Vehicle state estimation method, device, apparatus and storage medium

Abstract

The application relates to the automobile technical field and discloses a vehicle state estimation method, device, equipment and storage medium. The method comprises the following steps: acquiring a vehicle state parameter of a vehicle in a current working condition; constructing a vehicle dynamics model according to the vehicle state parameter; correcting a nonlinear dynamics model in the horizontal dynamics according to tire force and a side slip angle of the vehicle to obtain a corrected vehicle dynamics model; and performing adaptive Kalman filtering on a target state variable in the corrected vehicle dynamics model, and updating process noise information and measurement noise information in the process of adaptive Kalman filtering by using a regularization method to obtain estimation information of the target state variable at a current moment. The application can improve the vehicle state evaluation precision and inhibit filter divergence in the vehicle state process.

Description

Technical Field

[0001] This application relates to the field of automotive technology, and in particular to a vehicle state estimation method, apparatus, device, and storage medium. Background Technology

[0002] Kalman filtering is an optimal estimation method based on the state equations of a linear system, offering advantages such as high estimation accuracy and good real-time performance. In the automotive field, Kalman filtering has been widely applied to sensor data fusion and vehicle state estimation due to its relatively balanced accuracy and computational complexity.

[0003] However, in related technologies, vehicle state estimation methods based on Kalman filtering neglect the coupling effect between roll motion and the suspension system, leading to a significant decrease in estimation accuracy under complex conditions (such as emergency obstacle avoidance and high-speed cornering). Furthermore, the performance of vehicle state estimation methods based on Kalman filtering depends on the selection of process noise and measurement noise parameters. The fixed process noise and measurement noise information in Kalman filtering are difficult to adapt to dynamically changing driving environments, easily causing filter divergence problems.

[0004] This contradicts the uncertainties of the road environment and the errors in sensor measurement parameters during actual vehicle movement. Moreover, the Kalman filter observer with steady parameters requires repeated offline parameter adjustments, making it difficult to guarantee the robustness of state estimation. Summary of the Invention

[0005] The purpose of this application is to provide a vehicle state estimation method, apparatus, device, and storage medium, which aims to improve the accuracy of vehicle state assessment and suppress filter divergence in the vehicle state process.

[0006] This application provides a vehicle state estimation method, including:

[0007] Obtain the vehicle status parameters under the current operating conditions;

[0008] A vehicle dynamics model is constructed based on the vehicle state parameters; the vehicle dynamics model is a nonlinear dynamics model that describes the force balance relationship of the vehicle in the lateral, yaw and roll directions.

[0009] The nonlinear dynamics model is modified by lateral dynamics based on the tire force and sideslip angle of the vehicle to obtain the modified vehicle dynamics model.

[0010] An adaptive Kalman filter is applied to the target state variables in the modified vehicle dynamics model to obtain the estimated information of the target state variables at the current moment. The target state variables include the vehicle's center of gravity sideslip angle, yaw rate, roll rate, and roll angle. During the adaptive Kalman filtering process, a regularization method is used to update the process noise information and measurement noise information.

[0011] In some embodiments, the expression for the vehicle dynamics model is:

[0012] ,

[0013] ,

[0014] ,

[0015] in, Let m be the resultant force of the external forces acting on the vehicle along the Y-axis, and m be the mass of the vehicle. For the vehicle's lateral acceleration, Let be the longitudinal speed of the vehicle. Let yaw rate be the vehicle's angular velocity. For the sprung mass, This is the distance from the vehicle's roll center of gravity to the roll axis. The roll angle acceleration of the vehicle. The lateral force is from the front wheel. The lateral force is from the rear wheel. Let be the torque of the external force acting on the vehicle about the Z-axis. Let be the moment of inertia of the entire vehicle about its center of mass in the vertical direction. For the vehicle's yaw acceleration, This is the distance from the vehicle's center of gravity to the front axle. This is the distance from the vehicle's center of gravity to the rear axle. Let x be the torque of the external force acting on the vehicle about the X-axis. Let be the longitudinal rotational inertia of the entire vehicle about its center of mass. It is the gravitational constant. For equivalent roll stiffness, The vehicle's roll angle. For equivalent roll damping, This represents the vehicle's roll rate.

[0016] In some embodiments, the lateral dynamics correction of the nonlinear dynamics model based on the tire forces and sideslip angle of the vehicle includes:

[0017] The mechanical properties of the tires are fitted using a magic tire model to determine the tire forces of the vehicle.

[0018] The tire slip angle of the vehicle is calculated based on the vehicle's longitudinal speed, lateral speed, yaw rate, and front wheel steering angle.

[0019] The mapping relationship between the tire toe angle, tire camber angle, tire swerve angle and tire lateral force of the vehicle is determined based on the tire force and tire slip angle of the vehicle. The vehicle dynamics model is then modified by lateral dynamics based on the mapping relationship to obtain the modified vehicle dynamics model.

[0020] In some embodiments, the expression for the modified vehicle dynamics model is:

[0021] ,

[0022] ,

[0023] ,

[0024] Where m is the total vehicle mass. For the vehicle's lateral acceleration, Let be the longitudinal speed of the vehicle. Let yaw rate be the vehicle's angular velocity. For the sprung mass, This is the distance from the vehicle's roll center of gravity to the roll axis. The roll angle acceleration of the vehicle. The lateral force is from the front wheel. The lateral force is from the rear wheel. Let be the moment of inertia of the entire vehicle about its center of mass in the vertical direction. For the vehicle's yaw acceleration, This is the distance from the vehicle's center of gravity to the front axle. This is the distance from the vehicle's center of gravity to the rear axle. Let be the longitudinal rotational inertia of the entire vehicle about its center of mass. It is the gravitational constant. For equivalent roll stiffness, The vehicle's roll angle. For equivalent roll damping, This represents the vehicle's roll rate.

[0025] In some embodiments, performing adaptive Kalman filtering on the target state variables in the modified vehicle dynamics model includes:

[0026] Based on the estimated information of the target state variables at the previous moment and the front wheel steering angle of the vehicle, the prior estimated information of the target state variables at the current moment is determined;

[0027] The prior estimation error covariance information at the current time is determined based on the posterior estimation error covariance information and process noise information at the previous time, and the Kalman gain at the current time is determined based on the prior estimation error covariance information at the current time.

[0028] The current information sequence information is determined based on the measurement information of the target measurement variables at the current moment and the prior estimation information of the target state variables at the current moment; the target measurement variables include the yaw rate and lateral state variables of the vehicle.

[0029] The estimation information of the target state variable at the current moment is determined based on the prior estimation information, Kalman gain, and innovation sequence information of the target state variable at the current moment;

[0030] After obtaining the estimated information of the target state variables at the current moment, the process noise information and measurement noise information at the current moment are updated using a regularization method.

[0031] In some embodiments, a method for updating the process noise information and the measurement noise information using a regularization method includes:

[0032] Based on the regularization theory of uncertainty problems, a regularization parameter is introduced to describe the process noise information at the current moment;

[0033] A generalized cross-validation method is used to determine the first target regularization parameter; the first target regularization parameter is a regularization parameter that minimizes the process noise information at the current time.

[0034] Introducing factors that affect both the car body tilt angle and the vehicle speed, the first objective regularization parameter is updated based on the influencing factors to obtain the second objective regularization parameter;

[0035] The process noise information at the current moment is updated according to the second target regularization parameter to obtain the updated process noise information at the current moment, and the measurement noise information at the current moment is determined according to the updated process noise information at the previous moment.

[0036] In some embodiments, the expression describing the process noise information at the current moment using the regularization parameter is as follows:

[0037] ,

[0038] in, This provides the process noise information at the current moment. For regularization parameters, The Kalman gain at the current moment. for The singular values ​​of , where T denotes the transpose, i is a positive integer, i∈[1,n], and n is . The total number of singular values, for The left singular vector, for The right singular vector, For the innovation covariance information at the current moment;

[0039] The expression for determining the first target regularization parameter using the generalized cross-validation method is as follows:

[0040] ,

[0041] in, For the function output value, for The minimum value, The first objective is the regularization parameter. The operation for calculating the trace of a matrix;

[0042] The expression for obtaining the second objective regularization parameter is:

[0043] ,

[0044] in, For the second objective regularization parameter, This is a reference stable value for the carriage tilt angle. This is the longitudinal velocity reference stable value. The vehicle's roll angle. The longitudinal speed of the vehicle;

[0045] The expression for obtaining the updated process noise information at the current moment is:

[0046] ,

[0047] in, This is the updated process noise information at the current moment;

[0048] The expression for obtaining the measurement noise information at the current moment is:

[0049] ,

[0050] ,

[0051] in, This is the measurement noise information at the current moment. For the current observation information, This provides the prior estimate of the error covariance information at the current moment. This provides the state transition information for the current moment.

[0052] This application embodiment also provides a vehicle state estimation device, including:

[0053] The first module is used to obtain the vehicle status parameters under the current operating conditions.

[0054] The second module is used to construct a vehicle dynamics model based on the vehicle state parameters; the vehicle dynamics model is a nonlinear dynamics model that describes the force balance relationship of the vehicle in the lateral, yaw and roll directions.

[0055] The third module is used to perform lateral dynamic correction on the nonlinear dynamic model based on the tire force and sideslip angle of the vehicle, so as to obtain the corrected vehicle dynamic model.

[0056] The fourth module is used to perform adaptive Kalman filtering on the target state variables in the modified vehicle dynamics model to obtain the estimated information of the target state variables at the current moment. The target state variables include the vehicle's center of gravity sideslip angle, yaw rate, roll rate, and roll angle. In the process of adaptive Kalman filtering, a regularization method is used to update the process noise information and measurement noise information.

[0057] This application also provides an electronic device, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the above-described vehicle state estimation method.

[0058] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the vehicle state estimation method described above.

[0059] The beneficial effects of this application are as follows: A vehicle dynamics model is constructed based on the obtained vehicle state parameters. This vehicle dynamics model is a nonlinear dynamics model describing the force balance relationship of the vehicle in lateral, yaw, and roll conditions. Lateral dynamics correction is then performed on the nonlinear dynamics model based on the vehicle's tire forces and sideslip angle to obtain a corrected vehicle dynamics model. Adaptive Kalman filtering is applied to the target state variables in the corrected vehicle dynamics model. During the adaptive Kalman filtering process, a regularization method is used to update process noise and measurement noise information to obtain an estimate of the target state variables at the current moment. By constructing a vehicle dynamics model describing the force balance relationship of the vehicle in lateral, yaw, and roll conditions, the modeling accuracy of the vehicle's motion state under complex working conditions is significantly improved. The use of a regularization method to update process noise and measurement noise information during the adaptive Kalman filtering process overcomes the limitations of traditional fixed parameters, enabling the algorithm to adapt to the dynamic changes in road conditions and sensor errors in real time, effectively suppressing filter divergence, improving the accuracy of vehicle state assessment, and suppressing filter divergence during the vehicle state process. Attached Figure Description

[0060] Figure 1 This is a diagram illustrating the application environment of the vehicle state estimation method provided in the embodiments of this application.

[0061] Figure 2 This is a flowchart of the vehicle state estimation method provided in the embodiments of this application.

[0062] Figure 3 This is a schematic diagram of the vehicle dynamics model provided in the embodiments of this application.

[0063] Figure 4 A schematic diagram of the modified vehicle dynamics model provided in the embodiments of this application.

[0064] Figure 5 This is a schematic diagram of the vehicle state estimation device provided in the embodiments of this application.

[0065] Figure 6 This is a schematic diagram of the hardware structure of the electronic device provided in the embodiments of this application. Detailed Implementation

[0066] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0067] It should be noted that although functional modules are divided in the device schematic diagram and a logical order is shown in the flowchart, in some cases, the steps shown may be performed in a different order than the module division in the device or the order in the flowchart. The terms "first," "second," etc., in the specification, claims, and drawings are used to distinguish similar objects and are not used to describe a specific order or sequence.

[0068] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit this application.

[0069] The information, data, and signals involved in the embodiments of this application are all authorized by the relevant parties or fully authorized by all parties, and the collection, use, and processing of the relevant data comply with the relevant laws, regulations, and standards of the relevant countries and regions.

[0070] Figure 1 This diagram illustrates the application environment of the vehicle state estimation method provided in the embodiments of this application. (See also...) Figure 1This vehicle state estimation method is applied to a vehicle state estimation system. The system includes a terminal 110 and a server 120. The terminal 110 and server 120 are connected via a network. The terminal 110 can be a desktop terminal, a mobile terminal, or an in-vehicle terminal; a mobile terminal can be at least one of a mobile phone, tablet, or laptop. The server 120 can be a standalone server or a server cluster consisting of multiple servers. The terminal 110 sends the vehicle state parameters under the current operating condition to the server 120. The server 120 obtains the vehicle state parameters under the current operating condition, constructs a vehicle dynamics model based on these parameters, performs lateral dynamics correction on the nonlinear dynamics model based on the vehicle's tire forces and sideslip angle, obtaining a corrected vehicle dynamics model, and applies adaptive Kalman filtering to the target state variables in the corrected model to obtain an estimate of the target state variables at the current moment. The vehicle dynamics model is a nonlinear dynamic model that describes the force balance relationship of the vehicle in the lateral, yaw and roll directions. The target state variables include the vehicle's center of gravity sideslip angle, yaw rate, roll rate and roll angle. In the adaptive Kalman filtering process, a regularization method is used to update the process noise information and measurement noise information.

[0071] It should be understood that Figure 1 The application scenarios shown are merely examples. In practical applications, the vehicle state estimation method provided in this application embodiment can also be applied to other scenarios. For example, the above-mentioned vehicle state estimation method can be directly applied to terminal 110. Terminal 110 is used to obtain vehicle state parameters of the vehicle under the current operating condition, construct a vehicle dynamics model based on the vehicle state parameters, perform lateral dynamics correction on the nonlinear dynamics model based on the vehicle's tire force and sideslip angle, and obtain a corrected vehicle dynamics model. Adaptive Kalman filtering is then applied to the target state variables in the corrected vehicle dynamics model to obtain the estimated information of the target state variables at the current moment. The vehicle dynamics model is a nonlinear dynamics model describing the force balance relationship of the vehicle in the lateral, yaw, and roll directions. The target state variables include the vehicle's center of gravity sideslip angle, yaw rate, roll rate, and roll angle. During the adaptive Kalman filtering process, a regularization method is used to update process noise information and measurement noise information.

[0072] In the embodiments of this application, the vehicle state estimation method is executed by a terminal or a server.

[0073] Figure 2 This is a flowchart of the vehicle state estimation method provided in an embodiment of this application. (See attached document.) Figure 2 In some embodiments, the method includes, but is not limited to, steps S201 to S204.

[0074] Step S201: Obtain the vehicle status parameters under the current operating conditions.

[0075] Vehicle state parameters are key physical quantities used to describe the force relationships of a vehicle in the lateral, yaw, and roll directions. Vehicle state parameters can include vehicle attitude parameters (such as yaw angle, yaw rate, pitch angle, and roll angle), tire and powertrain parameters (such as tire slip ratio, tire slip angle, and drive torque / braking torque), and advanced state parameters (such as center of gravity slip angle, curvature of motion, and relative state with obstacles).

[0076] Obtaining vehicle state parameters under current operating conditions can be achieved by acquiring sensor signals from various vehicle sensors at the current moment. For example, an IMU (Inertial Measurement Unit) can be used to acquire the vehicle's acceleration information at the current moment. The acceleration signal indicates the vehicle's longitudinal acceleration, and the acceleration vibration amplitude is obtained from the acceleration signal. Sensor signals also include the brake pedal opening, steering wheel angle, and wheel speed signals at the current moment.

[0077] Step S202: Construct a vehicle dynamics model based on vehicle state parameters.

[0078] The vehicle dynamics model is a nonlinear dynamic model that describes the force balance relationship of a vehicle in lateral, yaw, and roll. It can be understood that the vehicle dynamics model is a three-degree-of-freedom coupled model that includes lateral force balance, yaw moment balance, and roll moment balance. Specifically, it can be achieved by establishing the kinematic equations for sprung and unsprung masses. This model can simultaneously reflect the interaction relationship between the vehicle's lateral acceleration, yaw motion, and roll.

[0079] In some embodiments, based on the 2-DOF (lateral and yaw) vehicle dynamics model, a roll degree of freedom is added to establish a 3-DOF vehicle dynamics model considering lateral, yaw, and roll. (See [reference]) Figure 3 For the vehicle dynamics model with lateral and yaw degrees of freedom, please refer to [reference needed]. Figure 3 (a) Vehicle dynamics model with roll degree of freedom, see [reference] Figure 3 (b); The vehicle dynamics model is based on the following assumptions: 1. The influence of the steering system is ignored, and the front wheel angle is directly used as the input, and the front wheel angle is relatively small; 2. The vehicle's forward speed along the longitudinal direction is considered constant; 3. The effects of aerodynamics are not considered, and the influence of load transfer is ignored.

[0080] The expression for the vehicle dynamics model is:

[0081] ,

[0082] ,

[0083] ,

[0084] in, Let m be the resultant force of the external forces acting on the vehicle along the y-axis, and m be the mass of the vehicle. For the vehicle's lateral acceleration, Let be the longitudinal speed of the vehicle. Let yaw rate be the vehicle's angular velocity. For the sprung mass, This is the distance from the vehicle's roll center of gravity to the roll axis. The roll angle acceleration of the vehicle. The lateral force is from the front wheel. The lateral force is from the rear wheel. Let be the torque of the external force acting on the vehicle about the Z-axis. Let be the moment of inertia of the entire vehicle about its center of mass in the vertical direction. For the vehicle's yaw acceleration, This is the distance from the vehicle's center of gravity to the front axle. This is the distance from the vehicle's center of gravity to the rear axle. Let x be the torque of the external force acting on the vehicle about the X-axis. Let be the longitudinal rotational inertia of the entire vehicle about its center of mass. It is the gravitational constant. For equivalent roll stiffness, The vehicle's roll angle. For equivalent roll damping, This represents the vehicle's roll rate.

[0085] The aforementioned vehicle dynamics model uses three nonlinear equations to describe the force balance relationships of the vehicle in the lateral, yaw, and roll directions. The lateral equation is based on the balance between tire lateral force and lateral acceleration, where the vehicle mass and sprung mass are separated to more accurately reflect the motion differences between sprung and unsprung mass. The yaw equation describes the dynamic characteristics of yaw angular acceleration by combining the contributions of the front and rear wheel lateral forces to the yaw moment with the moment of inertia about the vertical axis. The roll equation introduces equivalent roll stiffness and damping parameters, combining the roll motion of the sprung mass with the effect of gravity to characterize the coupling effect between the suspension system and the vehicle body roll. By solving these three equations simultaneously, the motion states in the lateral, yaw, and roll directions can be obtained simultaneously, thus avoiding neglecting the mutual influence between multiple degrees of freedom.

[0086] Step S203: Based on the tire force and sideslip angle of the vehicle, perform lateral dynamics correction on the nonlinear dynamics model to obtain the corrected vehicle dynamics model.

[0087] Lateral dynamics correction refers to error compensation of the model based on the tire mechanical characteristics and steering geometry. Specifically, it can be achieved by fitting the nonlinear lateral force using the magic formula tire model and combining it with the steering system kinematics to calculate the slip angle. This correction can eliminate model deviations caused by tire nonlinear characteristics.

[0088] Lateral dynamic corrections are made to the nonlinear dynamics model based on the vehicle's tire forces and slip angles. Specifically, after determining the vehicle's tire forces and slip angles, the mapping relationship between the changes in tire toe angle and camber angle caused by the vehicle's roll steering, roll camber, deformable steering, and deformable camber, and the wheel rotation angle and the lateral forces is analyzed. Based on the mapping relationship, the vehicle dynamics model is then laterally corrected to obtain the corrected vehicle dynamics model.

[0089] Step S204: Perform adaptive Kalman filtering on the target state variables in the corrected vehicle dynamics model to obtain the estimated information of the target state variables at the current moment.

[0090] The target state variables include the vehicle's central gravity (CG) sideslip angle, yaw rate, roll rate, and roll angle.

[0091] In the adaptive Kalman filtering process, a regularization method is used to update process noise and measurement noise information. Adaptive Kalman filtering refers to an estimation algorithm that dynamically adjusts filter parameters based on real-time operating conditions. Specifically, this can be achieved by constructing a new information covariance matrix and introducing regularization constraints. This method can suppress the interference of outlier observations on noise parameter estimation. The regularization method introduces mathematical constraints during the noise parameter update process. Specifically, this can be achieved by optimizing the regularization parameters using a generalized cross-validation criterion. This method can balance the weights of historical data and current observation information, avoiding filter divergence problems.

[0092] Adaptive Kalman filtering is applied to the target state variable in the modified vehicle dynamics model. Specifically, this can be improved by introducing regularization theory into the adaptive Kalman filtering method based on maximum likelihood estimation using noise statistics. The regularization principle of uncertainty problems is used to update the process noise and measurement noise information of the Kalman filtering algorithm in real time, so as to determine the estimated information of the target state variable at the current moment through the updated process noise and measurement noise information. Maximum likelihood estimation is a statistical solution based on the maximum likelihood principle. It provides a method for estimating model parameters based on experimental data. Once the probability distribution is determined, the optimal parameters that maximize the probability of the sample occurrence can always be found through independent and identical measurements. Therefore, maximum likelihood estimation can find the weights with the minimum error norm, which means that adaptive weight estimation and state estimation are complementary.

[0093] In some embodiments, the nonlinear dynamics model is laterally modified based on the vehicle's tire force and slip angle, including: fitting the mechanical properties of the tire using a magic tire model to determine the vehicle's tire force; calculating the vehicle's tire slip angle based on the vehicle's longitudinal velocity, lateral velocity, yaw rate, and front wheel steering angle; determining the mapping relationship between the vehicle's tire toe angle, tire camber angle, tire steering angle, and tire lateral force based on the vehicle's tire force and slip angle, and laterally modifying the vehicle dynamics model based on the mapping relationship to obtain the modified vehicle dynamics model.

[0094] The magic tire model is a nonlinear model that uses mathematical expressions to describe the mechanical behavior of a tire under different slip ratios, sideslip angles, and vertical loads. Specifically, it can be implemented using a combination of exponential functions and polynomials. Its purpose is to accurately fit the nonlinear characteristics of the tire under complex operating conditions, providing accurate tire force inputs for dynamic correction.

[0095] When a vehicle turns, the tires experience lateral forces, resulting in a slip angle. The Magic Tire Model calculates the tire lateral forces based on the tire's vertical load, slip ratio, and road surface friction coefficient. By measuring the steering wheel speed and the vehicle's longitudinal velocity, and combining geometric relationships, the real-time slip angle of each tire is derived. Based on the transformation relationship between the tire coordinate system and the vehicle coordinate system, the tire lateral forces are decomposed into vehicle lateral force components. Simultaneously considering the influence of suspension geometric parameters on tire camber and toe angles, a force transmission equation incorporating the suspension's dynamic characteristics is established. Substituting this equation into the lateral force balance equation of the original vehicle dynamics model corrects the model errors caused by neglecting suspension roll stiffness and damping, resulting in a corrected dynamics model that reflects the suspension-tire coupling effect.

[0096] When a car travels at high speed, the tires bear significant lateral dynamic loads, and the tires, especially the inner tires, experience substantial combined slippage, causing them to slide into the transition zone or friction zone of the slip curve. Furthermore, the vertical load on the tires and the road friction coefficient change with the car's cornering motion. Therefore, to simulate nonlinear tire forces, a magic tire model is used to fit the tire's mechanical properties. Based on the magic tire model, the expression for the vehicle's tire forces is determined as follows:

[0097] ,

[0098] ,

[0099] ,

[0100] ,

[0101] ,

[0102] in, For the tire force of the vehicle, The lateral force on a single front wheel. The lateral force on a single rear wheel. This refers to the vehicle's sideslip angle. The front wheel slip angle, Rear wheel slip angle, , , , , , , , , , and All are coefficients of the magic formula.

[0103] The expression for calculating the vehicle's slip angle is:

[0104] ,

[0105] ,

[0106] ,

[0107] in, For the sideslip angle of a single front wheel, Let be the lateral speed of the vehicle. For the front wheel steering angle, For the sideslip angle of a single rear wheel, This is the deflection angle of the vehicle's center of gravity.

[0108] Under complex operating conditions, vehicles often experience high lateral acceleration, and the nonlinear deformation of the suspension and steering systems leads to dynamic changes in four-wheel alignment parameters, significantly impacting the vehicle's lateral dynamics. Therefore, this study analyzes the vehicle's tire forces and slip angles to determine the mapping relationship between tire toe angle, tire camber angle, tire swerve angle, and tire lateral force. Based on this mapping relationship, the vehicle dynamics model is modified for lateral dynamics, resulting in the corrected vehicle dynamics model.

[0109] In suspension kinematics studies, the vehicle experiences significant body roll under extreme conditions, leading to substantial changes in wheel alignment parameters. Body roll causes the front steering wheels to rotate around the kingpin and the rear wheels around an axis perpendicular to the ground; that is, the wheel toe angle changes the roll steering angle. Figure 4 As shown in (a). Simultaneously, the tilting of the carriage will cause the front and rear wheels to rotate around the longitudinal axis, meaning the wheel camber angle will change, as shown in (a). Figure 4As shown in (b). In the study of suspension flexibility characteristics, the suspension guide rod elements deform under the action of lateral force or self-centering torque, causing the front wheel to rotate around the kingpin and the rear wheel to rotate around an axis perpendicular to the ground. That is, the wheel toe angle will produce a change in the deformation steering angle, such as... Figure 4 (c) Figure 4 As shown in (e). Simultaneously, this will also cause the front and rear wheels to rotate around the longitudinal axis, meaning the wheel camber angle will change due to deformation, such as... Figure 4 (d) Figure 4 As shown in (f). Considering the balance between improving model accuracy and ensuring computational efficiency, a modified vehicle dynamics model is established through vehicle lateral dynamics optimization. Without increasing the degrees of freedom, the changes in four-wheel alignment parameters caused by roll steering, roll camber, deformable steering, and deformable camber are compensated for in the wheel steering angle and camber angle. The wheel toe angle changes caused by roll steering and deformable steering are equivalent to geometric corrections of the wheel steering angle, and the vehicle's roll steering and deformable steering characteristics are introduced into the front and rear wheel slip angles of the dynamics model for correction; the wheel camber angle changes caused by roll camber and deformable camber are equivalent to force corrections of the lateral forces.

[0110] In some embodiments, it is assumed that the lateral force acts at the geometric center of the tire. However, in actual cornering, the vehicle will generate tire trail. and The corrected wheelbase distance from the center of mass to the front and rear axles is:

[0111] ,

[0112] ,

[0113] in, This is the corrected wheelbase from the center of gravity to the front axle. This is the corrected wheelbase from the center of mass to the rear axle.

[0114] The corrected front and rear wheel steering angles can be expressed as:

[0115] ,

[0116] ,

[0117] ,

[0118] ,

[0119] in, The corrected front wheel slip angle, The corrected rear wheel slip angle. The corrected lateral force of the front axle tires. The corrected lateral force of the rear axle tires. The corrected front axle tire return torque, The corrected rear axle tire return torque, , , , , and All of these are corrected weighting coefficients.

[0120] Camber causes the wheel to tend to roll along the intersection of its own axis of rotation and the ground. Due to the constraint of the axle, a camber lateral force is generated at the tire contact patch. Assuming that the lateral force consists only of lateral force and camber force, the change in wheel camber angle caused by vehicle roll and deformation camber is equivalent to a force correction for the lateral force. The corrected front and rear axle camber lateral forces are:

[0121] ,

[0122] ,

[0123] in, The corrected front axle camber lateral force, This is the corrected rear axle camber lateral force. For the camber stiffness of the front axle tires, For rear axle tire camber stiffness, , , , , and All of these are corrected weighting coefficients.

[0124] Therefore, the corrected expressions for the lateral forces of the front and rear axle tires can be obtained as follows:

[0125] ,

[0126] .

[0127] Based on this, the expression for the corrected vehicle dynamics model can be obtained as follows:

[0128] ,

[0129] ,

[0130] ,

[0131] Where m is the total vehicle mass. For the vehicle's lateral acceleration, Let be the longitudinal speed of the vehicle. Let yaw rate be the vehicle's angular velocity. For the sprung mass, This is the distance from the vehicle's roll center of gravity to the roll axis. The roll angle acceleration of the vehicle. The lateral force is from the front wheel. The lateral force is from the rear wheel. Let be the moment of inertia of the entire vehicle about its center of mass in the vertical direction. For the vehicle's yaw acceleration, This is the distance from the vehicle's center of gravity to the front axle. This is the distance from the vehicle's center of gravity to the rear axle. Let be the longitudinal rotational inertia of the entire vehicle about its center of mass. It is the gravitational constant. For equivalent roll stiffness, The vehicle's roll angle. For equivalent roll damping, This represents the vehicle's roll rate.

[0132] In some embodiments, adaptive Kalman filtering is performed on the target state variables in the modified vehicle dynamics model, including: determining prior estimation information of the target state variables at the current moment based on the estimated information of the target state variables at the previous moment and the front wheel steering angle of the vehicle; determining prior estimation error covariance information at the current moment based on the posterior estimation error covariance information and process noise information at the previous moment, and determining the Kalman gain at the current moment based on the prior estimation error covariance information; determining the innovation sequence information at the current moment based on the measurement information of the target measured variables at the current moment and the prior estimation information of the target state variables at the current moment; determining the estimation information of the target state variables at the current moment based on the prior estimation information of the target state variables at the current moment, the Kalman gain, and the innovation sequence information; and after obtaining the estimation information of the target state variables at the current moment, updating the process noise information and measurement noise information at the current moment using a regularization method. The target measured variables include the vehicle's yaw rate and lateral state variables.

[0133] Prior estimation information refers to the current state value predicted based on the previous state estimate and system input. Specifically, it can be calculated using the state transition equation combined with the front wheel steering angle input, and is used to initially infer the dynamic change trend of the vehicle. Process noise information refers to the error covariance caused by unmodeled dynamics or external disturbances in the system model. Specifically, it can be updated using the covariance propagation formula combined with historical error data, and is used to quantify the uncertainty of model predictions. Kalman gain refers to the correction weight of observation information on the state estimate. Specifically, it can be calculated based on the prior estimation error covariance and the observation noise covariance, and is used to balance the reliability of model predictions and sensor measurements. Innovation sequence information refers to the difference between the actual observed value and the predicted observed value. Specifically, it can be calculated using the difference between the current observation information and the prior estimated observation prediction, and is used to characterize the degree to which the system state deviates from the prediction.

[0134] The expression for determining the prior estimate of the target state variable at the current moment is:

[0135] ,

[0136] in, This provides prior estimates of the target state variables at the current moment. This is the estimated information of the target state variables from the previous time step. This represents the front wheel angle of the vehicle at the previous moment.

[0137] The expression for determining the prior estimate error covariance information at the current moment is:

[0138] ,

[0139] in, This provides the prior estimate of the error covariance information at the current moment. For state transition information, This provides the posterior estimation error covariance information from the previous time step. This is the process noise information updated from the previous moment.

[0140] The expression for determining the Kalman gain at the current moment is:

[0141] ,

[0142] in, The Kalman gain at the current moment. For the current observation information, This is the measurement noise information from the previous moment.

[0143] The expression for determining the current information sequence information is:

[0144] ,

[0145] in, This is the current information sequence. Measurement information for the target variable at the current moment.

[0146] The expression for determining the estimated information of the target state variable at the current moment is:

[0147] ,

[0148] in, This provides estimated information about the target state variables at the current moment.

[0149] After obtaining the estimated information of the target state variables at the current moment, the vehicle's state is estimated in real time using the estimated information of the target state variables at the current moment. Then, a regularization method is used to update the process noise information and measurement noise information at the current moment.

[0150] In some embodiments, the method for updating process noise information and measurement noise information using a regularization method includes: introducing a regularization parameter to describe the process noise information at the current moment based on the regularization theory of uncertainty problems; determining a first target regularization parameter using a generalized cross-validation method; the first target regularization parameter being a regularization parameter that minimizes the process noise information at the current moment; introducing factors affecting both the carriage tilt angle and the vehicle speed, updating the first target regularization parameter based on the factors to obtain a second target regularization parameter; updating the process noise information at the current moment based on the second target regularization parameter to obtain the updated process noise information at the current moment, and determining the measurement noise information at the current moment based on the updated process noise information at the previous moment.

[0151] Regularization parameters are adjustment coefficients used to balance the relationship between data fitting and model complexity during noise estimation. They can be implemented using matrix construction based on singular value decomposition, and their function is to suppress the ill-conditioned nature of noise covariance information. Generalized cross-validation is a numerical optimization method that determines the regularization parameter by minimizing the sum of squared prediction errors. This can be achieved by calculating the matrix trace, and its function is to automatically select the optimal regularization parameter value to avoid bias from manual parameter tuning. Influence factors are weighted coefficients reflecting the impact of the vehicle's roll angle and longitudinal speed on noise parameters. They can be implemented using the ratio of the roll angle reference stable value to the speed reference stable value, and their function is to dynamically adjust the regularization parameter according to the vehicle's dynamic characteristics to improve parameter adaptability.

[0152] When updating process noise, the singular value distribution of the Kalman gain is first calculated through singular value decomposition, and the update expression for process noise is constructed by combining it with regularization parameters. Then, the regularization parameters are iteratively optimized using a generalized cross-validation function to obtain the first objective regularization parameter that minimizes the noise estimation error. Furthermore, the ratio of the vehicle's side tilt angle to its longitudinal velocity is introduced as an influencing factor to dynamically correct the first objective regularization parameter, resulting in a second objective regularization parameter. This incorporates the actual vehicle motion state into the noise parameter update process. Finally, the process noise covariance information is recalculated based on the second objective regularization parameter, and the measurement noise information is derived based on the new covariance matrix. This method achieves dynamic adjustment of noise information through two parameter optimizations, enabling the filtering process to adapt to uncertainties under different driving conditions.

[0153] Since it is assumed that the noise has a Gaussian distribution and that the measurement at time step k is independent of other time steps, the measurement noise information can be calculated using partial derivative operations as follows:

[0154] ,

[0155] in, This is the measurement noise information at the current moment.

[0156] The same strategy can be used to obtain the process noise information at the current moment, based on the measurement noise information and approximate representation at the current moment:

[0157] ,

[0158] in, This represents the process noise information at the current moment.

[0159] The statistical characteristics of noise in real-world systems suffer from biases that are difficult to model accurately (such as non-Gaussianity and time-varying properties), and the strong coupling between system model errors and noise estimation leads to a fundamental mismatch between theoretical assumptions and actual dynamics. Therefore, the selection of process noise information is an uncertainty problem. Introducing regularization theory to optimize the value of process noise in the face of uncertainty can be transformed into the following linear least squares problem:

[0160] ,

[0161] .

[0162] The Tikhonov regularization method determines an effective value by replacing the above minimization problem with a penalized least squares problem expression, i.e.:

[0163] ,

[0164] in, This is the regularization parameter.

[0165] The regular solution defined by the above penalized least squares problem expression can be solved using the least squares method:

[0166] ,

[0167] in, It is an identity matrix.

[0168] According to the theory of singular value decomposition (SVD), Decomposed into:

[0169] ,

[0170] in, , , and They represent The left singular vector and the right singular vector, , for The singular values ​​of , and satisfy .

[0171] The above expression can be rewritten as:

[0172] ,

[0173] The solution can be obtained as:

[0174] ,

[0175] Based on the above, according to the Singular Value Decomposition (SVD) theory, The expression describing the process noise information at the current moment, after decomposition and introduction of regularization parameters, is as follows:

[0176] ,

[0177] in, This is a regularization operator.

[0178] For the Tikhonov regularization method, its regularization operator is described as follows:

[0179] .

[0180] The fundamental principle for selecting the optimal regularization parameter is to simultaneously consider the stability and accuracy of the solution, thereby achieving an optimal balance. Generalized cross-validation (GCV) is the most widely used method. The expression for determining the first objective regularization parameter using GCV is:

[0181] ,

[0182] in, For the function output value, for The minimum value, The first objective is the regularization parameter. The operation for calculating the trace of a matrix;

[0183] Considering that the coupling effect between roll motion and suspension system has a significant impact on vehicle dynamics under complex operating conditions (such as emergency obstacle avoidance and high-speed cornering), and the uncertainty of process noise increases under these conditions, a second objective regularization parameter is proposed. : After obtaining the first target regularization parameters Based on this, the influence factors of the carriage tilt angle and vehicle speed are added. The expression for the second objective regularization parameter is as follows:

[0184] ,

[0185] in, For the second objective regularization parameter, This is a reference stable value for the carriage tilt angle. This is the longitudinal velocity reference stable value. The vehicle's roll angle. The longitudinal speed of the vehicle;

[0186] Substituting the second objective regularization parameter into the expression describing the process noise information at the current moment, we can obtain the expression for obtaining the updated process noise information at the current moment:

[0187] ,

[0188] in, This is the updated process noise information at the current moment.

[0189] Based on the updated process noise information from the previous time step, the measurement noise information at the current time step is determined, and the expression for obtaining the measurement noise information at the current time step is:

[0190] ,

[0191] ,

[0192] in, This is the measurement noise information at the current moment. For the current observation information, This provides the prior estimate of the error covariance information at the current moment. This provides the state transition information for the current moment.

[0193] Please see Figure 5 This application also provides a vehicle state estimation device that can implement the above-described vehicle state estimation method. The device includes:

[0194] The first module 501 is used to obtain the vehicle status parameters under the current operating conditions.

[0195] The second module 502 is used to construct a vehicle dynamics model based on vehicle state parameters; the vehicle dynamics model is a nonlinear dynamics model that describes the force balance relationship of the vehicle in the lateral, yaw and roll directions.

[0196] The third module 503 is used to perform lateral dynamic correction on the nonlinear dynamic model based on the vehicle's tire force and sideslip angle to obtain the corrected vehicle dynamic model.

[0197] The fourth module 504 is used to perform adaptive Kalman filtering on the target state variables in the modified vehicle dynamics model to obtain the estimated information of the target state variables at the current moment. The target state variables include the vehicle's center of gravity sideslip angle, yaw rate, roll rate and roll angle. During the adaptive Kalman filtering process, a regularization method is used to update the process noise information and measurement noise information.

[0198] The specific implementation of this vehicle state estimation device is basically the same as the specific embodiment of the vehicle state estimation method described above, and will not be repeated here.

[0199] Figure 6 This is a block diagram illustrating an electronic device according to an exemplary embodiment.

[0200] The following reference Figure 6 To describe an electronic device 600 according to such an embodiment of the present disclosure. Figure 6 The electronic device 600 shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments disclosed herein.

[0201] like Figure 6 As shown, the electronic device 600 is presented in the form of a general-purpose computing device. The components of the electronic device 600 may include, but are not limited to: at least one processing unit 610, at least one storage unit 620, a bus 630 connecting different system components (including storage unit 620 and processing unit 610), a display unit 640, etc.

[0202] The storage unit stores program code, which can be executed by the processing unit 610, causing the processing unit 610 to perform the steps described in the above-described vehicle state estimation method section of this specification according to various exemplary embodiments of this disclosure.

[0203] Storage unit 620 may include a readable medium in the form of a volatile storage unit, such as random access memory (RAM) 6201 and / or cache memory 6202, and may further include a read-only memory (ROM) 6203.

[0204] Storage unit 620 may also include a program / utility 6204 having a set (at least one) program module 6205, such program module 6205 including but not limited to: operating system, one or more application programs, other program modules and program data, each or some combination of these examples may include an implementation of a network environment.

[0205] Bus 630 can represent one or more of several types of bus structures, including a memory cell bus or memory cell controller, a peripheral bus, a graphics acceleration port, a processing unit, or a local bus using any of the various bus structures.

[0206] Electronic device 600 can also communicate with one or more external devices 600' (e.g., keyboard, pointing device, Bluetooth device, etc.), and with one or more devices that enable a user to interact with electronic device 600, and / or with any device that enables electronic device 600 to communicate with one or more other computing devices (e.g., router, modem, etc.). This communication can be performed via input / output (I / O) interface 650. Furthermore, electronic device 600 can also communicate with one or more networks (e.g., local area network (LAN), wide area network (WAN), and / or public networks, such as the Internet) via network adapter 660. Network adapter 660 can communicate with other modules of electronic device 600 via bus 630. It should be understood that, although not shown in the figures, other hardware and / or software modules can be used in conjunction with electronic device 600, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0207] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the vehicle state estimation method described above.

[0208] The vehicle state estimation method, apparatus, device, and storage medium provided in this application construct a vehicle dynamics model based on the acquired vehicle state parameters. This vehicle dynamics model is a nonlinear dynamics model describing the force balance relationship of the vehicle in lateral, yaw, and roll conditions. Lateral dynamics correction is then applied to the nonlinear dynamics model based on the vehicle's tire forces and sideslip angle to obtain a corrected vehicle dynamics model. Adaptive Kalman filtering is then applied to the target state variables in the corrected vehicle dynamics model. During the adaptive Kalman filtering process, regularization methods are used to update process noise and measurement noise information to obtain the estimated information of the target state variables at the current moment. By constructing a vehicle dynamics model describing the force balance relationship of the vehicle in lateral, yaw, and roll conditions, the modeling accuracy of the vehicle's motion state under complex working conditions is significantly improved. The use of regularization methods to update process noise and measurement noise information during the adaptive Kalman filtering process overcomes the limitations of traditional fixed parameters, enabling the algorithm to adapt to the dynamic changes in road conditions and sensor errors in real time, effectively suppressing filter divergence, improving the accuracy of vehicle state assessment, and suppressing filter divergence during the vehicle state process.

[0209] From the above description of the embodiments, those skilled in the art will readily understand that the exemplary embodiments described herein can be implemented by software or by combining software with necessary hardware. Therefore, the technical solutions according to the embodiments of this disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, USB flash drive, external hard drive, etc.) or on a network, including several instructions to cause a computing device (such as a personal computer, server, or network device, etc.) to execute the methods described above according to the embodiments of this disclosure.

[0210] The program product may employ any combination of one or more readable media. A readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples (a non-exhaustive list) of readable storage media include: electrical connections having one or more wires, portable disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0211] Computer-readable storage media may include data signals propagated in baseband or as part of a carrier wave, carrying readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A readable storage medium may also be any readable medium other than a readable storage medium that can transmit, propagate, or transfer a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the readable storage medium may be transmitted using any suitable medium, including but not limited to wireless, wired, optical fiber, RF, etc., or any suitable combination thereof.

[0212] Those skilled in the art will understand that the above modules can be distributed in the device as described in the embodiments, or they can be modified accordingly and placed in one or more devices that are unique to this embodiment. The modules in the above embodiments can be combined into one module, or they can be further divided into multiple sub-modules.

[0213] Exemplary embodiments of this disclosure have been specifically shown and described above. It should be understood that this disclosure is not limited to the detailed structures, arrangements, or implementations described herein; rather, this disclosure is intended to cover various modifications and equivalent arrangements contained within the spirit and scope of the appended claims.

Claims

1. A vehicle state estimation method characterized by comprising: The method comprises the following steps: acquiring vehicle state parameters of the vehicle in a current working condition; constructing a vehicle dynamics model according to the vehicle state parameters; the vehicle dynamics model is a nonlinear dynamics model describing force balance relationship of the vehicle in lateral, yaw and roll directions; performing lateral dynamics correction on the nonlinear dynamics model according to tire force and side slip angle of the vehicle to obtain a corrected vehicle dynamics model; performing adaptive Kalman filtering on target state variables in the corrected vehicle dynamics model to obtain estimation information of the target state variables at a current time; the target state variables include side slip angle, yaw angular velocity, roll angular velocity and roll angle of the vehicle; in the process of adaptive Kalman filtering, a regularization method is used to update process noise information and measurement noise information; the lateral dynamics correction on the nonlinear dynamics model according to tire force and side slip angle of the vehicle comprises: fitting mechanical properties of the tire by using a magic tire model to determine tire force of the vehicle; calculating tire side slip angle of the vehicle according to longitudinal velocity, lateral velocity, yaw angular velocity and front wheel steering angle of the vehicle; determining mapping relationship between tire toe angle, tire camber angle, tire steering angle and tire lateral force of the vehicle according to tire force and tire side slip angle of the vehicle, and performing lateral dynamics correction on the vehicle dynamics model according to the mapping relationship to obtain the corrected vehicle dynamics model.

2. The vehicle state estimation method according to claim 1, characterized by, an expression of the vehicle dynamics model is: ， ， ， wherein, is the resultant of the external forces along the Y axis on the vehicle, m is the mass of the vehicle, is the lateral acceleration of the vehicle, is the longitudinal velocity of the vehicle, is the yaw angular velocity of the vehicle, is the sprung mass, is the distance of the vehicle roll center to the roll axis, is the roll angular acceleration of the vehicle, is the front wheel lateral force, is the rear wheel lateral force, is the moment of the external forces around the Z axis on the vehicle, is the moment of inertia of the vehicle around the vertical axis through the center of mass, is the yaw acceleration of the vehicle, is the distance of the vehicle center of mass to the front axle of the vehicle, is the distance of the vehicle center of mass to the rear axle of the vehicle, is the moment of the external forces around the X axis on the vehicle, is the moment of inertia of the vehicle around the longitudinal axis through the center of mass, is the gravitational constant, is the equivalent roll stiffness, is the roll angle of the vehicle, is the equivalent roll damping, is the roll angular velocity of the vehicle.

3. The vehicle state estimation method according to claim 1, characterized by, an expression of the corrected vehicle dynamics model is: ， ， ， where m is the vehicle mass, is the lateral acceleration of the vehicle, is the longitudinal velocity of the vehicle, is the yaw angular velocity of the vehicle, is the sprung mass, is the distance of the vehicle roll center to the roll axis, is the roll angular acceleration of the vehicle, is the front wheel lateral force, is the rear wheel lateral force, is the vehicle moment of inertia about the vertical through the center of mass, is the yaw acceleration of the vehicle, is the distance of the vehicle center of mass to the front axle of the vehicle, is the distance of the vehicle center of mass to the rear axle of the vehicle, is the vehicle moment of inertia about the longitudinal through the center of mass, is the gravitational constant, is the equivalent roll stiffness, is the roll angle of the vehicle, is the equivalent roll damping, is the roll angular velocity of the vehicle.

4. The vehicle state estimation method according to claim 1, characterized by, the adaptive Kalman filtering on the target state variables in the corrected vehicle dynamics model comprises: determining prior estimation information of the target state variables at a current time according to estimation information of the target state variables at a previous time and front wheel steering angle of the vehicle; determining prior estimation error covariance information at the current time according to prior estimation error covariance information at the previous time and process noise information, and determining Kalman gain at the current time according to the prior estimation error covariance information at the current time; determining innovation sequence information at the current time according to measurement information of target measurement variables and prior estimation information of the target state variables at the current time; the target measurement variables include yaw angular velocity and lateral state variables of the vehicle; determining estimation information of the target state variables at the current time according to the prior estimation information, Kalman gain and innovation sequence information of the target state variables at the current time; after obtaining the estimation information of the target state variables at the current time, updating process noise information and measurement noise information at the current time by using a regularization method.

5. The vehicle state estimation method according to claim 1, characterized by, the method for updating the process noise information and the measurement noise information by using the regularization method comprises: based on regularization theory of uncertainty problems, introducing a regularization parameter to describe process noise information at a current time; determining a first target regularization parameter by using a generalized cross-validation method; the first target regularization parameter is a regularization parameter that makes the process noise information at the current time reach a minimum value; Introducing an influencing factor affecting both the vehicle body roll angle and the vehicle speed, updating the first target regularization parameter according to the influencing factor to obtain a second target regularization parameter; Updating the process noise information at the current time according to the second target regularization parameter to obtain updated process noise information at the current time, and determining the measurement noise information at the current time according to the updated process noise information at the previous time.

6. The vehicle state estimation method according to claim 5, characterized by, The expression for introducing the regularization parameter to describe the process noise information at the current time is: ， wherein, is the process noise information at the current time, is a regularization parameter, is the Kalman gain at the current time, is a singular value of T denotes transpose, i is a positive integer, i ∈ [1, n], n is a total number of singular values of is a left singular vector of is a right singular vector of is the innovation covariance information at the current time; The expression for determining the first target regularization parameter by using the generalized cross-validation method is: ， wherein, is a function output value, is a minimum value of, is a first target regularization parameter, is an operation to calculate a trace of a matrix; The expression for obtaining the second target regularization parameter is: ， wherein, is a second target regularization parameter, is a lateral roll angle reference stability value, is a longitudinal speed reference stability value, is a lateral roll angle of the vehicle, is a longitudinal speed of the vehicle; The expression for obtaining the updated process noise information at the current time is: ， wherein, is the updated process noise information for the current time instant; The expression for obtaining the measurement noise information at the current time is: ， ， wherein is measurement noise information for the current time instant, is observation information for the current time instant, is prior estimation error covariance information for the current time instant, is state transition information for the current time instant.

7. A vehicle state estimation device characterized by comprising: It comprises: A first module for obtaining vehicle state parameters of a vehicle in a current working condition; A second module for constructing a vehicle dynamics model according to the vehicle state parameters; The vehicle dynamics model is a nonlinear dynamics model describing the force balance relationship of the vehicle in the lateral, yaw and roll directions; A third module for performing lateral dynamics correction on the nonlinear dynamics model according to the tire force and the side slip angle of the vehicle to obtain a corrected vehicle dynamics model; A fourth module for performing adaptive Kalman filtering on a target state variable in the corrected vehicle dynamics model to obtain estimation information of the target state variable at the current time; the target state variable includes the vehicle's center side slip angle, yaw angular velocity, roll angular velocity and roll angle, and the process noise information and the measurement noise information are updated by using a regularization method during the adaptive Kalman filtering; The lateral dynamics correction on the nonlinear dynamics model according to the tire force and the side slip angle of the vehicle comprises: The mechanical properties of the tire are fitted by using a magic tire model to determine the tire force of the vehicle; The tire side slip angle of the vehicle is calculated according to the longitudinal speed, lateral speed, yaw angular velocity and front wheel steering angle of the vehicle; The mapping relationship between the tire toe angle, tire camber angle, tire rotation angle and tire lateral force of the vehicle is determined according to the tire force and the tire side slip angle of the vehicle, and the vehicle dynamics model is corrected in the lateral dynamics according to the mapping relationship to obtain the corrected vehicle dynamics model.

8. An electronic device, comprising: The electronic device comprises a memory and a processor, the memory stores a computer program, and the processor implements the vehicle state estimation method of any one of claims 1 to 6 when executing the computer program.

9. A computer-readable storage medium storing a computer program, the computer program comprising instructions that, when executed by a computer, cause the computer to perform the method of any one of claims 1 to 8. The computer program is executed by the processor to implement the vehicle state estimation method of any one of claims 1 to 6.

Images