Vehicle transverse control method and system based on nonlinear vehicle model

Abstract

The invention discloses a vehicle transverse control method and system based on a nonlinear vehicle model, and relates to the field of vehicle control. Vehicle motion state data are acquired, and a kinematic model of a target vehicle is established; obtaining a vehicle target motion track, discretizing the kinematic model of the target vehicle in combination with the vehicle target motion track, and generating an optimal steering wheel angle instruction based on the kinematic model; scaling the optimal steering wheel angle instruction according to a preset coefficient to obtain a corresponding instruction sequence; obtaining historical state data of the vehicle, constructing a neural network vehicle model based on the historical state data of the vehicle, inputting the historical state data of the vehicle into the neural network vehicle model for training, and obtaining a final parameter value; and inputting the instruction sequence into the trained neural network vehicle model, obtaining a vehicle response output signal corresponding to the instruction sequence, determining a minimum cost value, and taking a steering wheel angle instruction corresponding to the minimum cost value as a final steering instruction.

Description

Technical Field

[0001] This invention relates to the field of vehicle control, and in particular to a vehicle lateral control method based on a nonlinear vehicle model, a vehicle lateral control system based on a nonlinear vehicle model, electronic equipment, and storage medium. Background Technology

[0002] The control of intelligent commercial vehicles is relatively difficult because commercial vehicles have poor response performance and long response delay time. Therefore, the control effect of intelligent commercial vehicles is relatively poor, mainly manifested in its low control accuracy, large actuator adjustment range, and high adjustment frequency.

[0003] To improve the control performance of intelligent vehicles, the most effective approach is to increase model complexity. This can be achieved by replacing the kinematic model that treats the vehicle as an ideal rigid body with a dynamic model that considers tire and suspension characteristics; and by considering the dynamic responses of the steering, drive, and braking systems. While these models significantly improve control performance in passenger vehicle control algorithms, linear dynamic models are less effective for commercial vehicles due to their strong nonlinear characteristics, offering limited improvement in control performance. In contrast, nonlinear models based on neural networks provide better modeling of the dynamic characteristics of commercial vehicles, resulting in a significant improvement in control performance.

[0004] However, using a nonlinear neural network vehicle model inevitably leads to the following problems: the model solution becomes complex or even impossible; the computing power required for the solution increases dramatically; and the model solution has poor stability. Summary of the Invention

[0005] In view of this, the purpose of this invention is to provide a vehicle lateral control method based on a nonlinear vehicle model, a vehicle lateral control system based on a nonlinear vehicle model, an electronic device and a storage medium. It aims to propose an algorithm suitable for intelligent control of highly complex commercial vehicles, which can comprehensively consider the accurate description of the dynamic characteristics of commercial vehicles by the nonlinear linear model, while ensuring the stable solution of the controller with low computing power consumption, and taking into account both the complexity of model solution and control effect.

[0006] This invention provides the following solution:

[0007] According to one aspect of the present invention, a vehicle lateral control method based on a nonlinear vehicle model is provided, comprising the following steps:

[0008] Acquire vehicle motion state data, and establish a kinematic model of the target vehicle based on the vehicle motion state data;

[0009] Obtain the target vehicle motion trajectory, and in conjunction with the target vehicle motion trajectory, discretize the kinematic model of the target vehicle to generate the optimal steering wheel angle command based on the kinematic model;

[0010] The optimal steering wheel angle command is scaled by a preset coefficient to obtain the corresponding command sequence;

[0011] Obtain historical vehicle status data, and construct a neural network vehicle model based on the historical vehicle status data. The neural network vehicle model includes: an input layer, two hidden layers, and an output layer.

[0012] The historical vehicle status data is input into a neural network vehicle model for training to obtain the final parameter values.

[0013] The instruction sequence is input into the trained neural network vehicle model to obtain the vehicle response output signal corresponding to the instruction sequence;

[0014] Based on each set of candidate steering wheel angle command sequences and their corresponding vehicle response output signals, the cost value corresponding to each command sequence is calculated. The minimum cost value is determined by screening and interpolation optimization of the cost value, and the steering wheel angle command corresponding to the minimum cost value is taken as the final steering command.

[0015] Furthermore, the vehicle motion data includes: vehicle longitudinal speed, front wheel steering angular rate, and the angle between the vehicle's longitudinal axis and the geodetic coordinate system.

[0016] Furthermore, the vehicle's historical status data includes: historical vehicle speed signal, historical steering wheel angle signal, vehicle weight signal, and historical vehicle acceleration signal.

[0017] Furthermore, including:

[0018] Each neuron from the input layer to the hidden layer and from the hidden layer to the output layer is connected to all neurons in the previous layer and contains a bias.

[0019] Furthermore, including:

[0020] During the training of the neural network vehicle model, the backpropagation method is used to calculate the error of each neural unit.

[0021] Specifically, the derivative of the error between the neural network output signal and the real signal collected from the actual vehicle is calculated to obtain the error of the output layer neural unit;

[0022] By recursively extrapolating the errors of the output layer neurons layer by layer, the errors of each neuron in the two hidden layers can be obtained.

[0023] Furthermore, including:

[0024] When calculating the cost value based on each set of candidate steering wheel angle command sequences and their corresponding vehicle response output signals, the objective function of MPC is used for calculation. The cost value characterizes the degree of deviation between the vehicle response corresponding to the candidate steering wheel angle command sequence and the expected trajectory.

[0025] Furthermore, including

[0026] The process of filtering and interpolating the cost value is as follows: select the two smallest cost values ​​from all candidate steering wheel angle instruction sequences;

[0027] The final minimum value is obtained by interpolating the two minimum values, and the steering wheel angle command corresponding to the final minimum value is determined.

[0028] According to two aspects of the present invention, a vehicle lateral control system based on a nonlinear vehicle model is provided, comprising:

[0029] The system comprises a kinematic model building module, a first instruction generation module, an instruction sequence acquisition module, a neural network model building module, a model training module, a response signal acquisition module, and a final steering instruction determination module.

[0030] The kinematic model building module is used to acquire vehicle motion state data and build a kinematic model of the target vehicle based on the vehicle motion state data.

[0031] The first instruction generation module is used to acquire the target vehicle motion trajectory, and in combination with the target vehicle motion trajectory, discretize the kinematic model of the target vehicle to generate the optimal steering wheel angle instruction based on the kinematic model.

[0032] The instruction sequence acquisition module is used to scale the optimal steering wheel angle instruction by a preset coefficient to obtain the corresponding instruction sequence;

[0033] The neural network model building module is used to acquire historical vehicle state data and build a neural network vehicle model based on the historical vehicle state data. The neural network vehicle model includes an input layer, two hidden layers, and an output layer.

[0034] The model training module is used to input historical vehicle state data into the neural network vehicle model for training and to obtain the final values ​​of the parameters.

[0035] The response signal acquisition module is used to input the instruction sequence into the trained neural network vehicle model and acquire the vehicle response output signal corresponding to the instruction sequence.

[0036] The final steering command determination module is used to calculate the cost value of each command sequence based on each set of candidate steering wheel angle command sequences and their corresponding vehicle response output signals. The minimum cost value is determined by filtering and interpolating the cost values, and the steering wheel angle command corresponding to the minimum cost value is taken as the final steering command.

[0037] According to three aspects of the present invention, an electronic device is provided, comprising: a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus;

[0038] The memory stores a computer program that, when executed by a processor, causes the processor to perform steps of a vehicle lateral control method based on a nonlinear vehicle model.

[0039] According to four aspects of the present invention, a computer-readable storage medium is provided that stores a computer program executable by an electronic device, which, when run on the electronic device, causes the electronic device to perform the steps of a vehicle lateral control method based on a nonlinear vehicle model.

[0040] Compared with the prior art, the present invention has the following advantages:

[0041] This application employs a highly complex intelligent control algorithm for commercial vehicles, using a nonlinear model to accurately describe the dynamic characteristics of commercial vehicles, while ensuring stable solution of the controller with relatively low computational power consumption. Attached Figure Description

[0042] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0043] Figure 1 This is a flowchart of a vehicle lateral control method based on a nonlinear vehicle model provided by one or more embodiments of the present invention.

[0044] Figure 2 This is a structural diagram of a vehicle lateral control system based on a nonlinear vehicle model provided by one or more embodiments of the present invention.

[0045] Figure 3 This is a schematic diagram illustrating the establishment of a vehicle kinematics model according to a specific embodiment of the present invention.

[0046] Figure 4This is an electronic device structural block diagram of a vehicle lateral control method based on a nonlinear vehicle model provided by one or more embodiments of the present invention. Detailed Implementation

[0047] The technical solution of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0048] The terminology used in the embodiments of this application is for the purpose of describing particular embodiments only and is not intended to limit the application. The singular forms “a,” “said,” and “the” used in the embodiments of this application and the appended claims are also intended to include the plural forms, and “multiple” generally includes at least two unless the context clearly indicates otherwise.

[0049] It should be understood that the term "and / or" used in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / " in this article generally indicates that the preceding and following related objects have an "or" relationship.

[0050] It should be understood that although the terms first, second, third, etc., may be used in the embodiments of this application, these descriptions should not be limited to these terms. These terms are only used to distinguish the descriptions. For example, first may also be referred to as second without departing from the scope of the embodiments of this application, and similarly, second may also be referred to as first.

[0051] Depending on the context, the words “if” or “suppose” as used here can be interpreted as “when” or “in response to determination” or “in response to detection.” Similarly, depending on the context, the phrases “if determination” or “if detection (of the stated condition or event)” can be interpreted as “when determination” or “in response to determination” or “when detection (of the stated condition or event)” or “in response to detection (of the stated condition or event).”

[0052] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that an article or device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such an article or device. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the article or device that includes said element.

[0053] It should be noted that any symbols and / or numbers present in the specification that are not marked in the accompanying drawings are not reference numerals.

[0054] Figure 1 This is a flowchart of a vehicle lateral control method based on a nonlinear vehicle model provided by one or more embodiments of the present invention.

[0055] like Figure 1 As shown, it includes the following steps:

[0056] Step S1: Obtain vehicle motion state data and establish a kinematic model of the target vehicle based on the vehicle motion state data;

[0057] Step S2: Obtain the target vehicle motion trajectory; combine the target vehicle motion trajectory to discretize the kinematic model of the target vehicle and generate the optimal steering wheel angle command based on the kinematic model.

[0058] Step S3: Scale the optimal steering wheel angle command according to a preset coefficient to obtain the corresponding command sequence;

[0059] Step S4: Obtain vehicle historical state data, and construct a neural network vehicle model based on the vehicle historical state data. The neural network vehicle model includes: an input layer, two hidden layers, and an output layer.

[0060] Step S5: Input the vehicle historical state data into the neural network vehicle model for training and obtain the final parameter values;

[0061] Step S6: Input the instruction sequence into the trained neural network vehicle model to obtain the vehicle response output signal corresponding to the instruction sequence;

[0062] Step S7: Based on each group of candidate steering wheel angle command sequences and their corresponding vehicle response output signals, calculate the cost value corresponding to each command sequence. Determine the minimum cost value by filtering and interpolating the cost values, and use the steering wheel angle command corresponding to the minimum cost value as the final steering command.

[0063] Furthermore, the vehicle motion data includes: vehicle longitudinal speed, front wheel steering angular rate, and the angle between the vehicle's longitudinal axis and the geodetic coordinate system.

[0064] Furthermore, the vehicle's historical status data includes: historical vehicle speed signal, historical steering wheel angle signal, vehicle weight signal, and historical vehicle acceleration signal.

[0065] Furthermore, including:

[0066] Each neuron from the input layer to the hidden layer and from the hidden layer to the output layer is connected to all neurons in the previous layer and contains a bias.

[0067] Furthermore, including:

[0068] During the training of the neural network vehicle model, the backpropagation method is used to calculate the error of each neural unit.

[0069] Specifically, the derivative of the error between the neural network output signal and the real signal collected from the actual vehicle is calculated to obtain the error of the output layer neural unit;

[0070] By recursively extrapolating the errors of the output layer neurons layer by layer, the errors of each neuron in the two hidden layers can be obtained.

[0071] Furthermore, including:

[0072] When calculating the cost value based on each set of candidate steering wheel angle command sequences and their corresponding vehicle response output signals, the objective function of MPC is used for calculation. The cost value characterizes the degree of deviation between the vehicle response corresponding to the candidate steering wheel angle command sequence and the expected trajectory.

[0073] Furthermore, including

[0074] The process of filtering and interpolating the cost value is as follows: select the two smallest cost values ​​from all candidate steering wheel angle instruction sequences;

[0075] The final minimum value is obtained by interpolating the two minimum values, and the steering wheel angle command corresponding to the final minimum value is determined.

[0076] Specifically, by using a simple kinematic model and solving it using MPC, different instruction sequences are obtained through scaling at different ratios. The corresponding costs are then calculated using a neural network vehicle model, and interpolation is used to obtain the final control command. This method allows for solving only a simple kinematic model, consuming less computational power and providing stable results. After simple scaling, the command values ​​calculated using a complex model do not consume much computational power. By comparing the costs generated by several different instruction sequences, a superior command value compared to the command obtained from the simple kinematic model can be identified.

[0077] Figure 2 This is a structural diagram of a vehicle lateral control system based on a nonlinear vehicle model provided by one or more embodiments of the present invention.

[0078] like Figure 2As shown, it includes: a kinematic model building module, a first instruction generation module, an instruction sequence acquisition module, a neural network model building module, a model training module, a response signal acquisition module, and a final steering instruction determination module;

[0079] The kinematic model building module is used to acquire vehicle motion state data and build a kinematic model of the target vehicle based on the vehicle motion state data.

[0080] The first instruction generation module is used to acquire the target vehicle motion trajectory, and in combination with the target vehicle motion trajectory, discretize the kinematic model of the target vehicle to generate the optimal steering wheel angle instruction based on the kinematic model.

[0081] The instruction sequence acquisition module is used to scale the optimal steering wheel angle instruction by a preset coefficient to obtain the corresponding instruction sequence;

[0082] The neural network model building module is used to acquire historical vehicle state data and build a neural network vehicle model based on the historical vehicle state data. The neural network vehicle model includes an input layer, two hidden layers, and an output layer.

[0083] The model training module is used to input historical vehicle state data into the neural network vehicle model for training and to obtain the final values ​​of the parameters.

[0084] The response signal acquisition module is used to input the instruction sequence into the trained neural network vehicle model and acquire the vehicle response output signal corresponding to the instruction sequence.

[0085] The final steering command determination module is used to calculate the cost value of each command sequence based on each set of candidate steering wheel angle command sequences and their corresponding vehicle response output signals. The minimum cost value is determined by filtering and interpolating the cost values, and the steering wheel angle command corresponding to the minimum cost value is taken as the final steering command.

[0086] It is worth noting that although only some basic functional modules are disclosed in this embodiment, it does not mean that the composition of this system is limited to the above-mentioned basic functional modules. On the contrary, what this embodiment intends to express is that, based on the above-mentioned basic functional modules, those skilled in the art can arbitrarily add one or more functional modules in combination with existing technology to form an infinite number of embodiments or technical solutions. That is to say, this system is open rather than closed. The fact that this embodiment only discloses a few basic functional modules does not mean that the scope of protection of the claims of this invention is limited to the disclosed basic functional modules. At the same time, for the convenience of description, the above device is described separately according to its functions as various units and modules. Of course, in implementing this invention, the functions of each unit and module can be implemented in one or more software and / or hardware.

[0087] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0088] One specific embodiment includes using simple kinematics for model predictive control. The control results obtained from the simple model calculation are corrected by combining a nonlinear neural network vehicle model with the results obtained from the model prediction calculation, thereby achieving high-precision solution results with low computing power consumption.

[0089] The specific implementation steps are as follows:

[0090] First, establish a simple vehicle kinematics model, such as Figure 3 As shown; then, based on this simple vehicle kinematics model, the MPC control algorithm is used to obtain the steering wheel control commands; the obtained control commands are scaled by a certain factor (such as multiplying by 0.8, 0.9, 1, 1.1, and 1.2 respectively) to obtain several command sequences of different sizes; the command sequences are input into the neural network vehicle model to obtain different output sequences; the above input sequences and output sequences are solved using the objective function of MPC to obtain several different cost values; the final steering command is obtained based on the several cost values.

[0091] Without considering tire and suspension elasticity, a vehicle kinematic model is established. The following relationships exist:

[0092]

[0093] Where v is the vehicle's longitudinal velocity, δ is the front wheel angular velocity, θ is the angle between the vehicle's longitudinal axis and the geodetic coordinate system, xy is the geodetic coordinate system, and L is the vehicle's wheelbase.

[0094] Specifically, an MPC control model is established, and the vehicle kinematics model above is discretized based on the desired vehicle trajectory, and expressed as follows:

[0095] (1)

[0096] In this formula, the state variable x represents the difference between the vehicle's motion and the desired vehicle trajectory, and the input variable u represents the required steering wheel angle.

[0097] Rewrite the model as an incremental form of the control variables:

[0098] (2)

[0099] in:

[0100]

[0101]

[0102]

[0103]

[0104]

[0105] Let the model prediction period be The control cycle is , .have

[0106] (3)

[0107]

[0108] (4)

[0109] From the above reasoning, we can conclude that... Always The output of the time-matter model is used for prediction, and can be summarized in the following form:

[0110] (5)

[0111] in:

[0112]

[0113]

[0114]

[0115]

[0116] 2) Scrolling optimization

[0117] The optimization objective can be written in the following form:

[0118] (6)

[0119] In the formula relaxation factor

[0120] set up:

[0121] (7)

[0122] Then equation (6) can be expressed as:

[0123] (8)

[0124] In the formula:

[0125]

[0126]

[0127]

[0128]

[0129] In an MPC controller, consider the constraints of the control variables:

[0130] (9)

[0131] (10)

[0132] Equation (9) represents the row expression of control increment constraints, but equation (10) is the row expression of control constraints. Since equation (8) is the row expression of control increments, equation (10) can be written as the row expression of control increments:

[0133] (11)

[0134] in:

[0135]

[0136]

[0137] , These are the minimum and maximum values ​​of the control quantity.

[0138] Combining equations (8), (9), and (10), we can solve the inequality-constrained optimization problem to obtain the following results.

[0139] (12)

[0140] Will

[0141] (13)

[0142] As the control input (steering wheel angle) of the current system.

[0143] Nonlinear vehicle model establishment based on neural network

[0144] Vehicle models based on neural networks can take many forms. One example is a neural network structure with two hidden layers, each containing 128 neurons. Its inputs include signals such as vehicle speed, steering wheel angle, longitudinal acceleration, driving force, braking force, current yaw rate, current lateral acceleration, and road gradient. The neural network outputs signals such as vehicle yaw rate, lateral acceleration, and sideslip angle.

[0145] The input layer of the neural network contains the following signals:

[0146] The vehicle speed signal for the past two seconds, one signal every 0.5 seconds, for a total of 5 signals;

[0147] The steering wheel angle signal for the past two seconds was one signal every 0.1 seconds, for a total of 21 signals;

[0148] Total vehicle weight signal, 1 signal in total.

[0149] The vehicle acceleration signal over the past two seconds, one signal every 0.2 seconds, for a total of 11 signals;

[0150] The input layer has a total of 38 signals, which pass through two hidden layers to obtain three signals in the output layer.

[0151] The input layer to the hidden layer and the hidden layer to the output layer both use a fully connected structure, meaning that each neuron in the later layer is connected to all neurons in the previous layer, and includes a bias value, which is expressed as follows:

[0152] ;

[0153] ;

[0154] in, , , This represents the output value of the previous neural unit. This represents the input value of a certain neural unit in this layer. This indicates the bias of a specific neural unit in this layer. , , This represents the output values ​​of each neuron in the previous layer up to... The weight, It is the output value of this neuron in this layer, which is used... Use the sigmoid function as the independent variable.

[0155] The above is a neural network model describing the nonlinear vehicle dynamics characteristics.

[0156] By combining data collected from actual vehicles, the parameters (including weight parameters and biases) of the above neural network are trained. The training method adopts the backpropagation method, that is, by calculating the derivative of the error between the neural network output and the actual vehicle signal, the error of the output layer neural unit is calculated, and the error of the hidden layer neural unit is obtained layer by layer. Based on the error of the neural unit, the partial derivatives of each parameter are obtained, and the gradient descent method is used to iteratively obtain the final value of each parameter.

[0157] The obtained neural network parameters can be used to describe the vehicle's dynamic characteristics more accurately.

[0158] Dynamic model correction

[0159] The steering wheel command (Equation 13) obtained by using a kinematic model and an MPC solver is insufficient to represent the optimal steering control command. This is because the actual vehicle response involves many components, and after being disturbed by several dynamic components, its response has deviated significantly from the optimal value.

[0160] To maintain the optimality of the instructions obtained by MPC, this embodiment proposes a solution that amplifies the solution obtained by (Equation 12) by 0.8, 0.9, 1.1, and 1.2 times respectively, resulting in two new solutions:

[0161] (16)

[0162] (17)

[0163] (18)

[0164] (19)

[0165] Equations 16 to 19, and Equation 13 are taken as five hypothetical steering wheel angle commands. Through calculation using a neural network vehicle model, the error value between the actual vehicle response and the desired trajectory can be obtained. Combining the five hypothetical steering wheel angle commands, five different cost values ​​can be calculated using Equation (6). The two smallest cost values ​​are selected for interpolation to obtain the minimum cost value, and the corresponding steering wheel angle command is calculated as the final steering wheel angle command. This command is the optimal solution that comprehensively considers the nonlinear response characteristics of the vehicle.

[0166] Figure 4 This is an electronic device structural block diagram of a vehicle lateral control method based on a nonlinear vehicle model provided by one or more embodiments of the present invention.

[0167] like Figure 4As shown, this application provides an electronic device, including: a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus;

[0168] The memory stores a computer program that, when executed by the processor, causes the processor to perform steps of a vehicle lateral control method based on a nonlinear vehicle model.

[0169] This application also provides a computer-readable storage medium storing a computer program executable by an electronic device, which, when run on the electronic device, causes the electronic device to perform steps of a vehicle lateral control method based on a nonlinear vehicle model.

[0170] For the sake of simplicity, the method embodiments are described as a series of actions. However, those skilled in the art should understand that the embodiments of the present invention are not limited to the described order of actions, because according to the embodiments of the present invention, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions involved are not necessarily essential to the embodiments of the present invention.

[0171] As can be seen from the above description of the embodiments, those skilled in the art can clearly understand that this application can be implemented by means of software plus necessary general-purpose hardware platforms. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in various embodiments or some parts of the embodiments of this application.

[0172] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A vehicle lateral control method based on a nonlinear vehicle model, characterized by, The method comprises the following steps: acquiring vehicle motion state data, and establishing a kinematic model of a target vehicle according to the vehicle motion state data; acquiring a target motion trajectory of the vehicle, discretizing the kinematic model of the target vehicle in combination with the target motion trajectory of the vehicle, and generating an optimal steering wheel angle instruction based on the kinematic model; scaling the optimal steering wheel angle instruction by a preset coefficient to obtain a corresponding instruction sequence; acquiring vehicle historical state data, and constructing a neural network vehicle model based on the vehicle historical state data, wherein the neural network vehicle model comprises an input layer, two hidden layers and an output layer; inputting the vehicle historical state data into the neural network vehicle model for training to obtain final parameter values; inputting the instruction sequence into the trained neural network vehicle model to obtain a vehicle response output signal corresponding to the instruction sequence; based on each set of candidate steering wheel angle instruction sequence and the corresponding vehicle response output signal, calculating the generation cost corresponding to each instruction sequence, and determining the minimum generation cost by screening and interpolation optimization of the generation cost, and taking the steering wheel angle instruction corresponding to the minimum generation cost as the final steering instruction.

2. The vehicle lateral control method based on a nonlinear vehicle model according to claim 1, characterized in that, The vehicle motion state data comprises a vehicle longitudinal speed, a front wheel angle rate, and an angle between a vehicle longitudinal axis and a coordinate system relative to the earth.

3. The vehicle lateral control method based on a nonlinear vehicle model according to claim 1, characterized in that, The vehicle historical state data comprises a historical vehicle speed signal, a historical steering wheel angle signal, a vehicle weight signal and a historical vehicle acceleration signal.

4. The vehicle lateral control method based on a nonlinear vehicle model according to claim 1, characterized in that, The method comprises the following steps: Each neural unit of the input layer to the hidden layer and the hidden layer to the output layer is connected with all neural units of the previous layer and contains a bias amount.

5. The vehicle lateral control method based on a nonlinear vehicle model according to claim 1, characterized in that, The method comprises the following steps: During the training of the neural network vehicle model, a back-propagation error method is used to calculate the error of each neural unit. Specifically, the derivative of the error between the neural network output signal and the real signal collected by the real vehicle is calculated to obtain the error of the output layer neural unit. The error of the output layer neural unit is recursively propagated back layer by layer to obtain the error of each neural unit in the two hidden layers.

6. The vehicle lateral control method based on a nonlinear vehicle model according to claim 1, characterized by, When calculating the generation cost based on each set of candidate steering wheel angle instruction sequence and the corresponding vehicle response output signal, an MPC objective function is used for calculation, and the generation cost represents the deviation degree between the vehicle response corresponding to the candidate steering wheel angle instruction sequence and the expected trajectory.

7. The vehicle lateral control method based on a nonlinear vehicle model according to claim 1, characterized by, The method comprises the following steps: The process of screening and interpolation optimization of the generation cost is as follows: the two smallest generation costs are selected from the generation costs corresponding to all candidate steering wheel angle instruction sequences; The final minimum generation cost is obtained by interpolation calculation of the two minimum generation costs, and the steering wheel angle instruction corresponding to the final minimum generation cost is determined.

8. A vehicle lateral control system based on a nonlinear vehicle model, characterized by, The method comprises the following steps: The method comprises the following steps: The kinematic model establishing module is used to acquire vehicle motion state data, and establish a kinematic model of a target vehicle according to the vehicle motion state data; The first instruction generation module is configured to obtain a target motion trajectory of the vehicle, discretize a kinematic model of the target vehicle in combination with the target motion trajectory of the vehicle, and generate an optimal steering wheel angle instruction based on the kinematic model. The instruction sequence acquisition module is configured to scale the optimal steering wheel angle instruction by a preset coefficient and acquire a corresponding instruction sequence. The neural network model construction module is configured to obtain historical state data of the vehicle, and construct a neural network vehicle model based on the historical state data of the vehicle. The neural network vehicle model comprises an input layer, two hidden layers, and an output layer. The model training module is configured to input the historical state data of the vehicle into the neural network vehicle model for training, and acquire final values of parameters. The response signal acquisition module is configured to input the instruction sequence into the trained neural network vehicle model, and acquire a vehicle response output signal corresponding to the instruction sequence. The final steering instruction determination module is configured to calculate a generation cost corresponding to each candidate steering wheel angle instruction sequence based on the candidate steering wheel angle instruction sequence and the corresponding vehicle response output signal, determine a minimum generation cost by screening and interpolation optimization on the generation costs, and determine a steering wheel angle instruction corresponding to the minimum generation cost as a final steering instruction.

9. An electronic device, comprising: The processor, the communication interface, the memory and the communication bus are in communication with each other. The memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the vehicle lateral control method based on the nonlinear vehicle model according to any one of claims 1-7. The memory stores a computer program executable by the electronic device, and when the computer program runs on the electronic device, the electronic device executes the steps of the vehicle lateral control method based on the nonlinear vehicle model according to any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, ​

Images