new equipment for 8×8 wheeled armored vehicles. However, the driving conditions of 8×8 wheeled armored vehicles are complex, and the performance is related to both design parameters and control parameters. The coupling of the two affects the vehicle's passability and maneuverability. It is difficult to achieve optimal performance by separating design and control. Therefore, this embodiment aims to provide a method for collaborative optimization of design parameters and control parameters of wheeled armored vehicles, and collaboratively optimizes the design parameters and control parameters to achieve the optimal performance of wheeled armored vehicles.

[0029] The present invention firstly uses the knowledge related to Lagrangian dynamics to perform high-precision, vectorized modeling of wheeled armored vehicles, including vehicle-mounted turrets, suspension models, tire models, etc. The specific process of dynamic modeling of wheeled armored vehicles based on Lagrangian dynamics is as follows:

[0030] The input of the multi-degree-of-freedom vehicle model is the steering wheel angle and wheel torque. The tire rotation is driven by the wheel torque and the longitudinal ground force. The input of the tire model module is the wheel rotation angular velocity and wheel center velocity, and the output is the tire force and torque. The vertical movement of the unsprung mass is generated by the vertical force of the tire and the suspension force. The movement of the vehicle body is generated by the longitudinal and lateral tire forces, aerodynamics, and suspension forces. The input and output interfaces of each submodule of the vehicle dynamics model are as follows: Figure 3 shown.

[0031] In order to improve the computational efficiency, vectorized programming method is applied to establish a vehicle dynamics model that supports vectorized computation. Figure 3 It is the whole vehicle dynamics model framework of wheeled armored vehicles, including the interaction between the body, suspension, unsprung mass and tire models.

[0032] The multi-degree-of-freedom vehicle model represents the dynamic behavior of a simplified vehicle composed of multiple rigid components. The vehicle body can move in the longitudinal, lateral and vertical directions, and rotate around the xyz axes. The 8 tires have 8 rotational degrees of freedom, and due to the existence of the suspension system, the 8 tires have 8 vertical degrees of freedom.

[0033] In the modeling process, O-XYZ represents the earth coordinate system, O1-X1Y1Z1 represents the vehicle coordinate system, and the generalized coordinates of multiple degrees of freedom are expressed by vectors as follows:

[0034]

[0035] Where: q represents the generalized coordinates of the vehicle model; q bv represents the position vector of the center of mass of the vehicle body, X A ,Y A ,Z A is the absolute displacement of the center of mass along the global coordinate system; q bω represents the rotation angle of the vehicle body’s center of mass, φ, ψ are the rotation angles of the vehicle body around the three axes of the global coordinate system; q uθ represents the angular velocity sequence generated by the rotation of the eight tires on both sides, θ i represents the angular velocity of tire rotation; q uz represents the vertical displacement sequence of the eight tires on both sides, z i Indicates the vertical displacement of the tire.

[0036] The xy coordinates of the 8 tire centers in the vehicle body coordinate system in the XY plane are:

[0037]

[0038] Where: l represents the longitudinal distance between the four axles of the vehicle and the center of mass, w represents the wheelbase of the four axles, l and w are the distances in the vehicle coordinate system, subscript f represents the first axle, s represents the second axle, t represents the third axle, and r represents the fourth axle.

[0039] Unsprung mass, i.e. the relative vertical position z of the tire in the vehicle coordinate system w It can be represented by its absolute coordinates, the vehicle body roll angle, pitch angle and the xy coordinates in the vehicle body coordinate system, as shown below:

[0040]

[0041] In the formula, z u Indicates the vertical position of the tire's center of mass.

[0042] The vehicle 22-degree-of-freedom dynamics equation is derived based on Lagrangian dynamics, as shown below:

[0043]

[0044] Where: T represents the kinetic energy of the system, Q b represents the generalized force on the sprung mass, Q u represents the generalized force on the unsprung mass, q b represents the generalized coordinates of the vehicle body, i.e., the sprung mass, q u Represents the generalized coordinates of the eight tires, i.e. the unsprung mass, including the tire angle and vertical position.

[0045] When calculating the kinetic energy of the sprung mass, the kinetic energy of the tire's lateral and longitudinal motion is taken into account, as shown in the following formula:

[0046]

[0047] Where: V b 、V ui are the components of the generalized velocity of the vehicle body and the velocity of the unsprung mass in the vehicle body coordinate system, M b 、M ui are the mass matrices of the vehicle body and the unsprung mass when calculating the lateral and longitudinal kinetic energy. Parameter V ui This is the aforementioned parameter V u,i Parameter M ui This is the aforementioned parameter M u,i .

[0048] There is a coordinate transformation between the vehicle coordinate system and the earth coordinate system, and its transformation matrix is:

[0049]

[0050] From the above formula, we can get the vehicle speed V b In the vehicle coordinate system:

[0051]

[0052] The lateral and longitudinal velocities of the eight wheels in the vehicle coordinate system can be determined from their relative positions h in the vehicle coordinate system. b,u,i The following vehicle speed is expressed as follows:

[0053]

[0054] Among them, the superscript i represents the sequential number of the 8 tires, which are the right wheel of axle one, the left wheel of axle one, the right wheel of axle two, the left wheel of axle two, the right wheel of axle three, the left wheel of axle three, the right wheel of axle four, and the left wheel of axle four.

[0055] Through coordinate transformation and matrix calculation, the kinetic energy of the sprung mass can be expressed as:

[0056]

[0057] Where M gb =[h A,b ] T [M b ][h A,b ]+Σ[h A,b ] T [h b,u,i ] T [M ui ][h b,u,i ][h A,b ].

[0058] The kinetic energy of the unsprung mass is composed of the rotational and vertical motion of the tire, as shown below:

[0059]

[0060] Where: V uz ,ω u are the vertical speed and angular velocity of the 8 unsprung tires, M u , J u They are the mass matrix and moment of inertia matrix for calculating vertical kinetic energy of unsprung mass. According to matrix calculation, the formula can be converted into the form of generalized coordinates and generalized mass matrix:

[0061]

[0062]

[0063] In the formula, m u,fr ,…,m u,rl represents the mass of eight tires, j u,fr ,…j u,rl Represents the rotational inertia of the 8 tires.

[0064] The vertical forces acting on the sprung mass (vehicle body) include the weight of the vehicle body, the suspension force, and the vertical air resistance. The longitudinal forces include the lateral force generated between the tire and the ground, the longitudinal force, and the air resistance in the global coordinate system. The lateral force includes the components of the lateral and longitudinal forces of the ground on the tire. The three-axis torques include the torques of the above forces on the center of mass, the tire self-aligning torque, the air resistance torque, and the reaction torque on the drive tire.

[0065] The vertical force on the tire is composed of the tire's own weight m u,i g and suspension force F bs,i constitute:

[0066] F z,i =F bs,i +m u,i g (13);

[0067] The longitudinal slip rate of the tire contact point with the ground is calculated as follows:

[0068]

[0069] Where: κ i is the tire slip rate, ω i is the tire rotation angular velocity, R e,i is the effective radius of the wheel, v wx,i 、v wy,i are the longitudinal and lateral speeds of the wheel center respectively, and the calculation method is shown in the following formula:

[0070]

[0071] Where: v ux,i , v uy,i The lateral and longitudinal speeds of the tire in the vehicle body coordinate system, δ i For tire corners.

[0072] Apply the magic formula to calculate the tire longitudinal force F x,i , lateral force F y,i , return torque M z,i , rolling resistance moment M y,i , yaw resistance moment M x,i .

[0073] The suspension force is composed of spring force and damping force. When the stiffness and damping ratio are constant values, the suspension force can be expressed as:

[0074]

[0075] In the formula, F bss,i is the suspension force, k bs,i is the suspension stiffness coefficient, Δl s.i is the suspension deformation, c d,i is the suspension damping coefficient, is the suspension deformation speed.

[0076] To calculate the air resistance, first calculate the air resistance slip angle. Assuming the air is still and the wind speed is 0, the air resistance slip angle is the vehicle's center of mass slip angle. The calculation method is as follows:

[0077]

[0078] Among them, α air is the air resistance slip angle, is the vehicle longitudinal velocity, is the lateral velocity of the vehicle body, and ψ is the yaw angle of the vehicle body.

[0079] Then, the air resistance coefficient is calculated by linear interpolation and cubic spline interpolation according to the air resistance side slip angle:

[0080]

[0081] C Fx ,C Fy , C Fz , C Mx , C My , C Mz Respectively represent the longitudinal, lateral, vertical air resistance and the air resistance moment around the x-axis, y-axis, and z-axis air resistance coefficients, interp represents the interpolation function, α air,refrepresents the interpolated air resistance side slip angle reference value, C Fx,ref ,C Fy,ref , C Fz,ref , C Mx,ref , C My,ref , C Mz,ref They represent the interpolated longitudinal, lateral, and vertical air resistance, as well as the reference values ​​of the air resistance coefficient around the x-axis, y-axis, and z-axis. method: linear represents the application of the linear interpolation method, and method: spline represents the application of the cubic spline interpolation method.

[0082] Calculate the air resistance factor Q:

[0083]

[0084] Where D is the air density.

[0085] In summary, put the above results into the calculation of air resistance:

[0086]

[0087] Where: A air is the frontal area of ​​the vehicle, L air Calculate the length for the vehicle's air resistance.

[0088] Based on the calculation of the tire force and suspension model above, the generalized force matrix of the sprung mass is as follows:

[0089]

[0090] Among them, δ i is the tire angle input, T d,i is the tire torque input.

[0091] The forces acting on the unsprung mass include suspension force, ground vertical force, and tire weight, and the torques acting on the unsprung mass include driving torque and tire rolling resistance torque. The generalized force matrix is ​​expressed as:

[0092]

[0093] Among them, δ i is the tire angle input, T d,i is the tire torque input, and g is the acceleration due to gravity.

[0094] Based on force analysis and the principle of virtual work, the generalized force can be derived as shown in the following equation:

[0095]

[0096] According to Lagrangian mechanics and D'Alembert's principle, the generalized equation of motion for the sprung mass is as follows:

[0097]

[0098] The generalized equation of motion for the unsprung mass is given by:

[0099]

[0100] The second-order differentials of all state variables can be obtained by equations (20) and (21), the first-order differentials of the state variables can be obtained by integration, and the values ​​of all state variables can be obtained by calculation.

[0101] After establishing the dynamic model of the wheeled armored vehicle, we further construct the problem of collaborative optimal design and optimal control. According to the needs, we consider multiple objectives such as mobility and vertical impact, establish the cost function, establish dynamic constraints (residual constraints) based on the dynamic model, extract the upper and lower bounds of the vehicle state according to the vehicle driving conditions, establish boundary constraints, and design path constraints and boundary constraints according to the needs. The path constraint is to limit the dynamic state of the vehicle such as displacement and speed when driving on the path to ensure that the vehicle follows a specific trajectory. The boundary constraint is to limit the state parameters (such as position, speed, acceleration, etc.) of the vehicle at the starting and ending times to ensure that the initial and final states of the vehicle meet specific requirements. Combine the design parameters and control parameters to establish the problem of collaborative optimal design and optimal control. Complete the objective function, upper and lower bounds of the constraints, and guess the initial value as needed. It is also necessary to set some optimization-related parameters, such as the transformation method m trans (The transformation method is a method for calculating residual constraints, Hermite-Simpson see formula (23)), differential method m diff , allocation point number N n , zoom settings scal, etc.

[0102] According to the established vehicle dynamics model, the input of the problem is provided, including the initial guess value x0 of the state quantity, the initial guess value u0 of the control quantity, and the terminal time t f0 And design variables p0. Design variables include suspension stiffness, tire radius, etc.

[0103]

[0104] Among them, N x,usr is the number of discrete points of the state variable provided by the user. Each discrete point can be considered as a moment, and the value of the state parameter of each discrete point needs to be determined; t is the time from 0 to t f0 Discretized into N x,usr A time series of points, N u,usr is the number of discrete points of the control variable provided by the user, n x is the number of state variables of the dynamic equation; nu is the number of control variables. If the terminal time is also a parameter to be optimized, then n tf is 1, otherwise it is 0, n p is the number of design parameters, Represents a two-dimensional matrix with i rows and j columns.

[0105] The above state variables, control variables, terminal time, upper and lower bounds of design variables and inequality constraints g are given by the following formula, n g is the number of path constraints:

[0106]

[0107] Among them, x min is the lower bound of the state parameter; x max is the upper bound of the state parameter; u min is the lower bound of the control parameter; u max is the upper bound of the control parameter; p min is the lower bound of the design parameter; p max is the upper bound of the design parameters; t fmin is the lower bound of the termination time; g min is the lower bound of the path constraint function; g max is the upper bound of the path constraint function.

[0108] In addition to the above initial values ​​and upper and lower bounds, the problem input also includes the Lagrangian term of the objective function. Meyer Item The first-order dynamic constraint function f, the path constraint function g, and the boundary constraint function b, the input and output dimensions of the above functions are shown in the following formula:

[0109]

[0110] Based on the dynamic model, a collaborative optimization model of wheeled vehicles is established. The optimization variables of the collaborative optimization model include state parameters, design parameters and control parameters. The collaborative optimization model includes cost functions and constraints. The constraints include dynamic constraints, upper and lower bounds of variable values, path constraints and boundary constraints. The dynamic constraints are constraints established based on the dynamic model. The collaborative optimization model is:

[0111]

[0112] Among them, J is the cost function; for Meyer Item; is the Lagrangian term; t0 is the starting time; t f is the end time; is the dynamic constraint; is the first-order derivative of the state parameter at time t; f[.] is the dynamic constraint function; x(t) is the state parameter at time t; u(t) is the control parameter at time t; p is the design parameter; x min ≤x(t)≤x max 、u min ≤u(t)≤u max 、p min ≤p≤p max The upper and lower bounds of the variable value are constrained; x min is the lower bound of the state parameter; x max is the upper bound of the state parameter; u min is the lower bound of the control parameter; u max is the upper bound of the control parameter; p min is the lower bound of the design parameter; p max is the upper bound of the design parameter; g min ≤g[x(t),u(t),t,p]≤g max is the path constraint; g min is the lower bound of the path constraint function; g[.] is the path constraint function; g max is the upper bound of the path constraint function; b min ≤b[x(t0),t0,x(t f ),t f ,p]≤b max is the boundary constraint; b min is the lower bound of the boundary constraint function; b[.] is the boundary constraint function; x(t0) is the state parameter at the starting time t0; x(t f ) is the termination time t f State parameters of b max is the upper bound of the boundary constraint function.

[0113] The state parameters are:

[0114]

[0115] The 8 tires are distinguished by subscripts. The subscript ufr represents the right tire of the first axle, ufl represents the left tire of the first axle, usr represents the right tire of the second axle, usl represents the left tire of the second axle, utr represents the right tire of the third axle, utl represents the left tire of the third axle, urr represents the right tire of the fourth axle, and url represents the left tire of the fourth axle.

[0116] Design parameters are defined based on requirements;

[0117] The control parameters are the rotation angle and torque of the eight tires;

[0118] After the collaborative optimal design and optimal control problems are constructed, the problem is further solved, that is, equation (29) is solved to obtain the optimal control parameters and optimal design parameters. The problem is first transformed into an NLP problem, and an efficient solution algorithm is obtained through sparse matrix operations, improved direct point matching method, automatic normalization (corresponding to the scaling part of (3) below), and high-performance cluster computing. Among them, sparse matrix operations and high-performance cluster computing are methods used in programming to improve the efficiency of problem solving, complete the efficient solution of the problem, and simultaneously solve the optimal design parameters and the optimal control parameters of the vehicle. The optimal design parameters are used in vehicle development, and the optimal control parameters are used to guide the development of vehicle control algorithms, and finally the collaborative optimization control of 8×8 wheeled vehicles is realized.

[0119] In the problem solving part, the problem needs to be converted into a large-scale NLP problem, and then the underlying solver, such as IPOPT, is called to solve the problem. In the problem conversion part, the format of variables and functions in the problem input is converted into the format required by the NLP solver. In this part, the local collocation method, differential method, and automatic scaling method are developed. The last part applies the NLP solver to solve the problem.

[0120] (1) Local point allocation method

[0121] (1.1) NLP solver variables

[0122] In the local collocation method, the continuous state and control quantities in the entire time interval can be discretized into N n nodes, the new variable is given by the following formula:

[0123]

[0124] Then, the discrete state variables, control variables and design variables are reconstructed into NLP column vectors as shown in the following formula.

[0125]

[0126] For the Hermite-Simpson method (local collocation method), when the control variable in each sampling interval is also selected as the parameter to be optimized, the reconstructed NLP variable is:

[0127]

[0128] When the state quantity in each sampling interval is also selected as the parameter to be optimized, the reconstructed NLP variable is as follows:

[0129]

[0130] In the formula, subscripts are used to represent different moments, and the underlined variables are obtained by calculating the corresponding two variables before and after them, that is, for and Calculated control parameters. for and Calculated state parameters.

[0131] The upper and lower bounds of the variables in formula (27) must also be reconstructed based on the reconstructed NLP variables.

[0132] (1.2) Constraints (this step is to process the constraints)

[0133] 1) Residual constraints of the Hermite-Simpson method

[0134] Based on the third-order Hermite interpolation, at x k With x k+1 The state variables between and its derivatives It can be deduced as follows:

[0135]

[0136] Where h is the time interval between discrete points, that is, the time interval between two adjacent moments (i.e., two adjacent nodes).

[0137] Based on the Hermite-Simpson method, The residual constraint at It can be expressed as:

[0138]

[0139] Apply linear interpolation to calculate the control parameter u at time k k and the control parameter u at time k+1 k+1 The control variables between as follows:

[0140]

[0141] In order to reduce repeated calculations in numerical calculations, matrix calculation method is applied, so the state variable and control variables It can be expressed as:

[0142]

[0143] is the dynamic equation of the system:

[0144]

[0145] Transformation matrix T s1 and T s2 for:

[0146]

[0147] The residual constraint ζ can be expressed by applying matrix calculation as:

[0148]

[0149] 2) Residual constraints of the Trapezoidal method

[0150] Residual Constraint ζ for Trapezoidal Method k It can be expressed as:

[0151]

[0152] The application matrix is ​​calculated as:

[0153]

[0154] When calculating the residual constraint, one of the residual constraints of the Hermite-Simpson method in 1) and the residual constraints of the Trapezoidal method in 2) can be selected.

[0155] 3) Path constraints and boundary constraints

[0156] The path constraint is a function of state parameters, control parameters, design parameters and time. The boundary constraints are the starting state (i.e., the state parameters at the starting time) x0 and the ending state (i.e., the state parameters at the ending time) x f The path constraints are as follows:

[0157]

[0158] Finally, after all constraints are calculated, the NLP constraints are expressed as:

[0159]

[0160] All residual constraints are bounded by 0.

[0161] (1.3) Jacobians matrix - Jacobian matrix (this step is to find the derivative of the above constraints with respect to the variables) 1) Jacobians matrix of residual constraints

[0162] Without considering the state parameters and control parameters at the midpoint of each discrete time interval, the residual constraint is x a ,ua ,x b ,u b ,t f ,p function, such as Figure 4 As shown, x a ,u a ,x b ,u b The specific values ​​are as follows:

[0163] To avoid repeated calculations, perform matrix calculations and first calculate the derivatives of ζ with respect to x, u, and p:

[0164]

[0165] Therefore, the Jacobians matrix can be simplified to:

[0166]

[0167] In the formula, is a matrix whose i-th column is 1 and the rest of the elements are all 0, T a ,T b Given by:

[0168]

[0169] in,

[0170] 2) Jacobians matrix of path constraints and boundary constraints

[0171] First, calculate the derivative of the path constraint g with respect to x, u, p, t as follows:

[0172]

[0173] Then calculate the boundary constraints b for x0,x f ,p,t’s derivative is as follows:

[0174]

[0175] (1.4) Objective function (cost function)

[0176] The input objective function consists of Mayer terms and Lagrange terms, which can be expressed as:

[0177]

[0178] (1.5) Gradient of the objective function

[0179] The objective function is [x,u,t f,p], so its gradient to state parameters and control parameters is as follows:

[0180]

[0181] Then the gradient of the objective function with respect to the terminal time and the design parameters is calculated, and finally the NLP gradient is obtained as follows:

[0182]

[0183] (2) Difference method (when calculating the Jacobians matrix, it is necessary to differentiate the variables, and this difference method is used to differentiate)

[0184] Most NLP solvers require the first-order derivatives of the constraints c(y), and accurate calculation of these derivatives helps to achieve faster convergence. Among them, the difference methods implemented include forward finite difference, central difference, complex step difference and analytical difference. You can use any of the above difference methods. The forward finite difference approximation method is recorded as:

[0185]

[0186] Where x is the state quantity used to calculate the derivative; h i For the time interval.

[0187] The central finite difference approximation method is:

[0188]

[0189] The complex step-size differentiation approximation is based on Taylor expansion:

[0190]

[0191] Ignore h 2 and higher-order terms, the approximation error is The first-order derivative of can be approximated as:

[0192] f'(x)=Im[f(x+ih)] / h(57);

[0193] In the formula, i is the imaginary number symbol.

[0194] Another method is the analytical method, which is to calculate through the symbolic toolbox to obtain the analytical form of the first-order derivative. This method can obtain accurate derivative values, but it may be difficult to obtain or impossible to obtain the analytical form for some complex problems.

[0195] (3) Scaling

[0196] (3.1) Variable scaling

[0197] Apply linear scaling methods to scale state variables, control variables, terminal times, and design variables:

[0198]

[0199]

[0200] Apply a gradient-based average norm method to scale the constraints for faster convergence:

[0201] The residual constraint is scaled as follows:

[0202]

[0203] Where N scal is the number of loops used to solve the scaling factor; ζ j,i is the i-th residual constraint in the j-th cycle.

[0204] The path constraint is scaled as follows:

[0205]

[0206] In formula (61), g j,i is the i-th path constraint in the j-th cycle.

[0207] The bounds constraints are scaled as follows:

[0208]

[0209] In formula (62), b j,i is the i-th boundary constraint in the j-th cycle.

[0210] The cost function is not scaled.

[0211] (3.3) Scaling of Jacobians

[0212] The relationship between scaling variables and scaling constraints is considered in the scaling of the Jacobians matrix. The scaling of the Jacobians matrix of the residual constraint is defined as follows:

[0213]

[0214] In the formula, J ζ,i is the Jacobian matrix of the residual constraint; ξ i is the row vector composed of the scaling factors in equation (59); is the scaling factor for the residual constraint.

[0215] Similarly, the scaling of the Jacobians matrix for path constraints is:

[0216]

[0217] In the formula, J g,i is the Jacobian matrix of the path constraint; is the scaling factor for the path constraint.

[0218]

[0219] In the formula, J b,i is the Jacobian matrix of the boundary constraints; is the scaling factor for the boundary constraints.

[0220] The gradient of the cost function is not scaled.

[0221] After the above problem transformation, the transformed problem is input into the underlying solver for solving, which can solve the optimal control parameters and the optimal design parameters at the same time.

[0222] Then in step S3, the collaborative optimization model is optimized and solved to obtain the optimal design parameters and optimal control parameters of the wheeled armored vehicle, which specifically include:

[0223] (1) Linear interpolation is performed on the state parameters and control parameters respectively to obtain the state parameters and control parameters at each interpolation time, so as to reconstruct the optimization variables and obtain the reconstructed optimization variables. The interpolation time includes the start time, the end time and several intermediate times between the start time and the end time. The reconstructed optimization variables include the design parameters and the state parameters and control parameters at each interpolation time.

[0224] The state parameters and the control parameters are linearly interpolated respectively, specifically including: using the Hermite-Simpson method to linearly interpolate the state parameters and the control parameters respectively.

[0225] The optimized variables after reconstruction are shown in formula (34).

[0226] (2) Calculate the gradient of the cost function with respect to the reconstructed optimized variable.

[0227] Calculating the gradient of the cost function with respect to the reconstructed optimization variable, specifically comprising: calculating the gradient of the cost function with respect to the reconstructed optimization variable using a complex step-size differentiation method.

[0228] The gradient is shown in formula (53).

[0229] (3) Based on the reconstructed optimization variables, the dynamic constraints, upper and lower bound constraints on variable values, path constraints, and boundary constraints are reconstructed respectively to obtain residual constraints, upper and lower bound constraints on reconstructed variable values, reconstructed path constraints, and reconstructed boundary constraints.

[0230] The method for determining the residual constraint includes: based on the reconstructed optimized variables, the dynamic constraint is reconstructed using the Hermite-Simpson method to obtain the residual constraint.

[0231] The residual constraint, reconstructed path constraint and reconstructed boundary constraint are shown in formula (45), and the upper and lower bound constraints of the reconstructed variable values ​​are shown in formula (29). It only requires changing the dimension of the optimized variable.

[0232] (4) Calculate the first Jacobian matrix of the residual constraint, the second Jacobian matrix of the reconstructed path constraint, and the third Jacobian matrix of the reconstructed boundary constraint respectively.

[0233] The first Jacobian matrix of the residual constraint, the second Jacobian matrix of the reconstructed path constraint and the third Jacobian matrix of the reconstructed boundary constraint are calculated respectively, specifically including: using the complex step differentiation method to calculate the first Jacobian matrix of the residual constraint, the second Jacobian matrix of the reconstructed path constraint and the third Jacobian matrix of the reconstructed boundary constraint respectively.

[0234] The first Jacobian matrix is ​​shown in formula (48), the second Jacobian matrix is ​​shown in formula (50), and the third Jacobian matrix is ​​shown in formula (51).

[0235] (5) The reconstructed optimization variables, cost function, gradient, residual constraints, upper and lower bounds of the reconstructed variables, reconstructed path constraints, reconstructed boundary constraints, the first Jacobian matrix, the second Jacobian matrix and the third Jacobian matrix are used as inputs, and the nonlinear programming solver is used to solve the problem, and the optimal design parameters and optimal control parameters of the wheeled armored vehicle are obtained.

[0236] The optimal design parameters and optimal control parameters of the wheeled armored vehicle are obtained by using the nonlinear programming solver with the reconstructed optimization variables, cost function, gradient, residual constraint, upper and lower bound constraints of the reconstructed variable value, reconstructed path constraint, reconstructed boundary constraint, first Jacobian matrix, second Jacobian matrix and third Jacobian matrix as input, and solving by nonlinear programming solver, including:

[0237] 1) Using the reconstructed optimization variables, cost function, gradient, residual constraints, upper and lower bound constraints of the reconstructed variable values, reconstructed path constraints, reconstructed boundary constraints, the first Jacobian matrix, the second Jacobian matrix and the third Jacobian matrix as input, a nonlinear programming solver is used to solve the problem and obtain the solution of the current iteration, which includes the state parameters, design parameters and control parameters of the current iteration.

[0238] 2) Determine whether the current number of iterations has reached the preset number of iterations.

[0239] 3) If not, return to the step of "using the reconstructed optimization variables, cost function, gradient, residual constraints, upper and lower bound constraints of the reconstructed variable values, reconstructed path constraints, reconstructed boundary constraints, the first Jacobian matrix, the second Jacobian matrix and the third Jacobian matrix as input, solving using the nonlinear programming solver to obtain the solution result of the current iteration".

[0240] 4) If so, based on the solution results of each iteration before the current iteration, calculate the first scaling factor of the residual constraint, the second scaling factor of the reconstructed path constraint, the third scaling factor of the reconstructed boundary constraint, the fourth scaling factor of the first Jacobian matrix, the fifth scaling factor of the second Jacobian matrix, and the sixth scaling factor of the third Jacobian matrix.

[0241] 5) Scale the reconstructed optimization variables of the current iteration using a linear scaling method to obtain scaled optimization variables, and scale the residual constraints, reconstructed path constraints, reconstructed boundary constraints, first Jacobian matrix, second Jacobian matrix and third Jacobian matrix of the current iteration based on the first scaling factor, the second scaling factor, the third scaling factor, the fourth scaling factor, the fifth scaling factor and the sixth scaling factor, respectively, to obtain the scaled residual constraints, the scaled path constraints, the scaled boundary constraints, the scaled first Jacobian matrix, the scaled second Jacobian matrix and the scaled third Jacobian matrix.

[0242] The scaled optimization variables are shown in formula (58), the scaled residual constraints are shown in formula (60), the scaled path constraints are shown in formula (61), the scaled boundary constraints are shown in formula (62), the scaled first Jacobian matrix is ​​shown in formula (63), the scaled second Jacobian matrix is ​​shown in formula (64), and the scaled third Jacobian matrix is ​​shown in formula (65).

[0243] 6) Using the scaled optimization variables, cost function, gradient, scaled residual constraints, upper and lower bound constraints of reconstructed variable values, scaled path constraints, scaled boundary constraints, scaled first Jacobian matrix, scaled second Jacobian matrix and scaled third Jacobian matrix as input, a nonlinear programming solver is used to solve the problem and obtain the solution result for the next iteration.

[0244] 7) Determine whether the iteration is completed.

[0245] If the solution minimizes the cost function value, the iteration ends.

[0246] 8) If the iteration is completed, the optimal design parameters and optimal control parameters of the wheeled armored vehicle are determined based on the solution results of the next iteration.

[0247] 9) If the iteration is not finished, the number of iterations is increased by 1, and the process returns to the step of "scaling the reconstructed optimized variables of the current iteration using a linear scaling method to obtain scaled optimized variables".

[0248] In this embodiment, for the rapid development and design of 8×8 wheeled armored vehicles, the present invention has developed for the first time a multi-objective optimization framework that can quickly solve large-scale complex mixed integer optimal design and optimal control problems. Compared with similar optimization frameworks in the world, it has a high success rate and fast speed, and has greatly improved the relevant performance of wheeled armored vehicles in terms of vehicle optimization design, component optimization design and control, as well as from the design and control perspectives. The overall process of the embodiment is as follows: Figure 1 As shown in the figure, firstly, a dynamic model of a wheeled vehicle is built based on the Lagrangian dynamics theory. The model consists of two parts: sprung mass and unsprung mass. Then, the above model is converted into residual constraints, boundary constraints are established according to driving conditions and working boundaries of each subsystem, and multi-objective functions are established according to task requirements. The design parameters and control parameters are used as optimization variables at the same time to complete the construction of the collaborative optimal design and optimal control problems. Finally, the problem is solved based on the improved collocation method, normalization method and scaling method. The optimal design parameters obtained are applied to vehicle development, and the optimal control parameters are used to guide the development of vehicle control algorithms.

[0249] The advantages of this embodiment are:

[0250] (1) A collaborative optimal design and optimal control framework for wheeled armored vehicles is constructed, which can optimize the design parameters and control parameters of wheeled armored vehicles based on the dynamic model of the wheeled armored vehicles. The optimal design parameters can be used for vehicle design, and the optimal control parameters can be used as a reference for the subsequent online control strategy formulation. The framework can apply gradient information to expand the optimization space and has good optimization effects, which can improve the development efficiency of wheeled armored vehicles. This is mainly related to the construction of optimal design and optimal control problems, the rationality of vehicle dynamics modeling, and the efficient solution of the problem.

[0251] (2) For large-scale multi-degree-of-freedom optimal design and optimal control problems, the solution stability and efficiency are greatly improved. This is mainly due to the vehicle vector dynamics modeling method. In the solution part of the optimal problem, the direct point matching method is improved, and the normalization method and scaling method are applied to achieve stable and fast solution to large-scale collaborative optimal problems.

[0252] Example 2

[0253] This embodiment provides a computer device, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the method for collaboratively optimizing design parameters and control parameters of a wheeled vehicle in the above-mentioned embodiment 1 is implemented.

[0254] Example 3

[0255] In this embodiment, a computer-readable storage medium is provided, which stores a computer program. When the computer program is executed by a processor, the method for collaboratively optimizing the design parameters and control parameters of a wheeled vehicle in the above-mentioned embodiment 1 is implemented.

[0256] Example 4

[0257] In this embodiment, a computer program product is provided, including a computer program, which, when executed by a processor, implements the method for collaboratively optimizing design parameters and control parameters of a wheeled vehicle in the above-mentioned embodiment 1.

[0258] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0259] This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method and core ideas of this application. At the same time, for those skilled in the art, according to the ideas of this application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.

Claims

1. A method for collaborative optimization of design parameters and control parameters of a wheeled vehicle, characterized in that: include: Establish a dynamic model of wheeled vehicles; Establishing a collaborative optimization model of wheeled vehicles based on the dynamic model; The optimization variables of the collaborative optimization model include state parameters, design parameters and control parameters; the collaborative optimization model includes a cost function and constraints, the constraints include dynamic constraints, upper and lower bounds of variable values, path constraints and boundary constraints, and the dynamic constraints are constraints established based on the dynamic model; The collaborative optimization model is optimized and solved to obtain the optimal design parameters and optimal control parameters of the wheeled vehicle.

2. The method for collaborative optimization of design parameters and control parameters of a wheeled vehicle according to claim 1, characterized in that: The dynamic model of the wheeled vehicle includes the generalized motion equations of the sprung mass of the wheeled vehicle and the generalized motion equations of the unsprung mass of the wheeled vehicle; The generalized equation of motion for the sprung mass of a wheeled vehicle is: Among them, M gb = [h A,b T [M b [h A,b + ∑[h A,b T [h b,u,i T [M ui [h b,u,i [h A,b ;​​​ The generalized equation of motion for the unsprung mass of a wheeled vehicle is: Among them, [M gu ] = diag{m u,fr ,…,m u,rl ,j u,fr ,…j u,rl }; In the formula, Q b It represents the generalized force on the sprung mass of the wheeled vehicle; Q u represents the generalized force acting on the unsprung mass of a wheeled vehicle; q b represents the generalized coordinates of the vehicle body; q u represents the generalized coordinates of the tire; and Indicates q b The first and second derivatives of ; and Indicates q u The first and second derivatives of ; Indicates M gb The first derivative of Indicates M gu The first derivative of b 、M ui are the mass matrices of the vehicle body and the unsprung mass when calculating the lateral and longitudinal kinetic energy; h b,u,i It represents the relative position of the 8 tires of the 8×8 wheeled vehicle in the body coordinate system. The subscript i represents the order of the 8 tires, which are the right wheel of the first axle, the left wheel of the first axle, the right wheel of the second axle, the left wheel of the second axle, the right wheel of the third axle, the left wheel of the third axle, the right wheel of the fourth axle, and the left wheel of the fourth axle; ψ is the body yaw angle; m u,fr ,…,m u,rl represents the mass of the eight tires of an 8×8 wheeled vehicle; j u,fr ,…j u,rl Represents the moment of inertia of the eight tires of an 8×8 wheeled vehicle.

3. The method for collaborative optimization of design parameters and control parameters of a wheeled vehicle according to claim 1, characterized in that: The collaborative optimization model for wheeled vehicles includes: Among them, J is the cost function; is the Mayer term; t0 is the starting time; t f is the end time; is the Lagrangian term; is the dynamic constraint; is the first-order derivative of the state parameter at time t; f[.] is the dynamic constraint function; x(t) is the state parameter at time t; u(t) is the control parameter at time t; p is the design parameter; x min ≤x(t)≤x max 、u min ≤u(t)≤u max 、p min ≤p≤p max The upper and lower bounds of the variable value are constrained; x min is the lower bound of the state parameter; x max is the upper bound of the state parameter; u min is the lower bound of the control parameter; u max is the upper bound of the control parameter; p min is the lower bound of the design parameter; p max is the upper bound of the design parameter; g min ≤g[x(t),u(t),t,p]≤g max is the path constraint; g min is the lower bound of the path constraint function; g[.] is the path constraint function; g max is the upper bound of the path constraint function; b min ≤b[x(t0),t0,x(t f ),t f ,p]≤b max is the boundary constraint; b min is the lower bound of the boundary constraint function; b[.] is the boundary constraint function; x(t0) is the state parameter at the starting time t0; x(t f ) is the termination time t f State parameters of b max is the upper bound of the boundary constraint function.

4. The method for collaborative optimization of design parameters and control parameters of a wheeled vehicle according to claim 1, characterized in that: The collaborative optimization model is optimized and solved to obtain the optimal design parameters and optimal control parameters of the wheeled vehicle, specifically including: Linearly interpolating the state parameters and the control parameters respectively to obtain the state parameters and the control parameters at each interpolation time, so as to reconstruct the optimization variables to obtain the reconstructed optimization variables; the interpolation time includes the starting time, the ending time and a plurality of intermediate times between the starting time and the ending time; the reconstructed optimization variables include the design parameters and the state parameters and the control parameters at each interpolation time; Calculating the gradient of the cost function with respect to the reconstructed optimization variable; Based on the reconstructed optimization variables, the dynamic constraints, the upper and lower bound constraints of the variable values, the path constraints and the boundary constraints are reconstructed respectively to obtain residual constraints, upper and lower bound constraints of the reconstructed variable values, reconstructed path constraints and reconstructed boundary constraints; respectively calculating a first Jacobian matrix of the residual constraint, a second Jacobian matrix of the reconstructed path constraint, and a third Jacobian matrix of the reconstructed boundary constraint; The optimal design parameters and optimal control parameters of the wheeled vehicle are obtained by using a nonlinear programming solver and taking the reconstructed optimization variables, the cost function, the gradient, the residual constraint, the upper and lower bound constraints of the reconstructed variable values, the reconstructed path constraint, the reconstructed boundary constraint, the first Jacobian matrix, the second Jacobian matrix and the third Jacobian matrix as inputs.

5. The method for collaborative optimization of design parameters and control parameters of a wheeled vehicle according to claim 4, characterized in that: Linear interpolation is performed on the state parameters and the control parameters respectively, specifically including: linear interpolation is performed on the state parameters and the control parameters respectively using the Hermite-Simpson method; The method for determining the residual constraint includes: based on the reconstructed optimization variables, reconstructing the dynamic constraint using the Hermite-Simpson method to obtain the residual constraint.

6. The method for collaborative optimization of design parameters and control parameters of a wheeled vehicle according to claim 4, characterized in that: Calculating the gradient of the cost function with respect to the reconstructed optimization variable, specifically comprising: calculating the gradient of the cost function with respect to the reconstructed optimization variable using a complex step-size differentiation method; The first Jacobian matrix of the residual constraint, the second Jacobian matrix of the reconstructed path constraint and the third Jacobian matrix of the reconstructed boundary constraint are calculated respectively, specifically including: using a complex step differentiation method to calculate the first Jacobian matrix of the residual constraint, the second Jacobian matrix of the reconstructed path constraint and the third Jacobian matrix of the reconstructed boundary constraint respectively.

7. The method for collaborative optimization of design parameters and control parameters of a wheeled vehicle according to claim 4, characterized in that: Taking the reconstructed optimization variables, the cost function, the gradient, the residual constraint, the upper and lower bound constraints of the reconstructed variable values, the reconstructed path constraint, the reconstructed boundary constraint, the first Jacobian matrix, the second Jacobian matrix and the third Jacobian matrix as input, a nonlinear programming solver is used to solve the problem, and the optimal design parameters and optimal control parameters of the wheeled vehicle are obtained, which specifically includes: Taking the reconstructed optimization variables, the cost function, the gradient, the residual constraint, the upper and lower bound constraints of the reconstructed variable values, the reconstructed path constraint, the reconstructed boundary constraint, the first Jacobian matrix, the second Jacobian matrix and the third Jacobian matrix as input, using a nonlinear programming solver to solve and obtain a solution result of the current iteration; Determine whether the current number of iterations reaches the preset number of iterations; If not, return to the step of "taking the reconstructed optimization variables, the cost function, the gradient, the residual constraint, the upper and lower bound constraints of the reconstructed variable values, the reconstructed path constraint, the reconstructed boundary constraint, the first Jacobian matrix, the second Jacobian matrix and the third Jacobian matrix as input, solving using a nonlinear programming solver to obtain a solution result of the current iteration"; If so, then based on the solution results of each iteration before the current iteration, calculate the first scaling factor of the residual constraint, the second scaling factor of the reconstructed path constraint, the third scaling factor of the reconstructed boundary constraint, the fourth scaling factor of the first Jacobian matrix, the fifth scaling factor of the second Jacobian matrix, and the sixth scaling factor of the third Jacobian matrix; Scale the reconstructed optimization variables of the current iteration by a linear scaling method to obtain scaled optimization variables, and scale the residual constraint, the reconstructed path constraint, the reconstructed boundary constraint, the first Jacobian matrix, the second Jacobian matrix, and the third Jacobian matrix of the current iteration based on the first scaling factor, the second scaling factor, the third scaling factor, the fourth scaling factor, the fifth scaling factor, and the sixth scaling factor, respectively, to obtain the scaled residual constraint, the scaled path constraint, the scaled boundary constraint, the scaled first Jacobian matrix, the scaled second Jacobian matrix, and the scaled third Jacobian matrix; Taking the scaled optimization variable, the cost function, the gradient, the scaled residual constraint, the upper and lower bound constraints of the reconstructed variable value, the scaled path constraint, the scaled boundary constraint, the scaled first Jacobian matrix, the scaled second Jacobian matrix and the scaled third Jacobian matrix as input, using a nonlinear programming solver to solve and obtain a solution result for the next iteration; Determine whether the iteration is completed; If the iteration is finished, the optimal design parameters and optimal control parameters of the wheeled vehicle are determined based on the solution results of the next iteration; If the iteration is not finished, the number of iterations is increased by 1, and the process returns to the step of "scaling the reconstructed optimized variables of the current iteration by using the linear scaling method to obtain the scaled optimized variables".

8. A computer device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method for collaboratively optimizing design parameters and control parameters of a wheeled vehicle according to any one of claims 1 to 7.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for collaboratively optimizing design parameters and control parameters of a wheeled vehicle according to any one of claims 1 to 7 is implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method for collaboratively optimizing design parameters and control parameters of a wheeled vehicle according to any one of claims 1 to 7 is implemented.

Images