Method and system for estimating downforce using half-vehicle model of uncertainty estimator

Abstract

A method for downforce control includes receiving a plurality of vehicle inputs from a vehicle. The plurality of vehicle inputs includes sensor data from a plurality of sensors of the vehicle. The method further comprises the following steps: (a) determining the uncertainty of the predicted half-vehicle model by adopting a neural network; and (b) determining a front normal force at the front axle and a rear normal force at the rear axle using the vehicle input, the predicted half-vehicle model uncertainty, and the half-vehicle model. The method further includes determining a first position of the first aerodynamic body relative to the vehicle body and a second position of the second aerodynamic body relative to the vehicle body based on a front normal force at the front axle and a rear normal force at the rear axle, respectively.

Description

Technical Field

[0001] The present disclosure relates to methods and systems for estimating downforce using a half-car model with a data-driven uncertainty estimator. Background Art

[0002] The background technology generally introduces the background of the present disclosure. In the description of the background technology, the achievements of the inventors mentioned at present and the description contents that may not conform to the prior art at the time of application are neither explicitly nor implicitly regarded as the prior art for the present disclosure.

[0003] Downforce refers to the vertical load generated by the aerodynamic components of a vehicle during motion. Some vehicles include actuators to control downforce. Summary of the invention

[0004] The present disclosure relates to a method for downforce control. In one aspect of the present disclosure, the method includes: receiving a plurality of vehicle inputs from a vehicle. The vehicle inputs include sensor data from sensors of the vehicle. The vehicle inputs may include a longitudinal velocity of the vehicle, a lateral acceleration of the vehicle, and a longitudinal acceleration of the vehicle. The vehicle includes a vehicle body, a front axle coupled to the vehicle body, a rear axle coupled to the vehicle body, a first aerodynamic actuator coupled to the vehicle body, and a second aerodynamic actuator coupled to the vehicle body. The first aerodynamic actuator includes a first aerodynamic body that can move relative to the vehicle body. The second aerodynamic actuator includes a second aerodynamic body that can move relative to the vehicle body. The method also includes: determining a downforce acting on the vehicle. Determining the downforce acting on the vehicle includes: determining a predicted half-vehicle model uncertainty using a neural network; and determining a front normal force at the front axle and a rear normal force at the rear axle using the vehicle inputs, the predicted half-vehicle model uncertainty, and the half-vehicle model. The method also includes: determining a first position of the first aerodynamic body relative to the vehicle body and a second position of the second aerodynamic body relative to the vehicle body based on the front normal force at the front axle and the rear normal force at the rear axle, respectively. The method further includes commanding a first aerodynamic actuator to move the first aerodynamic body to a first position relative to the vehicle body. The method further includes commanding a second aerodynamic actuator to move the second aerodynamic body to a second position relative to the vehicle body.

[0005] In one aspect of the present disclosure, the half-car model includes the following equation:

[0006]

[0007]

[0008] in:

[0009] M is the mass of the vehicle;

[0010] g is the gravitational constant;

[0011] is the rate of change of the vehicle's heave rate;

[0012] is the spring-based front normal force at the front axle;

[0013] F r，uncertanties is the uncertain force of the car model at the rear axle;

[0014] is the spring-based rear normal force at the rear axle;

[0015] F f，uncertainties is the uncertain force of the half-car model at the front axle;

[0016] It is the front downforce;

[0017] It is the rear downforce;

[0018] I yy is the inertia of the vehicle;

[0019] is the rate of change of the pitch angle;

[0020] L f is the distance from the vehicle’s center of gravity to the front axle;

[0021] L r is the distance from the vehicle’s center of gravity to the rear axle;

[0022] is the distance from the center of gravity of the vehicle to the first aerodynamic body;

[0023] is the distance from the center of gravity of the vehicle to the second aerodynamic body;

[0024] h pc，cg is the road height; and

[0025] a x is the longitudinal acceleration of the vehicle.

[0026] In one aspect of the present disclosure, the half-car model further comprises the following equation:

[0027]

[0028]

[0029] in:

[0030] is the spring-based front normal force at the front axle;

[0031] is the spring-based rear normal force at the rear axle;

[0032] is the front preload;

[0033] is the rear preload;

[0034] K f is the spring constant of the suspension in the front axle;

[0035] K r is the spring constant of the suspension in the rear axle;

[0036] z f It’s the undulation at the front axle;

[0037] z r It’s the undulation at the rear axle;

[0038] C f is the front suspension damping rate;

[0039] C r is the rear suspension damping rate;

[0040] is the rate of heave at the front axle; and

[0041] is the rate of change of heave at the rear axle.

[0042] The method further includes: using a Kalman filter to determine the downforce acting on the vehicle. In addition, the method further includes: developing a neural network to determine the half-vehicle model uncertainty forces at the front axle of the vehicle and the half-vehicle model uncertainty forces at the rear axle. The method further includes: training the neural network model by simulating the vehicle in a wind tunnel and / or by testing the vehicle on a road.

[0043] The present application also describes a tangible, non-transitory machine-readable medium containing machine-readable instructions. When executed by a processor, the machine-readable instructions cause the processor to perform the above-described method.

[0044] The present disclosure also describes a vehicle. The vehicle includes a vehicle body, a front axle connected to the vehicle body, a rear axle connected to the vehicle body, a plurality of sensors disposed in the vehicle body, and a vehicle controller disposed in the vehicle body. The vehicle controller communicates with the sensors. The vehicle controller is programmed to perform the above method.

[0045] Other applicable fields of the present disclosure will become apparent from the detailed description provided below.It should be understood that the detailed description and specific examples are only for illustrative purposes and are not intended to limit the scope of the present disclosure.

[0046] The above features, advantages and other features and advantages of the disclosed systems and methods are apparent from the detailed description (including the claims) and exemplary embodiments taken in conjunction with the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] The present disclosure will be more fully understood based on the detailed description and accompanying drawings, in which:

[0048] Figure 1 is a side view of a vehicle including a system for downforce control; and

[0049] Figure 2 is a flow chart of a method for downforce control. DETAILED DESCRIPTION

[0050] Several examples of the present disclosure illustrated in the accompanying drawings will now be discussed in detail. The same or similar reference numerals are used throughout the drawings and description to refer to the same or similar parts or steps.

[0051] See also Figure 1, the vehicle 10 includes (or is in communication with) a system 11 for downforce control. Although the system 11 is shown inside the vehicle 10, it is contemplated that the system 11 may be located outside the vehicle 10. As a non-limiting example, the system 11 may be in wireless communication with the vehicle 10. Although the vehicle 10 is shown as a two-door sedan, it is contemplated that the vehicle 10 may be another type of vehicle, such as a pickup truck, a sedan, a sport utility vehicle (SUV), a recreational vehicle (RV), etc. The system 11 may be used to accurately estimate the downforce applied by the active aerodynamic actuator. In the system 11, a hybrid delta model based on a half-car + machine learning is developed to estimate downforce with high reliability in real-time operation. A hybrid delta model based on a half-car + machine learning is developed to estimate downforce with high reliability in real-time operation. The system 11 calculates the uncertainty of the half-car model under different operating conditions. The calculated uncertainty is then used as the ground truth for training the machine learning uncertainty predictor. Model informed rate limiting min / max logic is used to prevent the machine learning uncertainty predictor from providing incorrect downforce estimates. In this system 11, a Kalman filter is designed on a half-car model to estimate front and rear downforce using spring-based normal force measurements. Due to the simplicity of the half-car model, downforce estimates may be inaccurate, especially during violent maneuvers. Therefore, system 11 is used to capture model uncertainty to improve downforce estimates.

[0052] In addition, the vehicle 10 includes a vehicle controller 34 and one or more sensors 40 in communication with the vehicle controller 34. The sensor 40 collects information and generates sensor data indicative of the collected information. As non-limiting examples, the sensor 40 may include a global navigation satellite system (GNSS) transceiver or receiver, an inertial measurement unit (IMU), a yaw rate sensor, a chassis height sensor, a speed sensor, a lidar, a radar, an ultrasonic sensor, and a camera, etc. The GNSS transceiver or receiver is configured to detect the position of the vehicle 10 on the earth. The speed sensor is configured to detect the speed of the vehicle 10. The yaw rate sensor is configured to determine the heading of the vehicle 10. The camera may have a field of view large enough to capture images of the front, rear, and sides of the vehicle 10. The chassis height sensor is configured to measure the chassis height of the vehicle 10. The ultrasonic sensor can detect static and / or dynamic objects.

[0053] The vehicle controller 34 is programmed to receive sensor data from the sensor 40 and includes at least one processor 44 and a non-transitory computer-readable storage device or medium 46. The processor 44 can be a custom processor, a central processing unit (CPU), a graphics processing unit (GPU), an auxiliary processor of several processors associated with the vehicle controller 34, a semiconductor-based microprocessor (in the form of a microchip or chipset), a macroprocessor, or a combination thereof, or generally a device for executing instructions. The computer-readable storage device or medium 46 can include volatile and non-volatile memory, such as read-only memory (ROM), random access memory (RAM), and keep-on memory (KAM). KAM is a persistent or non-volatile memory that can be used to store various operating variables when the processor 44 is powered off. The computer readable storage device or medium of the vehicle controller 34 can be implemented using a number of storage devices, such as PROM (programmable read-only memory), EPROM (electrically programmable read-only memory), EEPROM (electrically erasable programmable read-only memory), flash memory, or other electrical, magnetic, optical or combination storage devices capable of storing data (some of which represent executable instructions used by the vehicle controller 34 when controlling the vehicle 10).

[0054] The instructions may include one or more separate programs, each of which includes a series of ordered executable instructions for implementing logical functions. When executed by the processor 44, the instructions receive and process signals from the camera, execute logic, calculations, methods and / or algorithms for automatically controlling components of the vehicle 10, and generate control signals to actuators (e.g., the first aerodynamic actuator and / or the second aerodynamic actuator 42) to automatically control components of the vehicle 10 based on the logic, calculations, methods and / or algorithms. Although Figure 1 1, a single vehicle controller 34 is shown, but the system 11 may include multiple controllers 34 that communicate via a suitable communication medium or combination of communication media and cooperate to process sensor signals, execute logic, calculations, methods and / or algorithms, and generate control signals to automatically control the features of the system 11. In various embodiments, one or more instructions of the vehicle controller 34 are included in the system 11. The non-transitory computer-readable storage device or medium 46 contains machine-readable instructions (e.g., such as Figure 2 ), when executed by one or more processors, these machine-readable instructions cause the processor 44 to perform method 100 ( Figure 2 ).

[0055] The vehicle 10 includes a body 12, a first or rear axle 14, and a second or front axle 16. The front axle 14 and the rear axle 16 are coupled to the body 12. In addition, each of the front axle 14 and the rear axle 16 is configured to rotate relative to the body 12. The vehicle 10 further includes one or more first or rear tires 18 coupled to the rear axle 14 and one or more second or front tires 20 coupled to the front axle 16.

[0056] The vehicle 10 includes a first or rear aerodynamic actuator 41 and a second or front aerodynamic actuator 42, each of which communicates with the vehicle controller 34. The first aerodynamic actuator 41 includes a first aerodynamic body 48, and the second aerodynamic actuator 42 includes a second aerodynamic body 50. Each of the first aerodynamic body 48 and the second aerodynamic body 50 can be configured as a spoiler. In the present disclosure, the term "wing" is defined as having the shape of a wing (i.e., a wing having an airfoil shape defined by a streamlined cross-sectional shape that generates lift for flight or propulsion through a fluid). The term "spoiler" refers to an aerodynamic device that is capable of disrupting the movement of air passing through the vehicle 10 when the vehicle 10 is moving, thereby reducing drag and / or causing aerodynamic downforce on the vehicle 10. The term "downforce" refers to a component of force perpendicular to the relative direction of motion of the vehicle 10 (i.e., along the longitudinal direction, toward the road surface 13). For example, a spoiler can diffuse air by increasing the amount of turbulence flowing through it. The first aerodynamic actuator 41 is closer to the rear axle 14 than the front axle 16 to control a rear downforce 43 at or near the rear axle 14. The second aerodynamic actuator 42 is closer to the front axle 16 than the rear axle 14 to control a front downforce 45 at or near the front axle 16. The rear downforce 43 and the front downforce 45 may be determined using sensor data from a sensor 40 (e.g., a ride height sensor).

[0057] The first aerodynamic actuator 41 includes a bracket 52 that is directly coupled to the vehicle body 12 and one or more first pivots 54 (e.g., pivot pins, pivot mechanisms, etc.) that pivotally couple the first aerodynamic body 48 to the vehicle body 12. Thus, the first aerodynamic body 48 is movable (e.g., pivotable) relative to the vehicle body 12. The first aerodynamic actuator 41 includes a first motor 56 (or another suitable machine) that communicates with the vehicle controller 34 and is coupled to the first aerodynamic body 48 via the first pivot 54. Thus, the vehicle controller 34 is programmed to actuate the first motor 56 to move the first aerodynamic body 48 relative to the vehicle body 12.

[0058] The second aerodynamic actuator 42 is coupled to the vehicle body 12 and includes one or more second pivots 58 (e.g., pivot pins, pivot mechanisms, etc.) that pivotally couple the second aerodynamic body 50 to the vehicle body 12. Thus, the second aerodynamic body 50 is movable (e.g., pivotable) relative to the vehicle body 12. The second aerodynamic actuator 42 includes a second motor 60 (or another suitable machine) that communicates with the vehicle controller 34 and is coupled to the second aerodynamic body 50 via the second pivot 58. Thus, the vehicle controller 34 is programmed to actuate the second motor 60 to move the second aerodynamic body 50 relative to the vehicle body 12.

[0059] The vehicle 10 includes a user interface 23 that communicates with a vehicle controller 34. The user interface 23 can be, for example, a touch screen in a dashboard, and can include, but is not limited to, an alarm (such as one or more speakers that provide audible sounds, tactile feedback in a vehicle seat or other object, one or more displays, one or more microphones, one or more lights, and / or other devices suitable for providing notifications). The user interface 23 is in electronic communication with the vehicle controller 34 and is configured to receive input from a hearing-impaired vehicle occupant 25 (e.g., a vehicle user or a vehicle passenger). For example, the user interface 23 can include a touch screen and / or buttons that are configured to receive input from a vehicle occupant. Thus, the vehicle controller 34 is configured to receive input from a vehicle occupant via the user interface 23, and to provide output (e.g., auditory, tactile, and / or visual notifications) to the vehicle occupant.

[0060] Figure 2 1 is a flow chart of an active downforce control method 100 for downforce control. The method 100 begins at box 102. In box 102, the vehicle controller 34 receives vehicle inputs (such as sensor data from the sensor 40), driver inputs, stability triggers, and vehicle feedback. As non-limiting examples, the vehicle controller 34 may receive vehicle longitudinal acceleration, vehicle lateral acceleration, vehicle longitudinal speed, chassis height data, front vehicle angle, brake pedal pressure or position, throttle position, estimated friction coefficient between the road surface 13 and at least one of the first tire 18 or the second tire 20, steering wheel angle, normal force, downforce, tire trail, understeer angle, slip slope, and wheel ground speed. After the vehicle controller 34 receives the inputs, the method 100 continues to box 104.

[0061] In block 104, a half-car model is developed that incorporates pitch and heave degrees of freedom and an integrated suspension model. The equations of motion for the half-car model can be written in terms of heave and pitch motion. As a non-limiting example, the state space model of the half-car can be represented by the following equation:

[0062]

[0063]

[0064] in:

[0065] M is the mass of the vehicle 10;

[0066] g is the gravitational constant;

[0067] is the rate of change of the heave rate of change of the vehicle 10;

[0068] is the spring-based front normal force at the front axle 16;

[0069] F r，uncertainties is the uncertain force of the vehicle model at the rear axle 14;

[0070] is the spring-based rear normal force at the rear axle 14;

[0071] F f，uncertainties is the uncertain force of the half-car model at the front axle 16;

[0072] It is the front downforce;

[0073] It is the rear downforce;

[0074] I yy is the inertia of the vehicle 10;

[0075] is the rate of change of the pitch angle;

[0076] L f is the distance from the center of gravity of the vehicle 10 to the front axle;

[0077] L r is the distance from the center of gravity of the vehicle 10 to the rear axle;

[0078] is the distance from the center of gravity of the vehicle to the first aerodynamic body 48;

[0079] is the distance from the center of gravity of the vehicle to the second aerodynamic body 50;

[0080] h pc，cg is the road height; and

[0081] a x is the longitudinal acceleration of the vehicle 10 .

[0082] The half-car model further includes the following equations:

[0083]

[0084]

[0085] in:

[0086] is the spring-based front normal force at the front axle 16;

[0087] is the spring-based rear normal force at the rear axle 14;

[0088] is the front preload;

[0089] is the rear preload;

[0090] K f is the spring constant of the suspension in the front axle 16;

[0091] K r is the spring constant of the suspension in the rear axle 14;

[0092] z f It is the undulation at the front axle 16;

[0093] z r It is the undulation at the rear axle 14;

[0094] C f is the front suspension damping rate;

[0095] C r is the rear suspension damping rate;

[0096] is the rate of change of heave at the front axle 16; and

[0097] is the rate of change of the heave at the rear axle 14 .

[0098] After the half-car model is developed, the method 100 enters box 106. In box 106, a neural network is developed to determine the half-car model uncertainty forces at the front axle 16 and the rear axle 14. To this end, data needs to be collected to train the neural network. The training can be performed by testing the vehicle 10 on the road and / or by simulating the vehicle 10 in a wind tunnel. In order to estimate the uncertainty of the half-car model, several structures (e.g., input, number of layers, number of neurons, activation function) are trained and evaluated. As a non-limiting example, a shallow neural network can be developed. The shallow neural network may include 3 layers. One layer may include 15 neurons, another layer may include 10 neurons, and the third layer may include 2 neurons. The input of the neural network may include longitudinal velocity, longitudinal acceleration, lateral acceleration, the heave of the vehicle 10 at the front axle 16, the heave of the vehicle 10 at the rear axle 14, the pitch at the front axle 16, and the pitch at the rear axle 14. The output of the neural network is the half-car model uncertainty forces at the front axle 16 and the rear axle 14. The selected neural structure can then be used to train the neural network and make any necessary adjustments to optimize the structure. In order to train the uncertainty neural network, network input data and true value data (uncertainty load) are required. In order to generate the input (i.e., true value data), several simulations were performed using different driving scenarios. Since the magnitude of the downforce in these driving scenarios is known, a Kalman filter can be used to estimate the unmodeled load for a given input. The estimated unmodeled load can be used as the true value for a given input. An informed rate-limited min / max logic is used to prevent the machine learning uncertainty predictor from providing incorrect downforce estimates. The inputs to the informed rate-limited min / max logic can be chassis height measurements, longitudinal speeds, and the positions of the first aerodynamic actuator 41 and the second or front aerodynamic actuator 42. In the informed rate-limited min / max logic, the downforce at the front axle 16 (i.e., front downforce) and the downforce at the rear axle 14 (i.e., rear downforce) are determined using an aerodynamic map pattern. The rate of change of the front downforce and the rear downforce is calculated. Then, the maximum and minimum front and rear downforces and the maximum and minimum change rates of the front and rear downforces are calculated using the magnitudes of the front and rear downforces and their change rates. Then, the method 100 proceeds to frame 108 .

[0099] In box 108, the vehicle controller 34 predicts half-vehicle model uncertainties at the front axle 16 and the rear axle 16 using measurements and estimates from the vehicle 10 (e.g., inputs from the vehicle 10, such as sensor data) by employing a neural network. The predicted half-vehicle model uncertainties are then input into the half-vehicle model for Kalman filter downforce estimation. The estimated downforce is then passed through rate limiting and min / max logic as described above. The final estimated downforce is then used for downforce control. The method 100 then proceeds to box 110.

[0100] In block 110, the vehicle controller 34 determines the positions of the first aerodynamic body 48 and the second aerodynamic body 50 relative to the vehicle body 12 based on the final estimated downforce. The vehicle controller 34 then commands the first aerodynamic actuator 41 to move the first aerodynamic body 48 to the determined position relative to the vehicle body 12. In addition, the vehicle controller 34 commands the second aerodynamic actuator 42 to move the second aerodynamic body 50 to the determined position relative to the vehicle body 12.

[0101] Although exemplary embodiments are described above, this does not mean that these embodiments describe all possible forms included in the claims. The words used in this specification are descriptive words, not restrictive words, and it should be understood that various changes can be made without departing from the spirit and scope of the present disclosure. As described above, the features of various embodiments can be combined to form other embodiments that are not explicitly described or shown in the system and method of the present disclosure. Although various embodiments may have been described as providing advantages or being superior to other embodiments or prior art implementations in terms of one or more required characteristics, it should be recognized by those of ordinary skill in the art that one or more features or characteristics can be compromised to achieve the desired overall system properties, depending on the specific application and implementation. These properties may include, but are not limited to, cost, strength, durability, life cycle cost, marketability, appearance, packaging, size, applicability, weight, manufacturability, ease of assembly, etc. Therefore, embodiments described as being less ideal than other embodiments or prior art implementations in terms of one or more characteristics are not outside the scope of the present disclosure and may be ideal for specific applications.

[0102] The drawings are in simplified form and are not to exact scale. Directional terms (e.g., top, bottom, left, right, upward, directly above, above, below, directly below, rear side, and front side) may be used in the drawings for convenience and clarity only. These and similar directional terms should not be construed in any way as limiting the scope of the present invention.

[0103] Embodiments of the present disclosure are described herein. However, it should be understood that the disclosed embodiments are merely examples, and other embodiments may take various alternative forms. The drawings are not necessarily drawn to scale; some features may be enlarged or reduced to show the details of a particular component. Therefore, the specific structural and functional details disclosed herein should not be interpreted as being restrictive, but should only be interpreted as being a representative basis for teaching those skilled in the art to apply the systems and methods of the present disclosure in different ways. As will be understood by those of ordinary skill in the art, the various features shown and described in conjunction with any of the drawings may be combined with the features shown in one or more other drawings to produce embodiments that are not explicitly shown or described. The combination of features shown provides representative embodiments for typical applications. However, for specific applications or implementations, various combinations and modifications of features consistent with the teachings of the present disclosure may be required.

[0104] Embodiments of the present disclosure may be described herein according to functions and / or logic block components and various processing steps. It should be understood that these block components may be implemented by several hardware, software and / or firmware components configured to perform specified functions. For example, embodiments of the present disclosure may employ various integrated circuit components (e.g., storage elements, digital signal processing elements, logic elements, or lookup tables, etc.), which may perform various functions under the control of one or more microprocessors or other control devices. In addition, it will be appreciated by those skilled in the art that embodiments of the present disclosure may be practiced in conjunction with several systems, and the systems described herein are merely exemplary embodiments of the present disclosure.

[0105] For the sake of brevity, the present invention may not describe in detail the techniques related to other functional aspects of signal processing, data fusion, signaling, control and systems (and the various operating components of the systems). In addition, the lines shown in the various figures included herein are intended to represent exemplary functional relationships and / or physical connections between various elements. It should be noted that in an embodiment of the present disclosure, there may be alternative or additional functional relationships or physical connections.

[0106] This description is merely illustrative in nature and is in no way intended to limit the present disclosure, its application or use. The broad teachings of the present disclosure can be implemented in a variety of forms. Therefore, although the present disclosure includes specific examples, the true scope of the present disclosure should not be so limited, as other modifications will become apparent after studying the drawings, this specification and the appended claims.

Claims

1. A method for downforce control, comprising: receiving a plurality of vehicle inputs from a vehicle, wherein the plurality of vehicle inputs comprises sensor data from a plurality of sensors of the vehicle, the vehicle comprising a body, a front axle coupled to the body, a rear axle coupled to the body, a first aerodynamic actuator coupled to the body, and a second aerodynamic actuator coupled to the body, the first aerodynamic actuator comprising a first aerodynamic body movable relative to the body, and the second aerodynamic actuator comprising a second aerodynamic body movable relative to the body; Determining a downforce acting on the vehicle, wherein determining the downforce acting on the vehicle comprises: Using neural networks, the uncertainty of the predicted half-car model is determined; determining a front normal force at the front axle and a rear normal force at the rear axle using the vehicle input, the predicted half-vehicle model uncertainty, and the half-vehicle model; determining a first position of the first aerodynamic body relative to the vehicle body and a second position of the second aerodynamic body relative to the vehicle body based on a downforce acting on the vehicle; commanding the first aerodynamic actuator to move the first aerodynamic body to a first position relative to the vehicle body; and The second aerodynamic actuator is commanded to move the second aerodynamic body to a second position relative to the vehicle body.

2. The method of claim 1, wherein the half-car model comprises the following equation: in: M is the mass of the vehicle; g is the gravitational constant; is the rate of change of the heave rate of the vehicle; is the spring-based front normal force at the front axle; F r，uncertanties is the vehicle model uncertainty force at the rear axle; is the spring-based rear normal force at the rear axle; F f，uncertainties is the uncertain force of the half-car model at the front axle; It is the front downforce; It is the rear downforce; I yy is the inertia of the vehicle; is the rate of change of the pitch angle change rate; L f is the distance from the center of gravity of the vehicle to the front axle; L r is the distance from the center of gravity of the vehicle to the rear axle; is the distance from the center of gravity of the vehicle to the first aerodynamic body; is the distance from the center of gravity of the vehicle to the second aerodynamic body; h pc，cg is the road height; and a x is the longitudinal acceleration of the vehicle.

3. The method according to claim 2, wherein the half-car model further adopts the following equation: in: is the spring-based front normal force at the front axle; is the spring-based rear normal force at the rear axle; is the front preload; is the rear preload; K f is the spring constant of the suspension in the front axle; K r is the spring constant of the suspension in the rear axle; z f is the undulation at said front axle; z r is the undulation at the rear axle; C f is the front suspension damping rate; C r is the rear suspension damping rate; is the rate of change of heave at the front axle; and is the rate of change of heave at the rear axle.

4. The method according to claim 3, further comprising: A Kalman filter is used to determine downforce acting on the vehicle.

5. The method according to claim 4, further comprising: The neural network is developed to determine the half-vehicle model uncertainty forces at the front axle and the half-vehicle model uncertainty forces at the rear axle of the vehicle.

6. The method according to claim 4, further comprising: The neural network model is trained by simulating the vehicle in a wind tunnel.

7. The method according to claim 5, further comprising: The neural network model is trained by testing the vehicle on a road. 8 . The method of claim 6 , wherein the vehicle inputs include a longitudinal velocity of the vehicle, a lateral acceleration of the vehicle, and a longitudinal acceleration of the vehicle.

9. A tangible, non-transitory machine-readable medium comprising machine-readable instructions that, when executed by a processor, cause the processor to: receiving a plurality of vehicle inputs from a vehicle, wherein the plurality of vehicle inputs comprises sensor data from a plurality of sensors of the vehicle, the vehicle comprising a body, a front axle coupled to the body, a rear axle coupled to the body, a first aerodynamic actuator coupled to the body, and a second aerodynamic actuator coupled to the body, the first aerodynamic actuator comprising a first aerodynamic body movable relative to the body, and the second aerodynamic actuator comprising a second aerodynamic body movable relative to the body; Determining a downforce acting on the vehicle, wherein determining the downforce acting on the vehicle comprises: Using neural networks, the uncertainty of the predicted half-car model is determined; determining a front normal force at the front axle and a rear normal force at the rear axle using the vehicle input, the predicted half-vehicle model uncertainty, and the half-vehicle model; determining a first position of the first aerodynamic body relative to the vehicle body and a second position of the second aerodynamic body relative to the vehicle body based on a downforce acting on the vehicle; commanding the first aerodynamic actuator to move the first aerodynamic body to a first position relative to the vehicle body; and The second aerodynamic actuator is commanded to move the second aerodynamic body to a second position relative to the vehicle body.

10. The tangible, non-transitory machine-readable medium of claim 9, wherein the half-car model comprises the following equation: in: M is the mass of the vehicle; g is the gravitational constant; is the rate of change of the heave rate of the vehicle; is the spring-based front normal force at the front axle; F r，uncertainties is the vehicle model uncertainty force at the rear axle; is the spring-based rear normal force at the rear axle; F f，uncertainties is the uncertain force of the half-car model at the front axle; It is the front downforce; It is the rear downforce; I yy is the inertia of the vehicle; is the rate of change of the pitch angle; L f is the distance from the center of gravity of the vehicle to the front axle; L r is the distance from the center of gravity of the vehicle to the rear axle; is the distance from the center of gravity of the vehicle to the first aerodynamic body; is the distance from the center of gravity of the vehicle to the second aerodynamic body; h pc，cg is the road height; and a x is the longitudinal acceleration of the vehicle.

Images