Vehicle control method, device, apparatus, and storage medium

Abstract

The embodiment of the application relates to the technical field of vehicle-related technologies, and discloses a vehicle control method, device, equipment and storage medium, which comprises the following steps: acquiring real-time state information of a vehicle at a current moment in each control cycle; determining an optimization proposition equation corresponding to a nonlinear optimal control problem with a maximum vehicle speed as a target at the current moment based on the real-time state information and a pre-established vehicle dynamics model; taking each wheel torque and each wheel rotation angle as control variables to be solved for the optimization proposition equation; processing the optimization proposition equation to convert the nonlinear optimal control problem into a quadratic programming problem and solve the quadratic programming problem to determine an optimal control signal at the current moment; the optimal control signal comprises optimal torque signals and optimal rotation angle signals of each wheel; and the vehicle is controlled to perform motion in the current cycle according to the optimal control signal. The technical scheme of the application can solve the problem that control optimality and real-time performance cannot be considered simultaneously.

Description

Technical Field

[0001] This application relates to the field of vehicle-related technology, specifically to a vehicle control method, device, equipment, and storage medium. Background Technology

[0002] With the deep integration of electrical and electronic technologies and advanced control theories, distributed drive electric vehicles, due to the independent controllability of the torque and steering angle of each wheel, provide a physical basis for developing high-performance dynamic control functions. Especially when vehicles face extreme conditions such as moose tests, emergency obstacle avoidance, and high-speed lane changes, higher demands are placed on the vehicle's response speed, vehicle stability, and the extreme performance of the control system.

[0003] However, traditional manual driving relies on driver operation, which suffers from problems such as significant human interference, difficulty in approaching limits, high risks in extreme operating conditions, and poor repeatability, making it difficult to achieve precise control under extreme conditions. In existing technologies, some solutions rely on iterative methods to find control parameters under extreme conditions, but this results in low testing efficiency and cannot guarantee optimality. Other solutions rely on complex nonlinear models to solve the optimal control problem, but due to the large computational load and convergence difficulties, they cannot meet real-time control requirements and cannot fully explore the capability boundaries of the intelligent chassis. Summary of the Invention

[0004] In view of the above problems, embodiments of this application provide a vehicle control method, apparatus, device and storage medium to solve the problem in the prior art that it is impossible to balance test optimization and control real-time performance.

[0005] According to one aspect of the embodiments of this application, a vehicle control method is provided, the method comprising: performing the following steps in each control cycle:

[0006] Obtain the vehicle's real-time status information at the current moment;

[0007] Based on the real-time state information and the pre-established vehicle dynamics model, the optimization proposition equation corresponding to the nonlinear optimal control problem with the objective of maximizing vehicle speed at the current moment is determined; wherein, the optimization proposition equation uses the torque of each wheel and the steering angle of each wheel as the control variables to be solved.

[0008] The optimization proposition equation is processed to transform the nonlinear optimal control problem into a quadratic programming problem and solve it to determine the optimal control signal at the current moment; wherein, the optimal control signal includes the optimal torque signal and the optimal steering angle signal of each wheel;

[0009] Based on the optimal control signal, the vehicle is controlled to perform the motion of the current cycle.

[0010] According to another aspect of the embodiments of this application, a vehicle control device is provided, wherein each processing unit in the device is configured to execute a corresponding step in each control cycle, the device comprising:

[0011] The first processing unit is used to obtain the real-time status information of the vehicle at the current moment;

[0012] The second processing unit is used to determine the optimization proposition equation corresponding to the nonlinear optimal control problem with the objective of maximizing vehicle speed at the current moment based on the real-time state information and the pre-established vehicle dynamics model; wherein the optimization proposition equation uses the torque of each wheel and the steering angle of each wheel as the control variables to be solved.

[0013] The third processing unit is used to process the optimization proposition equation to transform the nonlinear optimal control problem into a quadratic programming problem and solve it to determine the optimal control signal at the current moment; wherein, the optimal control signal includes the optimal torque signal and the optimal steering angle signal of each wheel;

[0014] The fourth processing unit is used to control the vehicle to perform the motion of the current cycle according to the optimal control signal.

[0015] According to another aspect of the embodiments of this application, an electronic device is provided, including: a processor, a memory, a communication interface, and a communication bus, wherein the processor, the memory, and the communication interface communicate with each other through the communication bus;

[0016] The memory is used to store at least one executable instruction that causes the processor to perform the operation of the vehicle control method as described above.

[0017] According to another aspect of the embodiments of this application, a computer-readable storage medium is provided, the storage medium storing at least one executable instruction that causes an electronic device / apparatus to perform the operation of the vehicle control method as described above.

[0018] This application's embodiments, within each control cycle, dynamically construct a nonlinear optimal control problem based on the vehicle's real-time state and dynamics model. This problem uses highly flexible multi-wheel independent torque and steering angle as control variables, aiming to maximize vehicle speed. It then transforms this problem into a quadratic programming problem for solution, effectively addressing the core issue in existing technologies that cannot simultaneously achieve optimality and real-time performance. This enables automated and precise exploration of the vehicle's moose test limits under autonomous driving conditions. Compared to traditional manual testing methods that rely on driver experience and trial-and-error, this method offers higher testing efficiency, better test consistency, and safety. Furthermore, it can more objectively and scientifically uncover and calibrate the vehicle chassis's ultimate handling stability, providing a powerful solution for vehicle chassis tuning and active safety control development.

[0019] The above description is merely an overview of the technical solutions of the embodiments of this application. In order to better understand the technical means of the embodiments of this application and to implement them in accordance with the contents of the specification, and to make the above and other objects, features and advantages of the embodiments of this application more obvious and understandable, specific implementation methods of this application are described below. Attached Figure Description

[0020] The accompanying drawings are for illustrative purposes only and are not intended to limit the scope of this application. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings:

[0021] Figure 1 A schematic flowchart of a vehicle control method provided in an embodiment of this application;

[0022] Figure 2 A schematic flowchart illustrating another vehicle control method provided in an embodiment of this application;

[0023] Figure 3 An architecture diagram of a real-time optimal control system for a vehicle provided in an embodiment of this application;

[0024] Figure 4 This is a schematic diagram of the structure of a vehicle control device provided in an embodiment of this application;

[0025] Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0026] Exemplary embodiments of the present application will now be described in more detail with reference to the accompanying drawings. Although exemplary embodiments of the present application are shown in the drawings, it should be understood that the present application may be implemented in various forms and should not be limited to the embodiments set forth herein.

[0027] The moose test is an important international standard for measuring vehicle safety. The moose test simulates extreme conditions where a vehicle is traveling at high speed and suddenly needs to avoid an obstacle. It requires the vehicle to complete maneuvers such as sharp turns and straightening within a very short time while maintaining vehicle stability. Furthermore, in real-world driving situations such as emergency obstacle avoidance, high-speed lane changes, and extreme cornering, even higher demands are placed on the vehicle's response speed, vehicle stability, and the ultimate performance of its control systems.

[0028] The existing technologies in the field of vehicle extreme condition control mainly include the following three types of solutions: (1) The initial speed and path parameters of the vehicle are repeatedly adjusted manually or semi-automatically to gradually approach the optimal control result. This type of method relies on a large number of trial and error, has a long testing cycle, cannot guarantee global optimality, and is difficult to adapt to the complex control requirements of distributed drive systems. (2) A nonlinear optimal control problem is established using a high-precision vehicle dynamics model, the optimal trajectory is calculated offline by a solver, and then tracked online by the underlying controller. This type of method has high computational complexity, low solution efficiency, and is difficult to meet real-time requirements. Furthermore, model simplification or parameter errors can easily lead to deviations in control effect. (3) The control logic of traditional driving assistance systems is reused to generate control signals through preset rules. Although this type of solution has high computational efficiency, it cannot fully explore the performance boundary of distributed drive systems and lacks the ability to verify control strategies under extreme vehicle conditions.

[0029] It is evident that existing vehicle control methods under extreme conditions cannot simultaneously guarantee optimality (maximizing vehicle speed) while achieving real-time solution and effectively applying control commands to the distributed drive chassis to explore the vehicle's ultimate performance. Therefore, this application proposes a vehicle control method that constructs a nonlinear optimal control problem based on a high-precision dynamic model, then approximates it as a quadratic programming problem using a real-time iterative strategy and solves it in real-time to obtain the optimal control signal, thereby achieving real-time optimal control.

[0030] It should be noted that the vehicle control method provided in this application embodiment can not only be applied to unmanned moose test scenarios, but also widely applied to extreme working condition control scenarios such as emergency obstacle avoidance, high-speed lane change, and extreme cornering during actual vehicle driving, and has strong versatility and practicality.

[0031] The technical solution of this application and how the technical solution of this application solves the above-mentioned technical problems are described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of this application will now be described with reference to the accompanying drawings.

[0032] It should be noted that the vehicle control method provided in this application can be executed by a vehicle control device, which can be deployed in a vehicle on a controller with data processing and control functions, such as a vehicle control unit (VCU), a central controller, a smart driving domain controller, or other electronic devices. This application does not impose any limitations, and the method of this application can be implemented by software, hardware, or a combination of software and hardware.

[0033] Figure 1This is a flowchart illustrating a vehicle control method provided in an embodiment of this application. This embodiment is applicable to distributed drive electric vehicles, particularly distributed drive electric vehicle chassis platforms with four-wheel steering and / or four-wheel drive. This method can be used in scenarios where the vehicle is subjected to a moose test (i.e., obstacle avoidance test) under unmanned driving conditions, and also in extreme operating condition control scenarios such as emergency obstacle avoidance and high-speed lane changes encountered by the vehicle during actual driving, to explore the physical limits of the vehicle's performance under extreme conditions and obtain the maximum passing speed. This method can be executed by a vehicle control device. Figure 1 As shown, when a vehicle faces extreme operating conditions, the following steps must be performed within each control cycle (e.g., every 10 milliseconds, 20 milliseconds, or other preset time intervals):

[0034] Step 110: Obtain the vehicle's real-time status information at the current moment.

[0035] For example, real-time status information refers to a set of data that comprehensively describes the vehicle's motion state and pose at the current moment. At the beginning of each control cycle, the vehicle's real-time status information can be obtained through various sensors deployed on the vehicle (such as vehicle speed sensors, inertial navigation systems, wheel speed sensors, angle sensors, etc.). This real-time status information includes, but is not limited to, the vehicle's longitudinal speed (…). ), lateral velocity ( ), yaw rate ( ), the rotational speed of each wheel ( ), and the longitudinal distance of the vehicle in the Frenet coordinate system (also known as the natural coordinate system). ), lateral deviation ( ) and heading angle deviation ( Among them, longitudinal and lateral velocities can be obtained by combining an inertial measurement unit (IMU) with a vehicle speed sensor, yaw rate can be obtained by a gyroscope, wheel speed can be obtained by a wheel speed sensor, and position deviation in the Frenet coordinate system can be calculated by comparing a high-precision positioning system (such as GPS+IMU fusion) with a preset reference path (such as the cone path in the moose test or lane lines / avoidance trajectories in actual driving).

[0036] By acquiring the vehicle's real-time status information, an accurate initial state is provided for subsequently constructing the optimal control problem. This is the foundation for realizing closed-loop feedback control, ensuring that the control algorithm can be optimized based on the vehicle's real state, and avoiding model mismatch and error accumulation.

[0037] Step 120: Based on real-time state information and a pre-established vehicle dynamics model, determine the optimization proposition equation corresponding to the nonlinear optimal control problem with the objective of maximizing vehicle speed at the current moment; wherein, the optimization proposition equation uses the torque of each wheel and the steering angle of each wheel as the control variables to be solved.

[0038] For example, the vehicle dynamics model is a pre-established, high-precision model used to describe the mathematical relationship between the vehicle's motion state and the control input. For instance, a seven-degree-of-freedom or higher-degree-of-freedom vehicle dynamics model considering tire nonlinear characteristics can be used. Preferably, a high-precision vehicle dynamics model can be established in the Frenet coordinate system to more intuitively describe the relative relationship between the vehicle and the reference path. The state variables of the vehicle dynamics model include at least: the vehicle's longitudinal velocity (…). ), lateral velocity ( ), yaw rate ( ), the rotational speed of each wheel ( ), and the longitudinal distance of the vehicle in the Frenet coordinate system (also known as the natural coordinate system). ), lateral deviation ( ) and heading angle deviation ( The control variable is the torque of each wheel (); ) and the turning angle of each wheel ( ).

[0039] After acquiring real-time state information, the vehicle control device can determine an optimization equation corresponding to a nonlinear optimal control problem with the objective of maximizing vehicle speed, based on the real-time state information and the aforementioned vehicle dynamics model. The core objective function of this optimization equation is "maximizing vehicle speed," that is, maximizing the vehicle's longitudinal speed while ensuring that the vehicle can travel safely and stably along a predetermined path (such as a moose test path or an avoidance trajectory path). For example, minimizing the negative value of the initial vehicle speed can be used as the objective function, i.e., Minimize−v0, thereby achieving the optimization objective of maximizing vehicle speed. Then, constraints are set in the optimization equation, which include at least the following types: initial state constraints, dynamic equation constraints, path tracking constraints, and actuator constraints. Among them, initial state constraints are used to ensure that the optimization starts from the current real state; dynamic equation constraints are used to ensure that the solved control signal conforms to the physical motion law of the vehicle; path tracking constraints are used to ensure that the vehicle does not deviate from the preset driving path boundary during extreme operating condition control; and actuator constraints are used to ensure that the control signal is within the capability range of the physical actuator. When setting constraints, each constraint can be set separately based on its purpose; this application embodiment does not impose any restrictions. Finally, by combining the above objective function and various constraints, a complete optimization equation corresponding to the nonlinear optimal control problem can be obtained.

[0040] By constructing an optimal control problem aimed at maximizing vehicle speed, the exploration of physical limits is transformed into a mathematical optimization problem, laying the mathematical model foundation for subsequently solving for the truly optimal control signal. Simultaneously, employing multiple independent torques and steering angles as control variables fully utilizes the execution capabilities of modern drive-by-wire chassis, enabling the control algorithm to search for optimal solutions within a broader context.

[0041] Step 130: Process the optimization proposition equation to transform the nonlinear optimal control problem into a quadratic programming problem and solve it to determine the optimal control signal at the current moment; wherein, the optimal control signal includes the optimal torque signal and the optimal steering angle signal of each wheel.

[0042] For example, directly solving nonlinear optimal control problems typically involves enormous computational costs, making it difficult to meet the real-time requirements of vehicle control. Therefore, this step employs advanced control and optimization theories, such as the linear time-varying (LTV) method or real-time iterative algorithms in Model Predictive Control (MPC). Within each control cycle, the complex nonlinear problem constructed in step 120 is linearized and discretized near the current reference trajectory and operating point, thus transforming it into a standard quadratic programming problem. Quadratic programming is a convex optimization problem with efficient and fast solution algorithms (such as the interior-point method, the effective set method, and the OSQP solver). Furthermore, the transformed quadratic programming problem is solved quickly to obtain the optimal control input sequence that satisfies all constraints within the current control cycle. A portion of the control variables in this sequence is output as the "optimal control signal at the current moment," which contains explicit instructions: the "optimal torque signal" that each wheel should apply and the "optimal steering angle signal" that each wheel should execute.

[0043] In one example, the reference trajectory and corresponding Lagrange multipliers at the current moment can be obtained first; then, at the current reference trajectory, the objective function and constraints of the nonlinear optimal control problem are linearized to approximate the nonlinear optimal control problem as a quadratic programming problem; then, this quadratic programming problem is solved to obtain the optimal control increment at the current moment and the updated Lagrange multipliers; finally, the obtained optimal control increment is superimposed on the control signal in the reference trajectory to obtain the optimal control signal at the current moment, which should include the optimal torque signal of each wheel (…). ) and optimal turning angle signal ( ).

[0044] This application's embodiments approximate the nonlinear optimal control problem as a quadratic programming problem through a real-time iterative strategy, achieving a balance between solution efficiency and accuracy, and resolving the contradiction between optimality and real-time performance. On the one hand, it preserves the accuracy of the high-precision nonlinear model, ensuring the optimality of the solution result; on the other hand, the quadratic programming problem can be solved quickly in embedded systems, meeting the timeliness requirements of real-time control.

[0045] Step 140: Control the vehicle to perform the motion of the current cycle according to the optimal control signal.

[0046] For example, after obtaining the optimal torque and optimal steering angle signals calculated in step 130, the optimal torque signals of each wheel can be sent to the motor controller (for electric vehicles) or the engine / brake electronic control unit (for conventional vehicles), and the optimal steering angle signals of each wheel can be sent to the steer-by-wire actuator. The underlying actuator responds to these commands, generating corresponding torque and steering angle, thereby achieving precise control of the vehicle's movement and completing the vehicle's control actions under extreme conditions in the current control cycle. By precisely executing the optimized commands, the vehicle can accurately follow the predetermined driving path at the highest stable speed achievable in the current cycle.

[0047] In one example, to simulate the input method of a human driver and comprehensively test the control performance of the vehicle control unit (VCU), the optimal control signal can be converted into brake / drive pedal signals and steering wheel angle signals, and then input to the vehicle's VCU. The VCU then controls the actuators such as wheel hub motors and steering motors according to these instructions, driving the vehicle to complete the vehicle motion control of the current cycle. This comprehensively tests the VCU's response speed, control accuracy, and signal processing capabilities during extreme operating condition control.

[0048] The vehicle control method provided in this application achieves real-time optimal extreme condition control of a distributed drive electric vehicle by cyclically executing the following process within each control cycle: acquiring real-time state, determining the optimization proposition equation corresponding to the nonlinear optimal control problem, transforming it into a quadratic programming problem for solution, and executing the control. This closed-loop process enables real-time optimal extreme condition control of a distributed drive electric vehicle. On the one hand, this method ensures that the optimization proposition equation accurately describes the vehicle's real physical characteristics through a high-precision nonlinear vehicle dynamics model, thereby ensuring that the solved control signal can approximate the chassis's ultimate performance and maximize the vehicle's speed. On the other hand, by approximating the nonlinear optimal control problem, which is difficult to solve in real-time, into a quadratic programming problem, the computational complexity is significantly reduced, meeting the real-time requirements of embedded systems and effectively resolving the core contradiction in existing technologies that cannot simultaneously achieve optimality and real-time performance.

[0049] Figure 2This is a schematic flowchart illustrating another vehicle control method provided in an embodiment of this application. This embodiment... Figure 1 Based on the illustrated embodiment, the specific implementation method of applying the optimal control signal to the vehicle is further defined, in particular, the optimal control signal is converted into an input signal that can be recognized by the vehicle control unit through the reverse calibration method, so as to achieve comprehensive testing of the VCU control performance.

[0050] The vehicle control method provided in this embodiment can be executed by a vehicle control device, which can be implemented in software and / or hardware and integrated into the vehicle's control system, domain controller, or host computer. Figure 2 As shown, when a vehicle faces extreme operating conditions, the following steps must be performed in each control cycle:

[0051] Step 210: Obtain the vehicle's real-time status information at the current moment.

[0052] It should be noted that the specific implementation of step 210 can be referred to the description of step 110, and will not be repeated here.

[0053] Step 220: Based on real-time state information and a pre-established vehicle dynamics model, determine the optimization proposition equation corresponding to the nonlinear optimal control problem with the objective of maximizing vehicle speed at the current moment.

[0054] It should be noted that the specific implementation of step 220 can be referred to the description of step 120, and will not be repeated here.

[0055] For example, a high-precision vehicle dynamics model needs to be established first. This model can be built in the Frenet coordinate system to more intuitively describe the relative relationship between the vehicle and the reference path. Optionally, the state variables of the vehicle dynamics model include at least: the vehicle's longitudinal velocity, lateral velocity, yaw rate, wheel speeds, and the vehicle's longitudinal distance, lateral deviation, and heading angle deviation in the natural coordinate system (i.e., the Frenet coordinate system); the control variables of the vehicle dynamics model are the torque of each wheel and the steering angle of each wheel.

[0056] For example, a high-precision vehicle dynamics model established in the Frenet coordinate system can be as follows:

[0057]

[0058]

[0059]

[0060]

[0061]

[0062]

[0063] in, , and These represent the vehicle's longitudinal velocity, lateral velocity, and yaw rate in the vehicle coordinate system, respectively. This represents the longitudinal distance traveled by the vehicle in the Frenet coordinate system. This indicates the deviation between the vehicle's actual heading angle and the reference heading angle (i.e., the aforementioned heading angle deviation). This indicates the deviation between the vehicle's actual lateral position and the reference lateral position (i.e., the aforementioned lateral deviation). Indicates the curvature of the road about The function, Indicates the longitudinal resultant force. Indicates the lateral resultant force. Indicates yaw moment, This represents the vehicle's moment of inertia. Indicates the overall vehicle weight. This indicates air resistance.

[0064] Among them, longitudinal resultant force Lateral resultant force and yaw moment It is the torque of the hub motor of each wheel ( ) and wheel angle ( This is calculated using a tire model and a vertical load distribution model. The specific calculation method is as follows:

[0065]

[0066] Where, in the formula ; ; ; ; and These represent the distances from the front axle and rear axle to the center of mass, respectively. Indicates wheelbase. Indicates the first The turning angle of each wheel. , They represent the first The longitudinal and lateral forces of each wheel.

[0067] In this embodiment, a high-precision magic formula can be used to describe the nonlinear relationship between tire longitudinal and lateral forces and slip ratio and sideslip angle. The magic formula, by incorporating a combination of nonlinear function terms such as sine and arctangent, can accurately fit the mechanical characteristics of the tire near its adhesion limit, thereby ensuring the accuracy of the model under extreme conditions. Specifically, , The calculation method can be as follows:

[0068]

[0069]

[0070] in,

[0071]

[0072]

[0073]

[0074]

[0075]

[0076]

[0077]

[0078]

[0079]

[0080]

[0081]

[0082] in, and They represent the first Wheel slip angle and slip ratio; and They represent the first The longitudinal and lateral adhesion coefficients of the wheel; Indicates the first Vertical load on the wheel; , These represent the theoretically maximum longitudinal force and maximum lateral force that can be generated under the current road surface and vertical load, respectively. Indicates the first The rotational speed of the wheel; Indicates the first The longitudinal speed of the wheel; Indicates the first The wheel hub motor input torque. Other parameters are tire dynamics parameters; for example, B represents the stiffness factor, which determines the initial slope of the curve; C represents the shape factor, which determines whether the curve resembles a sine, cosine, or other shape; and E represents the curvature factor, which determines the shape and position near the curve's peak. , Similarly, tire characteristic parameters determine the speed and extent of the effect of the slip angle on longitudinal force; similarly, and It is also a tire characteristic parameter.

[0083] Wheel speed and The calculation method can be as follows:

[0084]

[0085]

[0086]

[0087]

[0088]

[0089]

[0090]

[0091]

[0092]

[0093]

[0094]

[0095]

[0096] Furthermore, without considering pitch and roll, the vertical load should also satisfy the following constraints:

[0097]

[0098]

[0099]

[0100]

[0101] The high-precision vehicle dynamics model of this application can be described by algebraic differential equations:

[0102]

[0103]

[0104] in,

[0105]

[0106]

[0107] In practical applications, because the vehicle control device runs in real time within an embedded system, the computer cannot directly process continuous differentials. Therefore, it must be transformed into a discrete algebraic equation, and the resulting discretized model can be as follows:

[0108]

[0109]

[0110] Based on the above regarding longitudinal resultant force Lateral resultant force and yaw moment The calculation formula, during the calculation process, can be based on the longitudinal velocity obtained from the state at the previous moment or from sensors. lateral velocity yaw rate Vehicle speed and the input torque obtained from the current control command. and wheel angle First, calculate the longitudinal force of each wheel. lateral force Then calculate the longitudinal force and lateral force Including the wheel corner Substituting these values ​​into the initial force composition formula, we can calculate the longitudinal resultant force at the vehicle's center of gravity. Lateral resultant force and yaw moment Furthermore, the longitudinal resultant force can be... Lateral resultant force and yaw moment Introduced into a high-precision vehicle dynamics model , , Formula, calculate the differential equation Mid-vector In , , Assuming the current time is k and the sampling time is ΔT (e.g., 0.01 seconds), the state at the next time k+1 can also be calculated, thus determining the discretization equation. The corresponding parameter value in the file.

[0111] After establishing the vehicle dynamics model, the optimization equations corresponding to the nonlinear optimal control problem with the objective of maximizing vehicle speed can be determined based on the aforementioned high-precision dynamics model of distributed electric vehicles.

[0112] Optionally, in one possible embodiment, based on real-time state information and a pre-established vehicle dynamics model, determining the optimization proposition equation corresponding to the nonlinear optimal control problem with the objective of maximizing vehicle speed at the current moment may include:

[0113] S1. Minimize the negative value of the initial vehicle speed as the objective function of the nonlinear optimal control problem;

[0114] S2. Based on real-time state information, vehicle dynamics model, preset path rules, and actuator capabilities of the distributed drive system, determine the constraints of the nonlinear optimal control problem.

[0115] S3. Combine the objective function and constraints to obtain the optimization equations corresponding to the nonlinear optimal control problem.

[0116] For example, the industry-standard evaluation criteria for extreme condition control, such as the moose test, is typically the maximum safe passing speed. Therefore, the optimization objective can be set as maximizing the initial speed, which is mathematically equivalent to minimizing the negative value of the initial speed, i.e. This objective function directly corresponds to the core evaluation index of control, giving the optimization problem a clear physical meaning.

[0117] Constraints are restrictions on the range of values ​​for the optimization variables, used to ensure that the solved control signal is physically feasible and conforms to driving rules. Real-time state information, vehicle dynamics model, preset path rules, and actuator capabilities of the distributed drive system reflect the actual state of the vehicle, physical laws, path requirements, and actuator limitations, respectively. Therefore, constraints for the nonlinear optimal control problem can be determined based on real-time state information, vehicle dynamics model, preset path rules, and actuator capabilities of the distributed drive system.

[0118] Optionally, in one possible embodiment, the constraints for determining the nonlinear optimal control problem based on real-time state information, vehicle dynamics model, preset path rules, and actuator capabilities of the distributed drive system may include:

[0119] S21. Use real-time state information as the initial state constraint for the nonlinear optimal control problem;

[0120] S22. The dynamic equations described by the vehicle dynamics model serve as dynamic equality constraints for the nonlinear optimal control problem.

[0121] S23. The linear inequality that the vehicle lateral position deviation determined based on the preset path rules must satisfy is used as the path tracking constraint for the nonlinear optimal control problem.

[0122] S24. The upper and lower limits of the torque and the upper and lower limits of the rotation angle of each wheel, determined by the actuator capability of the distributed drive system, are used as actuator constraints for the nonlinear optimal control problem.

[0123] For example, when setting constraints, the real-time state information obtained in step 210 can be used as the initial state constraints for the nonlinear optimal control problem, i.e. This constraint ensures that optimization begins with the vehicle's current true position and state, which is key to achieving feedback control.

[0124] The differential equations described by the pre-established high-precision vehicle dynamics model ( As an equality constraint. Therefore, during the optimization process, any control sequence considered ( The state at the next moment caused by ) All of these must be strictly equal to those derived from the dynamic model. The calculated results ensure that the solved control signal conforms to the actual physical motion of the vehicle.

[0125] The lateral deviation (e) of the vehicle in the Frenet coordinate system is limited to a very small range, i.e., e min ≤e k ≤e max This linear inequality constraint directly corresponds to the rule of "the vehicle does not deviate from the preset driving path boundary" in the extreme condition control process, ensuring that the vehicle can drive along the preset reference path without deviating. Among them, the lateral position deviation range can be determined according to the preset path rules such as the cone layout of the moose test or the lane line boundaries and obstacle safety distances in the actual driving scenario.

[0126] The torque (T) of each wheel i The torque is limited to the maximum torque that the motor can output, i.e., T. i,min ≤ T i,k ≤T i,max ; Adjust the rotation angle (δ) of each wheel i ) is limited to the mechanical limit range of the steering mechanism, i.e., δ i,min ≤ δ i,k ≤ δi,max These inequality constraints ensure that the solved control signal is physically executable. Specifically, the upper and lower limits of the torque for each wheel are determined by the physical output capability of the hub motor, and the upper and lower limits of the steering angle for each wheel are determined by the mechanical limits of the steering mechanism. In a specific control scenario, if the control scheme does not allow braking (e.g., only testing driving capability), the lower limit T of the torque for each wheel can be set... i,min Set to 0 to eliminate braking interference.

[0127] Finally, by integrating the above objective function and various constraints, a complete optimization equation corresponding to the nonlinear optimal control problem can be formed, the general form of which can be expressed as:

[0128]

[0129]

[0130]

[0131] Furthermore, the optimization variables in the above formula are integrated into... Vectors integrate equality constraints into Integrating inequality constraints into The optimization equation corresponding to the above nonlinear optimal control problem can be expressed as:

[0132]

[0133]

[0134]

[0135] This optional embodiment establishes the optimal propositional equations for the nonlinear optimal control problem aimed at maximizing vehicle speed, and fully considers multiple constraints such as vehicle dynamics, path rules, and actuator capabilities, laying a solid mathematical model foundation for subsequently solving for the truly optimal control signal. Furthermore, the application of the Frenet coordinate system allows path tracking constraints to be simply represented as linear inequalities, reducing problem complexity; and using the signals of the distributed-drive underlying actuators as control variables maximizes the control potential of the intelligent chassis.

[0136] Step 230: Process the optimization proposition equation to transform the nonlinear optimal control problem into a quadratic programming problem and solve it to determine the optimal control signal at the current moment; wherein, the optimal control signal includes the optimal torque signal and the optimal steering angle signal of each wheel.

[0137] It should be noted that the specific implementation of step 230 can be referred to the description of step 130, and will not be repeated here.

[0138] For example, since the optimization proposition equation determined in step 220 is difficult to solve directly and takes a long time to calculate, it is difficult to meet the real-time requirements of the embedded system. This step adopts a real-time iterative strategy to process it and transform it into a convex quadratic programming problem that can be solved quickly.

[0139] Optionally, in one possible embodiment, the optimization proposition equation is processed to transform the nonlinear optimal control problem into a quadratic programming problem and solved to determine the optimal control signal at the current time. This may include:

[0140] S10. Obtain the reference trajectory and corresponding Lagrange multipliers at the current moment;

[0141] S20. Linearize the objective function and constraints of the nonlinear optimal control problem at the reference trajectory, and transform the second-order term of the objective function into the Hessian matrix of the Lagrange function to determine a quadratic programming problem.

[0142] S30. Solve the quadratic programming problem to obtain the optimal control increment and the updated Lagrange multipliers;

[0143] S40. The optimal control increment is superimposed on the control signal in the reference trajectory to obtain the optimal control signal at the current moment.

[0144] For example, the reference trajectory is a vector set containing the current guessed values ​​of all state variables, control variables, and algebraic variables throughout the entire prediction time domain. In the first control cycle, the reference trajectory can be pre-generated through offline simulation or based on a simplified model; in subsequent cycles, the reference trajectory is typically obtained by time-domain shifting the optimal solution from the previous control cycle. Lagrange multipliers carry sensitivity information about the constraints, which helps accelerate convergence. When transforming the problem, the reference trajectory used for linearization at the current time step is first obtained (…). ) and the corresponding Lagrange multipliers ( and ).

[0145] Furthermore, in the current reference trajectory ( At point (), the objective function (J) and constraints (h, g) of the original nonlinear optimal control problem are expanded. The objective function is expanded to second order, and the constraints to first order. Simultaneously, the coefficient matrix of the second-order term of the objective function is approximated as the Hessian matrix (H) of the Lagrangian function with respect to the optimization variable (y). Through this process, the original nonlinear optimal control problem can be approximated as a quadratic programming problem of the following form:

[0146]

[0147]

[0148]

[0149] in, Indicates the optimization variable Reference value, Represents the Lagrange function with respect to variables The Hessian matrix, and Let represent the Lagrange multipliers. The Lagrange function is defined as: .

[0150] Optionally, in one possible embodiment, the objective function and constraints of the nonlinear optimal control problem are linearized at the reference trajectory, and the second-order terms of the objective function are transformed into the Hessian matrix of the Lagrangian function to determine a quadratic programming problem, which may include:

[0151] S201. Perform a second-order Taylor expansion on the objective function, retain the quadratic terms, and transform the coefficient matrix of the quadratic terms into the Hessian matrix of the Lagrangian function with respect to the optimization variables; wherein, the Lagrangian function is composed of the objective function and the constraints multiplied by Lagrange multipliers;

[0152] S202. Perform a first-order Taylor expansion on the constraints, retaining the linear terms;

[0153] S203. Based on the results of the second-order Taylor expansion and the first-order Taylor expansion, determine the quadratic programming problem.

[0154] For example, at the reference trajectory, the objective function in the nonlinear optimal control problem... A second-order Taylor expansion yields a coefficient matrix (i.e., the Hessian matrix) for the quadratic terms, which can be expressed using the Lagrange function. Approximating with the second derivative, we get: This approximation takes into account the influence of constraints on the curvature of the objective function, which can improve the convergence of the constructed quadratic programming problem.

[0155] Equality constraints and inequality constraints Perform a first-order Taylor expansion at the reference trajectory, i.e.: , Retaining linear terms ensures that the constraints remain linear after approximation, thus guaranteeing the convexity of the problem.

[0156] By combining the quadratic objective function obtained in S201 and the linear constraints obtained in S202, we can construct the standard quadratic programming problem form described above for the solver to solve.

[0157] When solving the above quadratic programming problem, since the problem is convex, various efficient algorithms (such as the interior-point method and the effective set method) can be used to quickly solve it in embedded systems to obtain the optimal control increment at the current moment. ) and the newer Lagrange multipliers ( and The optimal control increment obtained by solving (). The control section in ) The control signal superimposed on the reference trajectory ( By analyzing the data, the optimal control signal for the current moment can be obtained, that is: The optimal control signal includes the optimal torque signal for each wheel. ) and optimal turning angle signal ( ).

[0158] Furthermore, the reference trajectory and corresponding Lagrange multipliers for the next time step can be updated based on the solution results, preparing for the next control cycle. Specifically:

[0159]

[0160]

[0161]

[0162] in, Control the iteration step size.

[0163] This optional embodiment approximates the nonlinear optimal control problem as a quadratic programming problem through a real-time iterative strategy, achieving a balance between solution efficiency and accuracy. On the one hand, it preserves the accuracy of the high-precision nonlinear model, ensuring the optimality of the solution result; on the other hand, the quadratic programming problem can be solved quickly in embedded systems, meeting the timeliness requirements of real-time control. This is the core technical means to resolve the contradiction between "optimality and real-time performance".

[0164] Step 240: Using the reverse calibration method, convert the optimal torque signal into the vehicle's brake / drive pedal signal, and convert the optimal steering angle signal into the vehicle's steering wheel angle signal.

[0165] For example, by solving the above quadratic programming problem in each control cycle, the optimal control signal input to the vehicle at the current moment can be obtained:

[0166]

[0167] The optimal control signal acts directly on the vehicle. To demonstrate the role of the Vehicle Control Unit (VCU) during extreme operating conditions and to comprehensively test the control performance of the VCU and other vehicle control units, it is necessary to simulate the input of a human driver and input the control signal to the VCU. Therefore, this technical solution also proposes a reverse calibration method for signal conversion. The reverse calibration method refers to mapping the theoretically optimal four-wheel independent control signals in reverse to a unified control command generated by a human driver operating a conventional vehicle.

[0168] For the optimal torque signal, the independent optimal torque signals of the four wheels can be obtained. Through a pre-calibrated mapping relationship, the signals are synthesized and converted into a unified brake / drive pedal opening signal. For example, the corresponding pedal opening can be obtained by finding the preset pedal characteristic curve based on the average or weighted sum of the total demand driving force or braking force.

[0169] For the optimal steering angle signal, the independent optimal steering angle signals of the four wheels can be obtained. Through a pre-defined mapping relationship, the signals are comprehensively converted into a unified steering wheel angle signal. For example, for a four-wheel steering vehicle, the corresponding steering wheel angle can be calculated by using the geometric relationship between the front and rear wheel angles.

[0170] By using a reverse calibration method, the high-dimensional independent control signals of the four wheels are reduced in dimension and converted into pedal and steering wheel signals of the vehicle's standard interface. This not only enables the theoretically optimal control commands to be executed in a real vehicle, but more importantly, it simulates the real input scenarios of a human driver, creating conditions for testing the response speed and control accuracy of the vehicle control unit (VCU).

[0171] Step 250: Input the brake / drive pedal signal and steering wheel angle signal to the vehicle control unit to drive the vehicle to perform the motion of the current cycle.

[0172] For example, the brake / drive pedal signal obtained in step 240 ( ) and steering wheel angle signal ( These commands are sent as input instructions to the vehicle's vehicle control unit (VCU). As the core control unit of the vehicle, the VCU receives these commands, parses them according to its internal control logic and software algorithms, generates corresponding control signals for the underlying actuators, and ultimately drives actuators such as wheel hub motors and steering motors to perform actions, enabling the vehicle to complete the vehicle motion control actions of the current cycle according to the expected trajectory and speed.

[0173] By inputting control commands to the VCU for execution, the VCU is effectively incorporated into the control closed loop. During control under extreme conditions, the VCU's response speed, control accuracy, signal processing capabilities, and the robustness of its internal algorithms will be thoroughly tested. Compared to traditional control methods that directly send commands to the actuators, this method can simultaneously explore the physical limits of the chassis and verify the control performance of the upper-level controller (VCU) within a unified framework, achieving in-depth testing of the software-defined vehicle capabilities of intelligent electric vehicles.

[0174] The vehicle control method provided in this application implements a closed-loop process within each control cycle: acquiring real-time state, determining the optimization equation corresponding to the nonlinear optimal control problem, transforming it into a quadratic programming problem, back-calibrating, and inputting it into the VCU for execution. This achieves real-time optimal extreme condition control for distributed drive electric vehicles. This method ensures the accurate description of the vehicle's true physical characteristics by using a high-precision vehicle dynamics model and multiple constraints; it solves the real-time solution challenge by approximating the nonlinear problem as a quadratic programming problem through a real-time iteration strategy; and it achieves comprehensive testing of the VCU control performance by converting the optimal control signal into the standard input of the VCU through back-calibration. This method effectively resolves the core contradiction in existing technologies that cannot simultaneously achieve optimality and real-time performance, and expands the application depth and value of extreme condition control methods, providing an efficient and reliable testing method for the research and development of active safety performance of intelligent electric vehicles and the verification of domain control software.

[0175] For example, Figure 3 This is an architecture diagram of a real-time optimal control system for a vehicle provided in an embodiment of this application. Figure 3 As shown, within each control cycle, a nonlinear optimal control problem with the objective of maximizing vehicle speed is first determined based on the real-time feedback of vehicle state information and a pre-established high-precision vehicle dynamics model. This problem is then transformed into a quickly solvable quadratic programming problem using a real-time iterative strategy, yielding the optimal torque and steering angle signals for each wheel. Subsequently, through steering angle reverse calibration (MAP) and torque reverse calibration (MAP), the independent optimal steering angle and torque signals of the four wheels are converted into unified steering wheel angle and brake / drive pedal opening signals, respectively, and input to the vehicle control unit (VCU). Finally, the VCU drives the vehicle to execute the vehicle motion control action for the current cycle, while the actual vehicle state information is fed back to the optimization problem construction stage in real time, forming a closed-loop control. This allows for a comprehensive exploration of the physical limits of the vehicle chassis and the VCU control performance under extreme conditions. This application solves the real-time calculation problem of nonlinear optimal control under extreme conditions through real-time iterative technology and closely integrates theoretical optimal control with actual VCU testing through reverse calibration technology, thus systematically solving the problem of the inability to achieve both optimal and real-time performance.

[0176] The vehicle control method of this application is applicable to extreme vehicle control scenarios, including but not limited to: unmanned moose test scenarios, scenarios where a vehicle encounters an obstacle at high speed and needs to make emergency avoidance maneuvers, scenarios where a vehicle makes an emergency lane change or avoids an obstacle on a highway, and scenarios where a vehicle passes through a curve at its maximum speed. Through a high-precision nonlinear model and a real-time iterative solution strategy, optimal control is achieved not only in each scenario but also an efficient and universal technical platform for the development of active safety features and the iteration of domain control software for intelligent electric vehicles.

[0177] Figure 4 This is a schematic diagram of a vehicle control device provided in an embodiment of this application. Each processing unit in the vehicle control device of this application is used to execute corresponding steps in each control cycle, such as... Figure 4 As shown, the device 40 includes: a first processing unit 401, a second processing unit 402, a third processing unit 403, and a fourth processing unit 404.

[0178] The first processing unit 401 is used to obtain the real-time status information of the vehicle at the current moment;

[0179] The second processing unit 402 is used to determine the optimization proposition equation corresponding to the nonlinear optimal control problem with the objective of maximizing vehicle speed at the current moment based on real-time state information and a pre-established vehicle dynamics model; wherein the optimization proposition equation uses the torque of each wheel and the steering angle of each wheel as the control variables to be solved.

[0180] The third processing unit 403 is used to process the optimization proposition equation to transform the nonlinear optimal control problem into a quadratic programming problem and solve it to determine the optimal control signal at the current moment; wherein, the optimal control signal includes the optimal torque signal and the optimal steering angle signal of each wheel;

[0181] The fourth processing unit 404 is used to control the vehicle to perform the motion of the current cycle according to the optimal control signal.

[0182] In one alternative embodiment, the second processing unit 402 is specifically used for:

[0183] Minimizing the negative value of the initial vehicle speed is taken as the objective function of the nonlinear optimal control problem;

[0184] Based on real-time state information, vehicle dynamics model, preset path rules, and actuator capabilities of distributed drive system, the constraints of nonlinear optimal control problem are determined.

[0185] By combining the objective function and constraints, the optimization equations corresponding to the nonlinear optimal control problem are obtained.

[0186] In one alternative embodiment, the second processing unit 402 is specifically used for:

[0187] Real-time state information is used as the initial state constraint for the nonlinear optimal control problem;

[0188] The dynamic equations described by the vehicle dynamics model are used as dynamic equality constraints for the nonlinear optimal control problem.

[0189] The linear inequality that the vehicle's lateral position deviation, determined based on preset path rules, must satisfy is used as the path tracking constraint for the nonlinear optimal control problem.

[0190] The upper and lower limits of the torque and the upper and lower limits of the rotation angle of each wheel, determined by the actuator capability of the distributed drive system, are used as actuator constraints for the nonlinear optimal control problem.

[0191] In one alternative approach, the state variables of the vehicle dynamics model include at least: the vehicle's longitudinal velocity, lateral velocity, yaw rate, wheel speeds, and the vehicle's longitudinal distance, lateral deviation, and heading angle deviation in the natural coordinate system.

[0192] The control variables for the vehicle dynamics model are the torque and steering angle of each wheel.

[0193] In one alternative embodiment, the third processing unit 403 is specifically used for:

[0194] Obtain the reference trajectory and corresponding Lagrange multipliers at the current moment;

[0195] The objective function and constraints of the nonlinear optimal control problem are linearized at the reference trajectory, and the second-order term of the objective function is transformed into the Hessian matrix of the Lagrange function to determine a quadratic programming problem.

[0196] Solving the quadratic programming problem yields the optimal control increment and the updated Lagrange multipliers;

[0197] The optimal control increment is superimposed on the control signal in the reference trajectory to obtain the optimal control signal at the current moment.

[0198] In one alternative embodiment, the third processing unit 403 is specifically used for:

[0199] Perform a second-order Taylor expansion on the objective function, retain the quadratic terms, and transform the coefficient matrix of the quadratic terms into the Hessian matrix of the Lagrangian function with respect to the optimization variables; where the Lagrangian function is composed of the objective function and the constraints multiplied by Lagrange multipliers.

[0200] Perform a first-order Taylor expansion on the constraints, retaining the linear terms;

[0201] Based on the results of the second-order Taylor expansion and the first-order Taylor expansion, the convex quadratic programming problem is determined.

[0202] In one alternative approach, the method of this application is applied to vehicle extreme condition control scenarios, including unmanned moose test scenarios and vehicle emergency obstacle avoidance scenarios.

[0203] In one alternative embodiment, the fourth processing unit 404 is specifically used for:

[0204] The optimal torque signal is converted into the vehicle's brake / drive pedal signal and the optimal steering angle signal is converted into the vehicle's steering wheel angle signal using the reverse calibration method.

[0205] The brake / drive pedal signal and steering wheel angle signal are input to the vehicle control unit to drive the vehicle to perform the motion of the current cycle.

[0206] As can be seen from the above, the vehicle control device provided in this application realizes real-time optimal extreme condition control of distributed drive electric vehicles by cyclically executing the following process in each control cycle: acquiring real-time status, determining nonlinear optimal control problem, transforming it into a quadratic programming solution, and executing control. This effectively solves the core contradiction in the prior art that cannot balance optimality and real-time performance.

[0207] Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. The specific embodiments of this application do not limit the specific implementation of the electronic device.

[0208] like Figure 5 As shown, the electronic device may include: a processor 502, a communications interface 504, a memory 506, and a communications bus 508.

[0209] The processor 502, communication interface 504, and memory 506 communicate with each other via communication bus 508. Communication interface 504 is used to communicate with other network elements such as clients or other servers. The processor 502 executes program 510, specifically performing the relevant steps described above in the vehicle control method embodiment.

[0210] Specifically, program 510 may include program code, which includes computer-executable instructions.

[0211] Processor 502 may be a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of this application. The electronic device includes one or more processors, which may be processors of the same type, such as one or more CPUs; or they may be processors of different types, such as one or more CPUs and one or more ASICs.

[0212] Memory 506 is used to store program 510. Memory 506 may include high-speed RAM memory, and may also include non-volatile memory, such as at least one disk storage device.

[0213] Specifically, program 510 can be called by processor 502 to cause the electronic device to perform the relevant steps described above in the embodiment of the vehicle control method.

[0214] This application provides a computer-readable storage medium storing at least one executable instruction that, when executed on an electronic device, causes the electronic device to perform the vehicle control method in any of the above method embodiments.

[0215] The algorithms or displays provided herein are not inherently related to any particular computer, virtual system, or other device. Furthermore, the embodiments in this application are not directed to any particular programming language.

[0216] Numerous specific details are set forth in the specification provided herein. However, it will be understood that embodiments of this application may be practiced without these specific details. Similarly, for the purpose of simplification and aiding understanding of one or more aspects of the invention, in the above description of exemplary embodiments of this application, various features of the embodiments are sometimes grouped together in a single embodiment, figure, or description thereof. The claims, which follow the detailed description, are hereby expressly incorporated into that detailed description, wherein each claim itself is a separate embodiment of this application.

[0217] Those skilled in the art will understand that the modules in the device of the embodiment can be adaptively changed and placed in one or more devices different from that embodiment. Modules, units, or components in the embodiment can be combined into a single module, unit, or component, and further, they can be divided into multiple sub-modules, sub-units, or sub-components, except that at least some of such features and / or processes or units are mutually exclusive.

[0218] It should be noted that the above embodiments are illustrative of this application and not restrictive, and those skilled in the art can devise alternative embodiments without departing from the scope of the appended claims. In the claims, any reference signs placed between parentheses should not be construed as limiting the claims. The word "comprising" does not exclude the presence of elements or steps not listed in the claims. The word "a" or "an" preceding an element does not exclude the presence of a plurality of such elements. This application can be implemented by means of hardware comprising several different elements and by means of a suitably programmed computer. In the unit claims enumerating several means, several of these means may be embodied by the same item of hardware. The use of the words first, second, and third, etc., does not indicate any order. These words can be interpreted as names. The steps in the above embodiments, unless otherwise specified, should not be construed as limiting the order of execution.

Claims

1. A vehicle control method, characterized in that, The method includes performing the following steps within each control cycle: Obtain the vehicle's real-time status information at the current moment; Based on the real-time state information and the pre-established vehicle dynamics model, the optimization proposition equation corresponding to the nonlinear optimal control problem with the objective of maximizing vehicle speed at the current moment is determined; wherein, the optimization proposition equation uses the torque of each wheel and the steering angle of each wheel as the control variables to be solved. The optimization proposition equation is processed to transform the nonlinear optimal control problem into a quadratic programming problem and solve it to determine the optimal control signal at the current moment; wherein, the optimal control signal includes the optimal torque signal and the optimal steering angle signal of each wheel; Based on the optimal control signal, the vehicle is controlled to perform the motion of the current cycle.

2. The method according to claim 1, characterized in that, The optimization equations for the nonlinear optimal control problem aimed at maximizing the test vehicle speed at the current moment, based on the real-time state information and the pre-established vehicle dynamics model, are determined, including: Minimizing the negative value of the initial vehicle speed is taken as the objective function of the nonlinear optimal control problem. Based on the real-time status information, the vehicle dynamics model, the preset path rules, and the actuator capabilities of the distributed drive system, the constraints of the nonlinear optimal control problem are determined. By combining the objective function and the constraints, the optimization equation corresponding to the nonlinear optimal control problem is obtained.

3. The method according to claim 2, characterized in that, The determination of constraints for the nonlinear optimal control problem based on the real-time state information, the vehicle dynamics model, preset path rules, and the actuator capabilities of the distributed drive system includes: The real-time state information is used as the initial state constraint for the nonlinear optimal control problem. The dynamic equations described by the vehicle dynamics model are used as dynamic equality constraints for the nonlinear optimal control problem. The linear inequality that the vehicle lateral position deviation determined based on the preset path rules must satisfy is used as the path tracking constraint for the nonlinear optimal control problem. The upper and lower limits of the torque and the upper and lower limits of the steering angle of each wheel, determined based on the actuator capability of the distributed drive system, are used as actuator constraints for the nonlinear optimal control problem.

4. The method according to claim 1, characterized in that, The state variables of the vehicle dynamics model include at least: the vehicle's longitudinal velocity, lateral velocity, yaw rate, wheel speeds, and the vehicle's longitudinal distance, lateral deviation, and heading angle deviation in the natural coordinate system. The control variables of the vehicle dynamics model are the torque and steering angle of each wheel.

5. The method according to claim 1, characterized in that, The process of processing the optimization proposition equation to transform the nonlinear optimal control problem into a quadratic programming problem and solving it to determine the optimal control signal at the current moment includes: Obtain the reference trajectory and corresponding Lagrange multipliers at the current moment; The objective function and constraints of the nonlinear optimal control problem are linearized at the reference trajectory, and the second-order term of the objective function is transformed into the Hessian matrix of the Lagrange function to determine a quadratic programming problem. Solving the quadratic programming problem yields the optimal control increment and the updated Lagrange multipliers; The optimal control increment is superimposed on the control signal in the reference trajectory to obtain the optimal control signal at the current moment.

6. The method according to claim 5, characterized in that, The process of linearizing the objective function and constraints of the nonlinear optimal control problem at the reference trajectory, and transforming the second-order term of the objective function into the Hessian matrix of the Lagrangian function, determines a quadratic programming problem, including: The objective function is expanded using a second-order Taylor series, retaining the quadratic terms, and the coefficient matrix of the quadratic terms is transformed into a Hessian matrix of the Lagrange function with respect to the optimization variables; wherein the Lagrange function is composed of the objective function and the constraints multiplied by the Lagrange multipliers. Perform a first-order Taylor expansion on the constraints, retaining the linear terms; Based on the results of the second-order Taylor expansion and the first-order Taylor expansion, the quadratic programming problem is determined.

7. The method according to claim 1, characterized in that, The method is applied to vehicle extreme operating condition control scenarios, including unmanned moose test scenarios and vehicle emergency obstacle avoidance scenarios.

8. The method according to any one of claims 1-7, characterized in that, The step of controlling the vehicle to perform the motion of the current cycle according to the optimal control signal includes: The optimal torque signal is converted into a vehicle brake / drive pedal signal and the optimal steering angle signal is converted into a vehicle steering wheel angle signal using a reverse calibration method. The brake / drive pedal signal and the steering wheel angle signal are input to the vehicle control unit to drive the vehicle to perform the motion of the current cycle.

9. A vehicle control device, characterized in that, Each processing unit in the device is used to execute a corresponding step in each control cycle, and the device includes: The first processing unit is used to obtain the real-time status information of the vehicle at the current moment; The second processing unit is used to determine the optimization proposition equation corresponding to the nonlinear optimal control problem with the objective of maximizing vehicle speed at the current moment based on the real-time state information and the pre-established vehicle dynamics model; wherein the optimization proposition equation uses the torque of each wheel and the steering angle of each wheel as the control variables to be solved. The third processing unit is used to process the optimization proposition equation to transform the nonlinear optimal control problem into a quadratic programming problem and solve it to determine the optimal control signal at the current moment; wherein, the optimal control signal includes the optimal torque signal and the optimal steering angle signal of each wheel; The fourth processing unit is used to control the vehicle to perform the motion of the current cycle according to the optimal control signal.

10. An electronic device, characterized in that, include: The processor, memory, communication interface, and communication bus are provided, wherein the processor, memory, and communication interface communicate with each other via the communication bus. The memory is used to store at least one executable instruction that causes the processor to perform the operation of the vehicle control method as described in any one of claims 1-8.

11. A computer-readable storage medium, characterized in that, The storage medium stores at least one executable instruction, which, when executed on an electronic device, causes the electronic device to perform the operation of the vehicle control method as described in any one of claims 1-8.

Images