Vehicle driving control method, device, electronic device and storage medium

Abstract

The present application provides a vehicle driving control method, apparatus, device and computer program product. The method comprises: optimizing the vehicle state parameters of a target vehicle based on a first objective function coupled to a vehicle kinematic model; wherein the vehicle kinematic model includes a mathematical relationship between the vehicle driving control parameters and the vehicle state parameters. Based on a second objective function coupled to the vehicle kinematic model, the vehicle driving control parameters of the target vehicle are optimized with the optimized vehicle state parameters of the target vehicle as constraints. Based on the optimized vehicle driving control parameters of the target vehicle, the target vehicle is controlled. The present application can quickly optimize vehicle driving control parameters for vehicle driving control, thereby meeting the high real-time requirements of unmanned driving.

Description

Technical Field

[0001] The present application relates to the field of unmanned driving, and in particular to a vehicle driving control method, device, equipment and computer program product. Background Art

[0002] Vehicle control is the core of autonomous driving. Existing vehicle control methods directly optimize vehicle control parameters (such as speed and steering angle) based on a linearized vehicle kinematic model during each control cycle. This optimization method often yields significantly different optimal solutions across different control cycles, making it difficult to reuse the currently determined vehicle control parameters in the next control cycle, effectively preventing optimization acceleration for the next control cycle.

[0003] However, unmanned driving has very high real-time requirements for vehicle driving control, and the slow optimization of vehicle driving control parameters is a fatal flaw. Summary of the Invention

[0004] The purpose of this application is to provide a vehicle driving control method, device, equipment and computer program product that can quickly optimize vehicle driving control parameters for vehicle driving control, thereby meeting the high real-time requirements of unmanned driving.

[0005] In order to achieve the above objectives, the embodiments of the present application are implemented as follows:

[0006] In a first aspect, a vehicle driving control method is provided, comprising:

[0007] Optimizing vehicle state parameters of a target vehicle based on a first objective function coupled to a vehicle kinematic model comprising a mathematical relationship between vehicle driving control parameters and vehicle state parameters;

[0008] Based on a second objective function coupled to the vehicle kinematic model, and with the optimized vehicle state parameters of the target vehicle as constraints, optimizing vehicle driving control parameters of the target vehicle;

[0009] The target vehicle is controlled based on the optimized vehicle driving control parameters of the target vehicle.

[0010] In a second aspect, a vehicle driving control device is provided, comprising:

[0011] a first optimization module configured to optimize vehicle state parameters of a target vehicle based on a first objective function coupled to a vehicle kinematic model, wherein the vehicle kinematic model includes a mathematical relationship between vehicle driving control parameters and vehicle state parameters;

[0012] a second optimization module, configured to optimize vehicle driving control parameters of the target vehicle based on a second objective function coupled to the vehicle kinematic model and with the optimized vehicle state parameters of the target vehicle as constraints;

[0013] A driving control module is used to control the driving of the target vehicle based on the optimized vehicle driving control parameters of the target vehicle.

[0014] In a third aspect, an embodiment of the present application provides an electronic device, comprising: a processor; and a memory configured to store computer-executable instructions, wherein the computer-executable instructions, when executed, cause the processor to execute the method described in the first aspect.

[0015] According to a fourth aspect, a computer-readable storage medium is provided, wherein the computer-readable storage medium is used to store computer-executable instructions, and the computer-executable instructions implement the method described in the first aspect when executed by a processor.

[0016] The embodiment of the present application first optimizes the vehicle state parameters of the target vehicle based on a first objective function coupled with the vehicle kinematic model; then, with the optimized vehicle state parameters of the target vehicle as constraints, the vehicle driving control parameters of the target vehicle are optimized based on a second objective function coupled with the vehicle kinematic model. This is equivalent to narrowing the spatial distribution of the vehicle driving control parameters of subsequent control cycles through the optimized vehicle state parameters, thereby quickly determining the optimized vehicle state parameters to meet the high real-time requirements of unmanned driving for vehicle driving control. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments recorded in the embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0018] Figure 1 This is a flow chart of a vehicle driving control method according to an embodiment of the present application.

[0019] Figure 2 Schematic diagram of the first neural network for the vehicle dynamics model in the first objective function.

[0020] Figure 3 Schematic diagram of the neural network for the incremental penalty term of the vehicle control parameters in the first objective function.

[0021] Figure 4Schematic diagram of the neural network for the penalty term of the vehicle driving control parameters in the first objective function.

[0022] Figure 5 Schematic diagram of the neural network for the penalty term of the vehicle state parameter in the first objective function.

[0023] Figure 6 Schematic diagram of a first neural network for the restriction function of the vehicle driving control parameters in the first objective function.

[0024] Figure 7 Schematic diagram of the first neural network for the first objective function.

[0025] Figure 8 Schematic diagram of the second neural network for the first objective function.

[0026] Figure 9 Schematic representation of the multiplication operation for a neural network.

[0027] Figure 10 Schematic diagram of the second neural network for the vehicle dynamics model in the first objective function.

[0028] Figure 11 Schematic diagram of the second neural network for the restriction function of the vehicle driving control parameters in the first objective function.

[0029] Figure 12 Schematic diagram of the first neural network for the second objective function.

[0030] Figure 13 Schematic diagram of the second neural network for the second objective function.

[0031] Figure 14 Schematic diagram of the second neural network for the vehicle dynamics model in the second objective function.

[0032] Figure 15 This is a schematic diagram of the vehicle driving control method according to an embodiment of the present application performing vehicle driving control.

[0033] Figure 16 A schematic structural diagram of a vehicle driving control device according to an embodiment of the present application.

[0034] Figure 17 This is a schematic structural diagram of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0035] In order to enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this application will be clearly and completely described below in conjunction with the drawings in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this specification, not all of the embodiments. Based on the embodiments in this specification, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of this specification.

[0036] An embodiment of the present application provides a vehicle driving control method. Figure 1 FIG. 1 is a flow chart of the vehicle driving control method, comprising the following steps:

[0037] S102 , optimizing vehicle state parameters of the target vehicle based on a first objective function coupled to a vehicle kinematics model; wherein the vehicle kinematics model includes a mathematical relationship between vehicle driving control parameters and vehicle state parameters.

[0038] Specifically, in this embodiment, the optimization objective for optimizing the vehicle state parameters of the target vehicle includes achieving a corresponding desired value for at least one of the target vehicle's driving speed, steering angle, trajectory, and heading angle during the first control cycle. Furthermore, to ensure vehicle driving stability, the optimization objective may further include minimizing the increment of the vehicle driving control parameter within a preset range.

[0039] S104 , based on a second objective function coupled with the vehicle kinematic model, and with the optimized vehicle state parameters of the target vehicle as constraints, optimizing the vehicle driving control parameters of the target vehicle.

[0040] Specifically, the optimization objectives of optimizing the vehicle driving control parameters of the target vehicle in this embodiment include: making at least one of the driving speed, driving steering angle, driving trajectory and driving heading angle of the target vehicle in the second control cycle reach the corresponding expected value; in addition, in order to ensure the stability of vehicle driving, the optimization objectives may also include: minimizing the increment of the vehicle driving control parameters within a preset range.

[0041] It should be noted that in autonomous driving scenarios, to ensure the target vehicle's stable driving, driving control must be performed in each control cycle. Therefore, the vehicle's driving control parameters must be optimized within each control cycle. However, vehicle state parameters, as constraints that narrow the spatial distribution of vehicle driving control parameters, do not need to be optimized in each control cycle.

[0042] In this embodiment, the first control cycle represents a control cycle for optimizing the vehicle state parameters of the target vehicle, and the second control cycle represents a control cycle for optimizing the vehicle driving control parameters of the target vehicle. The second control cycle is no earlier than the first control cycle.

[0043] As an example, the first control cycle can be the same as the second control cycle. That is, if the hardware computing power is sufficient, each control cycle first optimizes the vehicle state parameters, and then optimizes the vehicle driving control parameters based on the optimized vehicle state parameters as constraints.

[0044] Alternatively, the first control cycle and the second control cycle are separated by other control cycles, and the first control cycle is the most recent control cycle before the second control cycle in which the vehicle state parameters of the target vehicle are optimized. That is, if hardware computing power is insufficient, the vehicle state parameters are optimized only during some control cycles, such as every two or every three control cycles. Correspondingly, this embodiment uses the vehicle state parameters obtained from the most recent optimization as a constraint to optimize the vehicle driving control parameters for the current control cycle.

[0045] S106 , performing driving control on the target vehicle based on the optimized vehicle driving control parameters of the target vehicle.

[0046] Specifically, taking the vehicle driving control parameters including driving speed and driving steering angle as an example, this embodiment can control the target vehicle to perform unmanned driving according to the optimized driving speed and driving steering angle.

[0047] To summarize, the method of this embodiment first optimizes the vehicle state parameters of the target vehicle based on the first objective function coupled with the vehicle kinematic model; then, with the optimized vehicle state parameters of the target vehicle as constraints, the vehicle driving control parameters of the target vehicle are optimized based on the second objective function coupled with the vehicle kinematic model. This is equivalent to narrowing the spatial distribution of the vehicle driving control parameters of subsequent control cycles through the optimized vehicle state parameters, thereby quickly determining the optimized vehicle state parameters to meet the high real-time requirements of unmanned driving for vehicle driving control.

[0048] The following describes in detail the application of the method of this embodiment in conjunction with specific implementation methods.

[0049] Here we first introduce the meaning of some mathematical symbols involved in the following text: the m-dimensional vector is represented as For vector , Represents the sum of the absolute values ​​of its elements ,and The activation functions related to neural network construction include: linear function (purelin), radial basis function (radbas), symmetric sigmoid function (tansig), and positive linear function (poslin). The expressions are as follows:

[0050] The vehicle kinematic model of this embodiment can be described as:

[0051]

[0052] in, for System status at the moment; for System input at any time.

[0053] In order to facilitate the subsequent analysis of vehicle state parameters, the above vehicle kinematic model can be converted into the following form:

[0054]

[0055] Based on the above, this embodiment sets up a model predictive controller, which is further composed of a main controller and a sub-controller. The main controller is responsible for optimizing vehicle state parameters; the sub-controller is responsible for optimizing vehicle driving control parameters.

[0056] 1. Introduction of main controller

[0057] For the main controller, the corresponding objective function is the first objective function mentioned above, which can be expressed as:

[0058]

[0059] st

[0060]

[0061]

[0062]

[0063] in, is the increment of vehicle driving control parameters; ; is the control domain of the second objective function, ; for The expected vehicle state parameters at the time are obtained by optimizing the first objective function; for The expected vehicle driving control parameters at time t; is the current vehicle status; is the current vehicle driving control parameter; is the weight of the vehicle driving control parameter; is the weight of the vehicle state parameter; The weight of the increment of the vehicle driving control parameter; is the vehicle driving control parameter The constraint set of is the vehicle state parameter Constraint set; is the increment of vehicle driving control parameters The constraint set.

[0064] Among them, the vehicle control parameters It can be regarded as a part of the vehicle state parameters. With the driving speed and the driving steering angle as input, the main goal of the optimization is to make the target vehicle accurately track the given path while maintaining a certain driving speed. Based on the geometric analysis of vehicle driving, it is known that when the turning curvature is When the vehicle's steering angle is approximately ,in is the turning radius, is the vehicle wheelbase. Given the desired driving path, the desired value of the steering angle can be determined. The non-equality constraints are further converted into soft constraints using penalty functions, and the first objective function can be converted into:

[0065]

[0066] st

[0067]

[0068]

[0069] in, and is the weight constant; ; ; and is called the limit function.

[0070] The above-mentioned first objective function is a 1-norm nonlinear problem. This embodiment can construct a neural network that represents the first objective function, and optimize the vehicle state parameters of the target vehicle in the neural network of the first objective function according to the neural network optimization algorithm.

[0071] As an example introduction, the vehicle dynamics model It can be expressed as Figure 2 The neural network shown. Figure 2In the figure, the neural network on the right is a simplified representation of the neural network on the left; the weights of the black connections are all defaulted to 1 (the following neural networks use the same representation); the system model is the vehicle kinematic model. This network can be based on a discretized nonlinear model It can be obtained by training the input and output data, or by constructing it. Later, we will introduce the method of constructing the corresponding neural network based on the vehicle dynamics model.

[0072] The incremental penalty term of the vehicle control parameter as input in the first objective function can be expressed as Figure 3 The neural network shown. Figure 3 In the figure, the neural network on the right is a simplified representation of the neural network on the left. The neurons in red and green belong to the input layer and output layer respectively. The activation function of the neurons is a linear function (purelin); the weights of the blue connections are the weights of the corresponding input increments. The network takes the increment of the vehicle driving control parameter as input and then multiplies it by the weight This network corresponds to the first objective function .

[0073] The penalty term of the vehicle driving control parameter in the first objective function can be expressed as Figure 4 The neural network shown. Figure 4 In the figure, the neural network on the right is a simplified representation of the neural network on the left. The neurons in red and green belong to the input layer and output layer respectively. The activation function of the neurons is a linear function (purelin); the weights of the blue connections are marked. and expected input As input, then subtract the two, and finally multiply the difference between the two by the weight This network corresponds to the first objective function .

[0074] The penalty term of the vehicle state parameter in the first objective function can be expressed as Figure 5 The neural network shown. Figure 5 In the figure, the neural network on the right is a simplified representation of the neural network on the left. The neurons in red and green belong to the input layer and output layer respectively. The activation function of the neurons is a linear function (purelin); the weights of the blue connections are marked. The network is in state and expected state As input, then subtract the two, and finally multiply the difference between the two by the weight This network corresponds to the first objective function .

[0075] The restriction function of the vehicle driving control parameters in the first objective function can be expressed as Figure 6 The neural network shown. Figure 6 In the figure, the neural network on the right is a simplified representation of the neural network on the left. The neurons in red and green belong to the input layer and output layer respectively. The activation function of the neurons in purple is , the activation function of other neurons is linear function (purelin); the weight of the blue connection is marked. The network is based on the vehicle state parameter As input, then limit the function to output the corresponding value, and finally multiply it by the weight This network corresponds to the first objective function In addition, the form of the neural network expression of the limit function of the increment of the vehicle driving control parameter is the same as this, and will not be repeated here. It should be noted that, in order to clearly express, Figure 6 The activation function of the purple neuron is an abstract representation and will be further specified later.

[0076] Correspondingly, the first objective function can be expressed as Figure 7 Alternatively, when the prediction horizon and the control horizon are large ( ), the first objective function can be equivalent to Figure 8 The neural network shown. Figure 7 and Figure 8 In the example, the state target is the expected value of the vehicle state parameter, the input is the vehicle control parameter, and the input target is the expected value of the vehicle control parameter. The neurons with red and green backgrounds belong to the input layer and output layer respectively. The input data are the states at time 1 and time k respectively. , input at time k-1 , the expected state at time k+1 And the expected input at time k All output data are set to 0 to minimize the first objective function. The activation function for all neurons is linear (excluding neurons in the rectangular module). The weights of red connections are undetermined, the weights of black connections are all defaulted to 1, and the weights of blue connections are -1.

[0077] This embodiment is based on the neural network optimization algorithm, only Figure 7 or Figure 8 The weights corresponding to the red connections are optimized, which is equivalent to the vehicle state parameters of the target vehicle. Thanks to the expressive power of neural networks and their good optimization strategies, the above constrained 1-norm nonlinear optimization problem can be solved quickly, and the first objective function also converges.

[0078] The following introduces the discrete model based on deformation Methods for building neural networks.

[0079] The vehicle kinematic model is decomposed into:

[0080]

[0081]

[0082]

[0083] It is then further converted to:

[0084]

[0085]

[0086]

[0087] Furthermore, the vehicle kinematic model is discretized at small time intervals:

[0088]

[0089]

[0090]

[0091] in, is the front wheel turning angle, is the longitudinal velocity at the center of mass of the vehicle, L is the vehicle wheelbase, is the heading angle of the vehicle, is the time interval.

[0092] Before building a neural network based on the above formula, let’s first introduce how to use neural networks to express multiplication operations. Figure 9 The example neural network can express multiplication operations. Figure 9 The neural network on the right is a simplified representation of the neural network on the left. The neurons in red and green belong to the input layer and output layer respectively. The activation function used by the four neurons in the middle hidden layer is the radial basis function (radbas), while the activation function used by the other neurons is the linear function (purelin). The weights of the sky blue connections are marked and meet the requirements.

[0093] In the above model, the function ,function and function The overall neural network is as follows Figure 10 shown. Figure 10 In the figure, the neurons with red and green backgrounds belong to the input layer and output layer respectively. The input data are the x coordinates of the vehicle , vehicle y coordinate , vehicle heading angle ; The output data are vehicle speed squared , the square of the vehicle speed multiplied by the square of the sine of the turning angle , vehicle y coordinate . Figure 10 The activation function for neurons in the figure is the tangent function (tan(x)), which is indicated on the corresponding neurons. The activation function for all other neurons is the linear function (purelin) (excluding the neurons in the rectangular module). The weights of blue connections are annotated, and the weights of black connections are all defaulted to 1.

[0094] Regarding the specific form of the restriction function and its neural network expression, the input increment and input amount constraints are considered here:

[0095]

[0096]

[0097]

[0098]

[0099] These constraints are formally the same, so here we use the velocity squared input constraint as an example, which is converted to the following penalty function:

[0100]

[0101] Obviously, when the variable satisfies the constraint When , the function output is 0, otherwise the function will output a positive value as a penalty, and the farther the variable is from the constraint range, the greater the value of the function output will be. The neural network expression corresponding to this function is as follows Figure 11 shown. Figure 11 In the figure, the neurons with red and green backgrounds belong to the input layer and output layer respectively. The circles represent neurons, and the hexagons express the bias values ​​of the corresponding neurons. The activation functions of neurons 1 and 3 are linear functions (purelin), and the activation function of neuron 2 is positive linear function (poslin); the weights of the blue connections are marked on the figure, M is a sufficiently large constant, and the weights of the black connections are all 1 by default.

[0102] The vehicle kinematics model, input increment constraint, and input quantity constraint described by the neural network are embedded into the vehicle kinematics model, input increment constraint, and input quantity constraint modules in the objective function neural network. Given the corresponding input and output data, the following solution to the objective function can be obtained after network training:

[0103]

[0104] st

[0105]

[0106]

[0107]

[0108]

[0109]

[0110]

[0111] in, and are the input increments of the vehicle speed squared and the vehicle speed squared multiplied by the square of the tangent of the steering angle, ; , To control the time domain, the optimization variables are the vehicle's x-coordinate sequence and heading angle sequence. and are the penalty weights for state, input, and input increment, respectively. and They are time , Expected values ​​of coordinates and heading angles. is the transformed discretized kinematic model described above. and They are The vehicle state parameters at the moment and The input amount at a time. are parameters related to the input increment and input quantity constraints. The first and second terms in the objective function aim to keep the vehicle speed and steering angle close to the given speed and steering angle. The third through fifth terms aim to ensure that the vehicle's trajectory and heading angle track the reference trajectory and desired heading as closely as possible. The sixth and seventh terms aim to minimize the vehicle's input increment, meaning the change in the control quantity is kept as small as possible, resulting in smoother control.

[0112] 2. Introduction of sub-controller

[0113] The objective function of the sub-controller is the second objective function mentioned above. The optimized vehicle state parameters can be obtained from the optimization of the main controller in the first part. If the relevant hardware computing power is sufficient, the optimization of the main controller can be completed once in each control cycle. Then, according to the obtained state, the first item in the optimization sequence (i.e. ), based on the model , directly calculate the control quantity and act on the system.

[0114] However, when the hardware computing power is insufficient to complete the optimization task of the main controller in each control cycle, each control cycle needs to measure the actual state of the system based on the cycle and the state optimization sequence obtained by the most recent main controller optimization, and optimize the control quantity of the current control cycle based on the algorithm introduced below.

[0115] The form of the second objective function used by the sub-controller is the same as the first objective function of the main controller, but the sub-controller directly optimizes the vehicle driving control parameters. The second objective function can be described as:

[0116]

[0117] st

[0118]

[0119]

[0120]

[0121] in, is the increment of vehicle driving control parameters; ; is the control domain of the second objective function, ; for The expected vehicle state parameters at the time are obtained by optimizing the first objective function; for The expected vehicle driving control parameters at time t; is the current vehicle status; is the current vehicle driving control parameter; is the weight of the vehicle driving control parameter; is the weight of the vehicle state parameter; The weight of the increment of the vehicle driving control parameter; is the vehicle driving control parameter The constraint set of is the vehicle state parameter Constraint set; is the increment of vehicle driving control parameters The constraint set.

[0122] The non-equality constraints are further converted into soft constraints using penalty functions, and the second objective function is converted into:

[0123]

[0124] st

[0125]

[0126]

[0127] in:

[0128] ;

[0129] .

[0130] Similarly, the above-mentioned second objective function is a 1-norm nonlinear problem. This embodiment can construct a neural network to represent the second objective function, and optimize the vehicle state parameters of the target vehicle in the neural network of the first objective function according to the neural network optimization algorithm.

[0131] The second objective function of the neural network can be expressed as Figure 12 Alternatively, when the prediction horizon and control horizon are large ( ) The neural network of the second objective function can be expressed as Figure 13 The neural network shown. Figure 12 and Figure 13 In the figure, the neurons with red and green backgrounds belong to the input layer and output layer respectively. The input data have the states at time 1 and time k respectively. , input at time k-1 , the expected state at time k+1 ; and all output data is set to 0, which means minimizing the second objective function. The activation function of all neurons in the figure is a linear function (purelin) (excluding the neurons in the rectangular module). The weights of the red connections are undetermined, and the weights of the black connections are all defaulted to 1. Similarly, this embodiment only optimizes the weights corresponding to the red connections. This weight corresponds to the vehicle form control parameter, that is, the input increment in the second objective function. Therefore, the weight of the red connection obtained after training the network is the optimal solution of the objective function (when ).

[0132] The following content introduces the method of building a neural network based on the vehicle kinematic model.

[0133] The vehicle kinematic model is decomposed into:

[0134]

[0135]

[0136]

[0137] Where L is the vehicle wheelbase, is the heading angle of the vehicle.

[0138] The vehicle kinematic model can be expressed as Figure 14 The neural network shown. Figure 14 In the figure, the red and green neurons belong to the input layer and output layer respectively. The input data are the longitudinal speed of the vehicle at time k , vehicle steering angle , X coordinate of the vehicle position , Y coordinate of the vehicle position , vehicle heading angle ; The output data are the X coordinates of the k+1 vehicle positions , Y coordinate of the vehicle position and vehicle heading angle The activation functions for the neurons in the figure are sine (sin(x)), cosine (cos(x)), and tangent (tan(x)), which are indicated on the corresponding neurons. The activation functions for all other neurons (excluding the neurons in the rectangular module) are linear functions. The weights of the blue connections are indicated on the figure, while the weights of the black connections are all assumed to be 1 by default.

[0139] The following describes the specific form of the above-mentioned restriction function and its neural network expression, which takes into account the input increment and input amount constraints:

[0140]

[0141]

[0142]

[0143]

[0144] The neural network expressions corresponding to these constraints are the same as those of the main controller, so they will not be repeated here.

[0145] The vehicle kinematic model, input increment constraint, and input quantity constraint described by the neural network are embedded in the system model, input increment restriction, and input quantity restriction modules of the objective function neural network. Given the corresponding input and output data, the following solution to the objective function can be obtained after network training:

[0146]

[0147] st

[0148]

[0149]

[0150]

[0151]

[0152]

[0153]

[0154] in, and are the input increments of the vehicle longitudinal velocity and steering angle, and are also the optimization variables. ; , is the control time domain of the sub-controller, and . and are the penalty weights for state, input, and input increment, respectively. and They are time , The expected values ​​of coordinates and heading angles are optimized by the main controller. This is the kinematic model described above. and are the latest system status and input quantity respectively. These are parameters related to input increment and input quantity constraints.

[0155] 3. Introduction to Model Predictive Controller Application in Vehicle Driving Control

[0156] As shown in Reference 15 , during each control cycle, the model predictive controller calculates the optimized driving steering angle and driving speed at time k based on the desired path, desired heading sequence, and vehicle state information measured and estimated by the signal filtering and fusion processor, such as position, longitudinal velocity, lateral velocity, roll angular velocity, heading angle, and steering angle. These values ​​serve as the desired values ​​for the steering controller and speed controller of the target vehicle. The steering controller then calculates the angle control variable based on the desired driving steering angle and the steering output of the steering system measured in real time, and applies it to the steering system. Similarly, the speed controller calculates the speed control variable based on the desired driving and the speed output of the speed system measured in real time, and applies it to the speed system. This process is then repeated for each control moment.

[0157] In addition, corresponding to Figure 1 In addition to the method shown in FIG. 1 , another embodiment of this embodiment further provides a vehicle driving control device. Figure 16 FIG. 1 is a schematic structural diagram of the vehicle driving control device 1600, comprising:

[0158] a first optimization module 1610 for optimizing vehicle state parameters of a target vehicle based on a first objective function coupled to a vehicle kinematic model comprising a mathematical relationship between vehicle driving control parameters and vehicle state parameters;

[0159] a second optimization module 1620 for optimizing vehicle driving control parameters of the target vehicle based on a second objective function coupled to the vehicle kinematic model and with the optimized vehicle state parameters of the target vehicle as constraints;

[0160] The driving control module 1630 is configured to control the driving of the target vehicle based on the optimized vehicle driving control parameters of the target vehicle.

[0161] Based on the device of this embodiment, considering that the hyperlinks of illegal web pages are phishing links, they generally cannot obtain the hyperlink names of regular websites. Therefore, when it is necessary to determine whether the target web page is an illegal web page, the potential impersonating web page corresponding to the target web page is first determined, and the hyperlink domain name naming rules of the potential impersonating web page are determined based on the hyperlink records of the potential impersonating web page; then, based on whether the hyperlink domain name of the target web page meets the hyperlink domain name naming rules, the target web page is determined to be an illegal web page, and the illegal web page judgment result of the target web page is obtained. Since the entire scheme does not rely on the web page content of the target web page, it can also effectively identify illegal web pages for web pages that are configured with anti-crawling functions and / or have scrambled page content layouts.

[0162] The device of this embodiment first optimizes the vehicle state parameters of the target vehicle based on a first objective function coupled with the vehicle kinematic model; then, with the optimized vehicle state parameters of the target vehicle as constraints, the vehicle driving control parameters of the target vehicle are optimized based on a second objective function coupled with the vehicle kinematic model. This is equivalent to narrowing the spatial distribution of the vehicle driving control parameters of subsequent control cycles through the optimized vehicle state parameters, thereby quickly determining the optimized vehicle state parameters to meet the high real-time requirements of unmanned driving for vehicle driving control.

[0163] Optionally, the optimization goal of optimizing the vehicle state parameters of the target vehicle includes: making at least one of the driving speed, driving steering angle, driving trajectory and driving heading angle of the target vehicle in the first control cycle reach the corresponding expected value; the optimization goal of optimizing the vehicle driving control parameters of the target vehicle includes: making at least one of the driving speed, driving steering angle, driving trajectory and driving heading angle of the target vehicle in the second control cycle reach the corresponding expected value; wherein, the second control cycle is no earlier than the first control cycle.

[0164] Optionally, the first control cycle and the second control cycle are the same control cycle; or, the first control cycle and the second control cycle are separated by other control cycles, and the first control cycle is the most recent control cycle before the second control cycle for optimizing the vehicle state parameters of the target vehicle.

[0165] Optionally, the optimization goal of optimizing the vehicle state parameters and / or vehicle driving control parameters of the target vehicle further includes: minimizing the increment of the vehicle driving control parameters within a preset interval.

[0166] Optionally, the first optimization module 1610 optimizes vehicle state parameters of the target vehicle based on a first objective function coupled to the vehicle kinematic model, including: constructing a neural network representing the first objective function, and optimizing the vehicle state parameters of the target vehicle in the neural network of the first objective function according to a neural network optimization algorithm;

[0167] Optionally, the second optimization module 1620 optimizes the vehicle driving control parameters of the target vehicle based on a second objective function coupled with the vehicle kinematic model, with the optimized vehicle state parameters of the target vehicle as constraints, including: constructing a neural network that represents the second objective function, and according to the neural network optimization algorithm, in the neural network of the second objective function, with the optimized vehicle state parameters of the target vehicle as input parameters, to optimize the vehicle driving control parameters of the target vehicle.

[0168] Optionally, the first objective function is:

[0169] ;

[0170] st

[0171]

[0172] ;

[0173] ;

[0174] ;

[0175] in, is the increment of vehicle driving control parameters; To control the time domain; is the vehicle state parameter used as the optimization variable; for Expected vehicle state parameters at time t; for Expected vehicle driving control parameters at all times; for Vehicle state parameters at the moment; for Vehicle driving control parameters at the moment; is the weight of the vehicle driving control parameter; is the weight of the vehicle state parameter; The weight of the increment of the vehicle driving control parameter; is the vehicle driving control parameter The constraint set of is the vehicle state parameter Constraint set; is the increment of vehicle driving control parameters The constraint set of and is the weight constant.

[0176] Optionally, the second objective function is:

[0177] ;

[0178] st

[0179] ;

[0180] ;

[0181] ;

[0182] ;

[0183] in, is the increment of vehicle driving control parameters; ; is the control domain of the second objective function, ; for The expected vehicle state parameters at the time are obtained by optimizing the first objective function; for The expected vehicle driving control parameters at time t; is the current vehicle status; is the current vehicle driving control parameter; is the weight of the vehicle driving control parameter; is the weight of the vehicle state parameter; The weight of the increment of the vehicle driving control parameter; is the vehicle driving control parameter The constraint set of is the vehicle state parameter Constraint set; is the increment of vehicle driving control parameters The constraint set of and is the weight constant.

[0184] It should be noted that the vehicle driving control device of this embodiment can be used as Figure 1 The execution subject of the method shown can thus realize Figure 1 The steps and functions in the method shown.

[0185] Figure 17 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. Figure 17 At the hardware level, the electronic device includes a processor and, optionally, an internal bus, a network interface, and memory. The memory may include internal memory, such as high-speed random-access memory (RAM), or non-volatile memory, such as at least one disk drive. Of course, the electronic device may also include other hardware required for its services.

[0186] The processor, network interface, and memory can be interconnected via an internal bus, which can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus, or an EISA (Extended Industry Standard Architecture) bus. The bus can be divided into an address bus, a data bus, a control bus, and the like. For ease of representation, Figure 17 Only one bidirectional arrow is used in the diagram, but this does not mean that there is only one bus or one type of bus.

[0187] The memory is used to store computer programs. Specifically, the computer program may include program code, and the program code includes computer operation instructions. The memory may include internal memory and non-volatile memory, and provides the computer program to the processor. The processor reads the corresponding computer program from the non-volatile memory into the internal memory and then runs it, forming the above-mentioned logic layer. Figure 16 The vehicle driving control device shown. Correspondingly, the processor executes the program stored in the memory and is specifically used to perform the following operations:

[0188] The vehicle state parameters of the target vehicle are optimized based on a first objective function coupled to a vehicle kinematic model, wherein the vehicle kinematic model includes a mathematical relationship between vehicle driving control parameters and vehicle state parameters.

[0189] Based on a second objective function coupled to the vehicle kinematic model, the vehicle driving control parameters of the target vehicle are optimized with the optimized vehicle state parameters of the target vehicle as constraints.

[0190] The target vehicle is controlled based on the optimized vehicle driving control parameters of the target vehicle.

[0191] The electronic device of this embodiment first optimizes the vehicle state parameters of the target vehicle based on a first objective function coupled with the vehicle kinematic model; then, with the optimized vehicle state parameters of the target vehicle as constraints, based on a second objective function coupled with the vehicle kinematic model, the vehicle driving control parameters of the target vehicle are optimized. This is equivalent to narrowing the spatial distribution of the vehicle driving control parameters of subsequent control cycles through the optimized vehicle state parameters, thereby quickly determining the optimized vehicle state parameters to meet the high real-time requirements of unmanned driving for vehicle driving control.

[0192] The above is as in this manual Figure 1The methods disclosed in the illustrated embodiments can be applied to and implemented by a processor. The processor may be an integrated circuit chip with signal processing capabilities. During implementation, each step of the above method can be performed by hardware integrated logic circuits within the processor or by software instructions. The above processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. The methods, steps, and logic block diagrams disclosed in the embodiments of this application can be implemented or executed. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of this application can be directly implemented and executed by a hardware decoding processor, or by a combination of hardware and software modules within the decoding processor. The software module can be located in a storage medium well-known in the art, such as random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, etc. The storage medium is located in the memory, and the processor reads the information in the memory and, in conjunction with its hardware, completes the steps of the above method.

[0193] Of course, in addition to software implementation, the electronic device in this specification does not exclude other implementation methods, such as logic devices or a combination of software and hardware, etc. That is to say, the execution subject of the following processing flow is not limited to each logic unit, but can also be hardware or logic devices.

[0194] In addition, the embodiment of the present application also proposes a computer-readable storage medium, which stores one or more computer programs, wherein the one or more computer programs include instructions. When the above instructions are executed by a portable electronic device including multiple applications, the portable electronic device can execute Figure 1 The steps in the method shown include:

[0195] The vehicle state parameters of the target vehicle are optimized based on a first objective function coupled to a vehicle kinematic model, wherein the vehicle kinematic model includes a mathematical relationship between vehicle driving control parameters and vehicle state parameters.

[0196] Based on a second objective function coupled to the vehicle kinematic model, the vehicle driving control parameters of the target vehicle are optimized with the optimized vehicle state parameters of the target vehicle as constraints.

[0197] The target vehicle is controlled based on the optimized vehicle driving control parameters of the target vehicle.

[0198] Those skilled in the art will appreciate that the embodiments of this specification may be provided as methods, systems, or computer program products. Therefore, this specification may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware. Furthermore, this specification may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0199] The foregoing description of this specification describes specific embodiments. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims can be performed in an order different from that described in the embodiments and still achieve the desired results. Furthermore, the processes depicted in the accompanying drawings do not necessarily require the specific order shown or the sequential order to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0200] The above are merely examples of the present invention and are not intended to limit this specification. For those skilled in the art, various modifications and variations of this specification are possible. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of this specification shall be included within the scope of the claims of this specification. In addition, all other embodiments obtained by those of ordinary skill in the art without creative effort shall fall within the scope of protection of this document.

Claims

1. A vehicle driving control method, characterized in that: include: Constructing a neural network representing a first objective function, and optimizing vehicle state parameters of a target vehicle in the neural network of the first objective function according to a neural network optimization algorithm; wherein optimizing the vehicle state parameters of the target vehicle includes causing at least one of a driving speed, a driving steering angle, a driving trajectory, and a driving heading angle of the target vehicle in a first control period to reach a corresponding desired value; wherein the vehicle kinematic model includes a mathematical relationship between the vehicle driving control parameters and the vehicle state parameters; A neural network representing the second objective function is constructed, and according to a neural network optimization algorithm, the vehicle state parameters of the target vehicle after optimization are used as input parameters in the neural network of the second objective function to optimize the vehicle driving control parameters of the target vehicle; wherein the optimization objectives of optimizing the vehicle driving control parameters of the target vehicle include: causing at least one of the driving speed, driving steering angle, driving trajectory, and driving heading angle of the target vehicle in a second control period to reach a corresponding expected value; and the second control period is no earlier than the first control period. Performing driving control on the target vehicle based on the optimized vehicle driving control parameters of the target vehicle; The first objective function is: ； st ； ； ； ； in, is the increment of vehicle driving control parameters; To control the time domain; is the vehicle state parameter used as the optimization variable; for Expected vehicle state parameters at time t; for Expected vehicle driving control parameters at all times; for Vehicle state parameters at the moment; for Vehicle driving control parameters at the moment; is the weight of the vehicle driving control parameter; is the weight of the vehicle state parameter; The weight of the increment of the vehicle driving control parameter; is the vehicle driving control parameter The constraint set of is the vehicle state parameter Constraint set; is the increment of vehicle driving control parameters The constraint set of and is the weight constant; The second objective function is: ； st ； ； ； ； in, is the increment of vehicle driving control parameters; ; is the control domain of the second objective function, ; for The expected vehicle state parameters at the time are obtained by optimizing the first objective function; for The expected vehicle driving control parameters at time t; is the current vehicle status; is the current vehicle driving control parameter; is the weight of the vehicle driving control parameter; is the weight of the vehicle state parameter; The weight of the increment of the vehicle driving control parameter; is the vehicle driving control parameter The constraint set of is the vehicle state parameter Constraint set; is the increment of vehicle driving control parameters The constraint set of and is the weight constant.

2. The method according to claim 1, characterized in that The first control period and the second control period are the same control period; or, The first control cycle is separated from the second control cycle by other control cycles, and the first control cycle is a control cycle that optimizes the vehicle state parameters of the target vehicle most recently before the second control cycle.

3. The method according to claim 2, characterized in that The optimization goal of optimizing the vehicle state parameters and / or vehicle driving control parameters of the target vehicle also includes: minimizing the increment of the vehicle driving control parameters within a preset interval.

4. A vehicle driving control device, characterized in that: include: a first optimization module configured to construct a neural network representing a first objective function and, in accordance with a neural network optimization algorithm, optimize vehicle state parameters of a target vehicle within the neural network of the first objective function; wherein optimizing the vehicle state parameters of the target vehicle includes causing at least one of a driving speed, a driving steering angle, a driving trajectory, and a driving heading angle of the target vehicle in a first control period to reach a corresponding desired value; wherein the vehicle kinematic model includes a mathematical relationship between the vehicle driving control parameters and the vehicle state parameters; a second optimization module, configured to construct a neural network representing the second objective function, and optimize, in accordance with a neural network optimization algorithm, vehicle driving control parameters of the target vehicle using the optimized vehicle state parameters of the target vehicle as input parameters in the neural network of the second objective function; wherein the optimization objectives of optimizing the vehicle driving control parameters of the target vehicle include: causing at least one of the driving speed, driving steering angle, driving trajectory, and driving heading angle of the target vehicle in a second control period to reach a corresponding expected value; and the second control period shall not be earlier than the first control period; a driving control module, configured to control the driving of the target vehicle based on the optimized vehicle driving control parameters of the target vehicle; The first objective function is: ； st ； ； ； ； in, is the increment of vehicle driving control parameters; To control the time domain; is the vehicle state parameter used as the optimization variable; for Expected vehicle state parameters at time t; for Expected vehicle driving control parameters at all times; for Vehicle state parameters at the moment; for Vehicle driving control parameters at the moment; is the weight of the vehicle driving control parameter; is the weight of the vehicle state parameter; The weight of the increment of the vehicle driving control parameter; is the vehicle driving control parameter The constraint set of is the vehicle state parameter Constraint set; is the increment of vehicle driving control parameters The constraint set of and is the weight constant; The second objective function is: ； st ； ； ； ； in, is the increment of vehicle driving control parameters; ; is the control domain of the second objective function, ; for The expected vehicle state parameters at the time are obtained by optimizing the first objective function; for The expected vehicle driving control parameters at time t; is the current vehicle status; is the current vehicle driving control parameter; is the weight of the vehicle driving control parameter; is the weight of the vehicle state parameter; The weight of the increment of the vehicle driving control parameter; is the vehicle driving control parameter The constraint set of is the vehicle state parameter Constraint set; is the increment of vehicle driving control parameters The constraint set of and is the weight constant.

5. An electronic device comprising: processor; and a memory arranged to store computer-executable instructions, wherein the executable instructions, when executed, cause the processor to perform the method according to any one of claims 1 to 3.

6. A computer program product comprising a computer-readable storage medium storing a computer program, characterized in that: The computer program is operable to cause a computer to perform the method according to any one of claims 1 to 3.

Images