Vehicle control method, device, vehicle and storage medium

Abstract

The present application relates to a vehicle control method, device, vehicle and storage medium. The method includes: obtaining a control reference quantity of a preset vehicle control parameter through a linear quadratic solver based on a vehicle feedback matrix and a vehicle state matrix; determining a control sequence containing multiple control quantities according to a first setting rule based on the control reference quantity; traversing the control quantities in the control sequence to obtain the first loss function value of each control quantity corresponding to the prediction time domain; selecting the control quantity corresponding to the first loss function value that meets the preset conditions as the control determination quantity of the preset vehicle control parameter based on the first loss function value of each control quantity in the control sequence corresponding to the prediction time domain; and controlling the vehicle according to the control determination quantity of the preset vehicle control parameter. The solution provided by the present application can avoid the swing and overshoot of the controlled object caused by vehicle control based on a linear quadratic solver.

Description

Technical Field

[0001] The present application relates to the field of autonomous driving technology, and in particular to a vehicle control method, device, vehicle, and storage medium. Background Art

[0002] In related technologies, lateral control for autonomous driving is solved using an LQR (Linear Quadratic Regulator) to obtain a control variable within a fixed time domain. LQR solves over a fixed time domain and has only one optimal solution. However, LQR does not consider the error between the vehicle's state and the planned path at future moments, which can easily cause the autonomous vehicle to oscillate and overshoot. Summary of the Invention

[0003] In order to solve or partially solve the problems existing in the related art, the present application provides a vehicle control method, device, vehicle and storage medium, which can avoid the swing and overshoot of the controlled object caused by vehicle control based on a linear quadratic solver.

[0004] A first aspect of the present application provides a vehicle control method, the method comprising:

[0005] According to the vehicle feedback matrix and the vehicle state matrix, a control reference quantity of the preset vehicle control parameters is obtained through a linear quadratic solver;

[0006] Determining a control sequence including a plurality of control quantities according to a first setting rule based on the control reference quantity;

[0007] Traversing the control quantities in the control sequence, and obtaining a first loss function value of each control quantity corresponding to a prediction time domain;

[0008] According to the first loss function value of each control variable in the control sequence corresponding to the prediction time domain, selecting the control variable corresponding to the first loss function value that meets the preset condition as the control determination variable of the preset vehicle control parameter;

[0009] The vehicle is controlled according to the control determination amount of the preset vehicle control parameter.

[0010] In one embodiment, determining a control sequence including a plurality of control quantities according to a first setting rule based on the control reference quantity includes:

[0011] According to the control reference amount, a plurality of control amounts are selected according to a first set adjustment step, and the control sequence is composed of the plurality of control amounts.

[0012] In one embodiment, traversing the control quantities in the control sequence to obtain a first loss function value of each control quantity corresponding to a prediction time domain includes:

[0013] Determining a plurality of prediction moments in the prediction time domain;

[0014] According to each control variable in the control sequence and the vehicle dynamics model, respectively obtaining the lateral error and the heading error of the vehicle at each prediction moment in the prediction time domain;

[0015] According to the lateral error and the heading error of the vehicle at each prediction moment in the prediction time domain, a first loss function value of each control variable corresponding to the prediction time domain is obtained respectively.

[0016] In one embodiment, before obtaining the control reference value of the preset vehicle control parameter by a linear quadratic solver based on the vehicle feedback matrix and the vehicle state matrix, the method includes:

[0017] Obtaining the current state information of the vehicle, determining a target reference point corresponding to the current state in the vehicle reference trajectory, and obtaining the vehicle state information of the target reference point;

[0018] A vehicle feedback matrix is ​​obtained through a linear quadratic solver according to the vehicle's current state information, the vehicle state information of the target reference point, and the vehicle dynamics model.

[0019] A second aspect of the present application provides another vehicle control method, the method comprising:

[0020] Calculate the vehicle's reference feedback matrix using a linear quadratic solver;

[0021] Determining a feedback matrix sequence including a plurality of feedback matrices according to a second setting rule based on the reference feedback matrix;

[0022] Traversing the feedback matrices in the feedback matrix sequence, and obtaining a second loss function value corresponding to a prediction time domain for each feedback matrix;

[0023] According to the second loss function value of each feedback matrix in the feedback matrix sequence corresponding to the prediction time domain, selecting the feedback matrix corresponding to the second loss function value that meets the preset conditions as the vehicle feedback matrix;

[0024] According to the vehicle feedback matrix and the vehicle state matrix, the control amount of the preset vehicle control parameters is obtained by the linear quadratic solver;

[0025] The vehicle is controlled according to the control amount of the preset vehicle control parameters.

[0026] In one embodiment, determining a feedback matrix sequence including a plurality of feedback matrices according to a second setting rule based on the reference feedback matrix includes:

[0027] Selecting corresponding parameter values ​​for each parameter value of the reference feedback matrix according to a second set adjustment step size, and determining a parameter value series for each parameter value of the reference feedback matrix;

[0028] According to the parameter value series of each parameter value of the reference feedback matrix, a plurality of feedback matrices are determined by permutation and combination, and the plurality of feedback matrices constitute the feedback matrix sequence.

[0029] In one embodiment, traversing the feedback matrices in the feedback matrix sequence to obtain the second loss function value of each feedback matrix corresponding to the prediction time domain includes:

[0030] Determining a plurality of prediction moments in the prediction time domain;

[0031] Obtaining, according to each feedback matrix in the feedback matrix sequence and a vehicle dynamics model, a lateral error and a heading error of the vehicle at each prediction moment in a prediction time domain;

[0032] According to the lateral error and the heading error of the vehicle at each prediction moment in the prediction time domain, a second loss function value of each feedback matrix in the feedback matrix sequence corresponding to the prediction time domain is obtained respectively.

[0033] A third aspect of the present application provides a vehicle control device, the device comprising:

[0034] A first calculation module is used to obtain a control reference value of a preset vehicle control parameter through a linear quadratic solver according to a vehicle feedback matrix and a vehicle state matrix;

[0035] a first determining module, configured to determine a control sequence comprising a plurality of control quantities according to a first setting rule based on the control reference quantity obtained by the first calculating module;

[0036] a second calculation module, configured to traverse the control variables in the control sequence determined by the first determination module, and obtain a first loss function value of each control variable corresponding to a prediction time domain;

[0037] a selection module configured to select, based on the first loss function value corresponding to each control variable in the control sequence obtained by the second calculation module and corresponding to the prediction time domain, a control variable corresponding to the first loss function value that meets a preset condition as the control determination variable of the preset vehicle control parameter;

[0038] A control module is used to control the vehicle according to the control determination amount of the preset vehicle control parameter selected by the selection module.

[0039] A fourth aspect of the present application provides another vehicle control device, the device comprising:

[0040] A matrix calculation module, used to obtain a reference feedback matrix of the vehicle through a linear quadratic solver;

[0041] a matrix sequence module, configured to determine a feedback matrix sequence comprising a plurality of feedback matrices according to a second setting rule based on the reference feedback matrix obtained by the matrix calculation module;

[0042] a loss calculation module, configured to traverse the feedback matrices in the feedback matrix sequence determined by the matrix sequence module, and obtain a second loss function value of each feedback matrix corresponding to a prediction time domain;

[0043] A matrix selection module, configured to select, based on the second loss function value of each feedback matrix in the feedback matrix sequence obtained by the loss calculation module corresponding to the prediction time domain, a feedback matrix corresponding to the second loss function value that meets a preset condition as the vehicle feedback matrix;

[0044] a control quantity calculation module, configured to obtain a control quantity of a preset vehicle control parameter through the linear quadratic solver based on the vehicle feedback matrix and the vehicle state matrix selected by the matrix selection module;

[0045] The vehicle control module is used to control the vehicle according to the control amount of the preset vehicle control parameters obtained by the control amount calculation module.

[0046] A fifth aspect of the present application provides a vehicle, comprising:

[0047] processor; and

[0048] The memory stores executable codes thereon, and when the executable codes are executed by the processor, the processor is caused to execute the method described above.

[0049] In a sixth aspect, the present application provides a computer-readable storage medium having executable code stored thereon. When the executable code is executed by a processor, the processor is caused to execute the method described above.

[0050] The technical solution provided by this application may have the following beneficial effects:

[0051] According to some embodiments of the present application, by serializing the output of the linear quadratic solver, an output series including multiple outputs is obtained; the loss function value of each output in the output series corresponding to the prediction time domain is calculated, and the output corresponding to the loss function value that meets the preset conditions is selected as the final output; by optimizing the output obtained by the linear quadratic solver, the optimal output in the prediction time domain is obtained from the multiple outputs in the output sequence; based on the optimal output, the control quantity of the preset vehicle control parameter is obtained; and the vehicle is controlled based on the control quantity of the preset vehicle control parameter, which can avoid the swing and overshoot of the controlled object caused by the vehicle control based on the linear quadratic solver.

[0052] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] The above and other objects, features and advantages of the present application will become more apparent by describing in more detail exemplary embodiments of the present application in conjunction with the accompanying drawings, wherein the same reference numerals generally represent the same components in the exemplary embodiments of the present application.

[0054] Figure 1 is a flow chart of a vehicle control method according to an embodiment of the present application;

[0055] Figure 2 is a flow chart of a vehicle control method according to another embodiment of the present application;

[0056] Figure 3 is a flow chart of a vehicle control method according to another embodiment of the present application;

[0057] Figure 4 is a flow chart of a vehicle control method according to another embodiment of the present application;

[0058] Figure 5 is a structural diagram of a vehicle control device according to an embodiment of the present application;

[0059] Figure 6 is a structural schematic diagram of a vehicle control device according to another embodiment of the present application;

[0060] Figure 7 It is a structural schematic diagram of a vehicle according to an embodiment of the present application. DETAILED DESCRIPTION

[0061] The following describes embodiments of the present application in more detail with reference to the accompanying drawings. Although the accompanying drawings illustrate embodiments of the present application, it should be understood that the present application can be implemented in various forms and should not be limited by the embodiments described herein. Rather, these embodiments are provided to make the present application more thorough and complete, and to fully convey the scope of the present application to those skilled in the art.

[0062] The terms used in this application are for the purpose of describing specific embodiments only and are not intended to limit this application. As used in this application and the appended claims, the singular forms "a," "an," "the," and "the" are intended to include the plural forms, unless the context clearly indicates otherwise. It should also be understood that the term "and / or" as used herein refers to and encompasses any and all possible combinations of one or more of the associated listed items.

[0063] It should be understood that although the terms "first", "second", "third", etc. may be used in this application to describe various information, this information should not be limited to these terms. These terms are only used to distinguish information of the same type from each other. For example, without departing from the scope of this application, the first information may also be referred to as the second information, and similarly, the second information may also be referred to as the first information. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of this application, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0064] In related technologies, autonomous driving lateral control is solved using LQR, which yields a control variable within a fixed time domain. LQR solves over a fixed time domain, and there is only one optimal solution for a fixed time domain. However, LQR does not consider the error between the vehicle's state at future moments and the planned path, which can easily cause the autonomous vehicle to oscillate and overshoot.

[0065] In response to the above problems, an embodiment of the present application provides a vehicle control method that can avoid swinging and overshooting of an autonomous driving vehicle.

[0066] The technical solutions of the embodiments of the present application are described in detail below with reference to the accompanying drawings.

[0067] Figure 1 It is a flowchart of a vehicle control method according to an embodiment of the present application.

[0068] See also Figure 1 , a vehicle control method, comprising:

[0069] In S110 , a control reference value of a preset vehicle control parameter is obtained through a linear quadratic solver according to the vehicle feedback matrix and the vehicle state matrix.

[0070] In one embodiment, the LQR may calculate a control amount of a preset vehicle control parameter for controlling the vehicle based on the vehicle feedback matrix and the vehicle state matrix, and use the control amount of the preset vehicle control parameter as a control reference amount of the preset vehicle control parameter.

[0071] In S120 , a control sequence including a plurality of control variables is determined according to a first setting rule based on the control reference variable.

[0072] In one embodiment, a first setting rule may be used to select a plurality of corresponding control quantities for a control reference quantity of a preset vehicle control parameter, and the plurality of control quantities may be used to form a control sequence.

[0073] In S130 , the control variables in the control sequence are traversed to obtain the first loss function value of each control variable corresponding to the prediction time domain.

[0074] In one embodiment, for each control quantity in the control sequence, the vehicle state prediction is performed in the prediction time domain, and the error of each control quantity each time the vehicle prediction state is updated in the prediction time domain is calculated; based on the error of each control quantity each time the vehicle prediction state is updated in the prediction time domain, the first loss function value of each control quantity corresponding to the prediction time domain is obtained.

[0075] In S140 , based on the first loss function value of each control variable in the control sequence corresponding to the prediction time domain, the control variable corresponding to the first loss function value that meets the preset conditions is selected as the control determination variable of the preset vehicle control parameter.

[0076] In one embodiment, based on the first loss function value of each control variable in the control sequence corresponding to the prediction time domain, the control variable corresponding to the minimum first loss function value can be determined as the control determination variable of the preset vehicle control parameter.

[0077] In S150 , the vehicle is controlled according to a control determination amount of a preset vehicle control parameter.

[0078] In one embodiment, the vehicle is controlled accordingly according to the determined control determination amount of the preset vehicle control parameter.

[0079] According to an embodiment of the present application, a control reference quantity of a preset vehicle control parameter is obtained through a linear quadratic solver; a control sequence including multiple control quantities is determined according to a first setting rule based on the control reference quantity; based on the first loss function value of each control quantity in the control sequence corresponding to the prediction time domain, the control quantity corresponding to the first loss function value that meets the preset conditions is selected as the control determination quantity of the preset vehicle control parameter; the control reference quantity of the preset vehicle control parameter obtained by the linear quadratic solver can be optimized, the optimal control quantity in the prediction time domain can be obtained from multiple control quantities in the control sequence, and the vehicle can be controlled according to the optimal control quantity of the preset vehicle control parameter, which can avoid the swing and overshoot of the controlled object caused by the vehicle control based on the linear quadratic solver.

[0080] Figure 2 It is a flowchart of a vehicle control method according to another embodiment of the present application. Figure 2 Relative to Figure 1 The technical solution of this application is described in more detail.

[0081] See also Figure 2 , a vehicle control method, comprising:

[0082] In S210 , the vehicle state information at the current moment is obtained, a target reference point corresponding to the current moment in the vehicle reference trajectory is determined, and the vehicle state information of the target reference point is obtained.

[0083] In one embodiment, a reference trajectory is determined based on a path planned by an autonomous vehicle. When planning a path, an autonomous vehicle typically calculates state information such as the vehicle's position, speed, and curvature at each trajectory point. By determining the target reference point corresponding to the current moment in the vehicle's reference trajectory, the vehicle's state information at that target reference point can be obtained. The vehicle's current state information includes, for example, the vehicle's speed and position.

[0084] In some embodiments, after the vehicle planning obtains the path trajectory, a required section is selected from the path trajectory as a reference trajectory based on the delay time caused by the vehicle information processing and transmission process and the prediction time domain of the subsequent prediction algorithm to be used.

[0085] In some embodiments, the target reference point is the point among multiple planned reference points in the reference trajectory that is closest to the vehicle position at the current moment. The vehicle status information of the target reference point includes, for example, part or all of the position, heading, speed, acceleration, and curvature of the reference trajectory of the target reference point.

[0086] In some embodiments, the target reference point may be the point in the reference trajectory curve that is closest to the vehicle position at the current moment. The position and status information of the target reference point may be obtained, for example, by interpolating the position and status information of adjacent planned reference points.

[0087] In S220 , a vehicle feedback matrix is ​​obtained through a linear quadratic solver based on the vehicle's current state information, the vehicle state information at the target reference point, and the vehicle dynamics model.

[0088] In one embodiment, the vehicle feedback matrix K is a 4×1 matrix. The vehicle feedback matrix K[k0, k1, k2, k3] is calculated using LQR based on the vehicle's current state information, the vehicle state information of the target reference point, and the vehicle dynamics model.

[0089] In S230 , a control reference value of a preset vehicle control parameter is obtained through a linear quadratic solver according to the vehicle feedback matrix and the vehicle state matrix.

[0090] In one embodiment, the control reference Cx of the preset vehicle control parameter can be obtained by LQR calculation based on the calculated vehicle feedback matrix K and vehicle state matrix. The preset vehicle control parameter can be, for example, the front wheel angle or the throttle opening.

[0091] In one embodiment, the vehicle state matrix is ​​a 4×1 matrix including lateral error, lateral error change rate, heading error, and heading error change rate, which can be a matrix [lateral error, lateral error change rate, heading error, heading error change rate].

[0092] The lateral error (lat_eror) is the error between the vehicle's current position and the vehicle position of the target reference point corresponding to the current moment in the L direction in the frent coordinate system.

[0093] The heading error (yaw_error) is the error between the vehicle's current heading and the heading of the target reference point corresponding to the current moment in the Frent coordinate system.

[0094] Lateral error change rate = vehicle current speed * heading error.

[0095] Heading error change rate = vehicle heading change rate - expected vehicle heading change rate. The vehicle heading change rate is obtained by the sensor. The expected vehicle heading change rate = expected vehicle speed x curvature.

[0096] In S240 , a plurality of prediction moments in the prediction time domain are determined.

[0097] In one embodiment, multiple prediction moments in the prediction time domain are determined based on the starting point of the prediction time domain and the prediction step size. The prediction step size can be determined based on the length of the prediction time domain and the computing power of the vehicle hardware platform. Longer prediction time domains require smaller prediction step sizes, which in turn increases the robustness of the vehicle hardware platform and the required computing power. It will be appreciated that, in one embodiment, the first prediction moment in the prediction time domain can be the current moment.

[0098] In one embodiment, adjacent prediction moments in the prediction time domain are separated by a prediction step length, and the number of prediction moments included in the prediction time domain is the number of iterative calculations to be performed by the prediction algorithm.

[0099] In S250 , a control sequence including a plurality of control quantities is determined according to a first setting rule based on a control reference quantity of a preset vehicle control parameter.

[0100] In one embodiment, multiple control variables are selected based on a control reference value of a preset vehicle control parameter using a first predetermined adjustment step size, and a control sequence is formed from the multiple control variables. Based on the control reference value Cx of the preset vehicle control parameter output by the LQR, the control reference value Cx is selected using a first predetermined adjustment step size J to obtain multiple control variables, which form a control sequence C[C1, C2, Cx, C3, C4]. The control variables in the control sequence are determined by the control reference value and the first predetermined adjustment step size, where C2 = Cx-J, C1 = Cx-2J, C3 = Cx+J, and C4 = Cx+2J. For example, in one embodiment, if the control reference value Cx of the vehicle's front wheel angle output by the LQR is 0.5 and the first predetermined adjustment step size J is 0.1, the control sequence is C[0.3, 0.4, 0.5, 0.6, 0.7]. Different control sequences C can be obtained according to different values ​​of the first setting adjustment step J. For example, if the control reference Cx is still 0.5 and the first setting adjustment step J is 0.2, the control sequence can be C[0.1, 0.3, 0.5, 0.7, 0.9].

[0101] In S260 , the lateral error and heading error of the vehicle at each prediction moment in the prediction time domain are obtained according to each control variable in the control sequence and the vehicle dynamics model.

[0102] In one embodiment, for each control variable in the control sequence, the lateral error and heading error of the vehicle at each prediction time in the prediction time domain can be obtained based on the vehicle dynamics model. For a control variable Ci in the control sequence, the lateral error and heading error of the vehicle at the prediction time tj in the prediction time domain can be obtained as follows:

[0103] According to the vehicle's current state information and the control variable Ci in the control sequence, the predicted state information of the vehicle at the predicted time tj is obtained through the vehicle dynamics model;

[0104] Determine the target reference point corresponding to the predicted time tj in the vehicle reference trajectory, and obtain the vehicle state information of the target reference point;

[0105] Based on the predicted state information of the vehicle at the predicted time tj and the vehicle state information of the target reference point corresponding to the predicted time tj, the lateral error and heading error of the vehicle at the predicted time tj are calculated.

[0106] In some embodiments, the lateral error lat_error of the vehicle at the predicted time tj is the error between the predicted position of the vehicle at the predicted time tj and the vehicle position of the target reference point corresponding to the predicted time tj in the L direction on the reference trajectory in the frent coordinate system. The heading error yaw_error of the vehicle at the predicted time tj is the error between the predicted heading of the vehicle at the predicted time tj and the vehicle heading of the target reference point corresponding to the predicted time tj in the frent coordinate system on the reference trajectory.

[0107] In S270 , first loss function values ​​of respective control variables corresponding to the prediction time domain are obtained respectively according to the lateral error and heading error of the vehicle at each prediction moment in the prediction time domain.

[0108] In one embodiment, for each control quantity Ci in the control sequence, the first loss function value of the control quantity Ci corresponding to the prediction time domain is obtained based on the lateral error and heading error of the vehicle at each prediction moment in the prediction time domain. The first loss function value is the sum of the prediction errors of the lateral error and heading error at each prediction moment in the prediction time domain.

[0109] According to the lateral error and heading error of the vehicle, the loss function ∑cost is designed:

[0110] cost=Q×lat_error^2+R×yaw_error^2.

[0111] Where cost is the prediction error at the prediction time tj, lat_error is the lateral error at the prediction time tj, yaw_error is the heading error at the prediction time tj, Q is the coefficient of the lateral error, and R is the coefficient of the heading error. Q and R can be adjusted according to the control accuracy. The higher the accuracy requirement for the lateral error and heading error, the larger the coefficient of the lateral error and the coefficient of the heading error can be adjusted accordingly.

[0112] For example, in the control sequence C[C1, C2, Cx, C3, C4], the first set adjustment step J is set between each control variable in the control sequence. The prediction time domain starts at the current time t0 and includes the prediction times t0, t1, t2, t3, and t4. The time length between each time is the prediction step.

[0113] In one embodiment, based on the vehicle state information V0 at the current time t0 and the control variable C1 in the control sequence, the vehicle state information at the predicted time t1, t2, t3, and t4 is predicted through a vehicle dynamics model or a kinematics model; based on the vehicle state information at the predicted time t0, t1, t2, t3, and t4, the lateral error lat_error at each predicted time t0, t1, t2, t3, and t4 is calculated respectively. C1-t0 , lat_error C1-t1 , lat_error C1-t2 , lat_error C1-t3 , lat_error C1-t4 , calculate the heading error yaw_error at each prediction time t0, t1, t2, t3, and t4 respectively C1-t0 , yaw_error C1-t1 , yaw_error C1-t2 , yaw_error C1-t3 , yaw_error C1-t4 ,; According to the lateral error and heading error at each prediction time t0, t1, t2, t3, t4, the prediction error cost at time t0, t1, t2, t3, t4 is calculated according to the loss function C1-t0 、cost C1-t1 、cost C1-t2 、cost C1-t3 、cost C1-t4 ,

[0114] cost C1-t0 =Q×(lat_error C1-t0 )^2+R×(yaw_error C1-t0 )^2,

[0115] cost C1-t1 =Q×(lat_error C1-t1 )^2+R×(yaw_error C1-t1 )^2,

[0116] cost C1-t2 =Q×(lat_error C1-t2 )^2+R×(yaw_error C1-t2 )^2,

[0117] cost C1-t3 =Q×(lat_error C1-t3 )^2+R×(yaw_error C1-t3 )^2,

[0118] cost C1-t4=Q×(lat_error C1-t4 )^2+R×(yaw_error C1-t4 )^2,

[0119] Based on the prediction error cost C1-t0 、cost C1-t1 、cost C1-t2 、cost C1-t3 、cost C1-t4 , calculate the control quantity C1 corresponding to the error and ∑cost in the prediction time domain C1 =cost C1-t0 +cost C1-t1 +cost C1-t2 +cost C1-t3 +cost C1-t4 .

[0120] In S280 , based on the first loss function value of each control variable in the control sequence corresponding to the prediction time domain, the control variable corresponding to the first loss function value that meets the preset conditions is selected as the control determination variable of the preset vehicle control parameter.

[0121] In one embodiment, for each control quantity in the control sequence of the preset vehicle control parameters, the error and ∑cost corresponding to the prediction time domain can be obtained, and the control quantity corresponding to the smallest value of each error and ∑cost can be selected as the control determination quantity of the preset vehicle control parameter.

[0122] In S290 , the vehicle is controlled according to a control determination amount of a preset vehicle control parameter.

[0123] In one embodiment, the vehicle is controlled accordingly based on the determined control determination amount of the preset vehicle control parameter. For example, the front wheel steering angle of the vehicle is controlled based on the control determination amount of the front wheel steering angle.

[0124] According to an embodiment of the present application, a control reference quantity of a preset vehicle control parameter is obtained through a linear quadratic solver; a control sequence including multiple control quantities is determined according to a first setting rule based on the control reference quantity; based on the first loss function value of each control quantity in the control sequence corresponding to the prediction time domain, the control quantity corresponding to the first loss function value that meets the preset conditions is selected as the control determination quantity of the preset vehicle control parameter; the control reference quantity of the preset vehicle control parameter obtained by the linear quadratic solver can be optimized, the optimal control quantity in the prediction time domain can be obtained from multiple control quantities in the control sequence, and the vehicle can be controlled according to the optimal control quantity of the preset vehicle control parameter, which can avoid the swing and overshoot of the controlled object caused by the vehicle control based on the linear quadratic solver.

[0125] Furthermore, in an embodiment of the present application, based on the lateral error and heading error of the vehicle at each prediction moment in the prediction time domain, the first loss function value of each control quantity corresponding to the prediction time domain is obtained respectively. At the first loss function value, different coefficients are set for the lateral error and heading error at each prediction moment in the prediction time domain. By adjusting the coefficient of the heading error or the coefficient of the lateral error, the accuracy of the control quantity of the preset vehicle control parameters in the heading or lateral direction can be adjusted.

[0126] Figure 3 It is a flowchart of a vehicle control method according to another embodiment of the present application.

[0127] See also Figure 3 , a vehicle control method, comprising:

[0128] In S310 , a reference feedback matrix of the vehicle is calculated using a linear quadratic solver.

[0129] In one embodiment, a feedback matrix of the vehicle is obtained by a linear quadratic solver based on the vehicle state information, the vehicle state information of the target reference point, and the vehicle dynamics model, and the feedback matrix is ​​used as the reference feedback matrix of the vehicle.

[0130] In S320, a feedback matrix sequence including a plurality of feedback matrices is determined according to a second setting rule based on the reference feedback matrix.

[0131] In one embodiment, different values ​​are selected for each parameter value according to each parameter value in the reference feedback matrix to obtain multiple feedback matrices, and the multiple feedback matrices form a feedback matrix sequence.

[0132] In S330 , the feedback matrices in the feedback matrix sequence are traversed to obtain the error and the second loss function value corresponding to each feedback matrix in the feedback matrix sequence in the prediction time domain calculated according to the set function.

[0133] In one embodiment, for each feedback matrix in the feedback matrix series, a vehicle state prediction is performed in the prediction time domain, and the error when the vehicle predicted state is updated each time in the prediction time domain is calculated; based on the error when the vehicle predicted state is updated each time in the prediction time domain, a second loss function value corresponding to the prediction time domain of each feedback matrix is ​​obtained.

[0134] In S340, based on the second loss function value corresponding to each feedback matrix in the feedback matrix sequence in the prediction time domain, the feedback matrix corresponding to the second loss function value that meets the preset conditions is selected as the vehicle error sum, and the optimal feedback matrix is ​​determined.

[0135] In one embodiment, based on the second loss function value of each feedback matrix in the feedback matrix sequence corresponding to the prediction time domain, the feedback matrix corresponding to the minimum second loss function value can be determined as the vehicle feedback matrix.

[0136] In S350 , a control variable of a preset vehicle control parameter is obtained through a linear quadratic solver according to the vehicle feedback matrix and the vehicle state matrix.

[0137] In one embodiment, the LQR can calculate the control amount of the preset vehicle control parameters for controlling the vehicle based on the vehicle feedback matrix and the vehicle state matrix.

[0138] In S360 , the vehicle is controlled according to a control amount of a preset vehicle control parameter.

[0139] In one embodiment, the vehicle is controlled accordingly according to a control amount of a preset vehicle control parameter.

[0140] According to an embodiment of the present application, a reference feedback matrix of a vehicle is obtained through a linear quadratic solver; a feedback matrix sequence including multiple feedback matrices is determined according to a second setting rule based on the reference feedback matrix; based on the second loss function value of each feedback matrix in the feedback matrix sequence corresponding to the prediction time domain, a feedback matrix corresponding to the second loss function value that meets the preset conditions is selected as the vehicle feedback matrix; based on the vehicle optimal feedback matrix and the vehicle state matrix, a control amount of a preset vehicle control parameter is obtained through a linear quadratic solver; the vehicle is controlled according to the control amount of the preset vehicle control parameter; the optimal vehicle feedback matrix obtained in the prediction time domain can be obtained from multiple feedback matrices in the feedback matrix sequence by summing the errors of the second loss function values ​​corresponding to the prediction time domain of each feedback matrix in the feedback matrix sequence, and the control amount of the preset vehicle control parameter is obtained according to the optimal vehicle feedback matrix, which can avoid the swing and overshoot of the controlled object caused by the vehicle control based on the linear quadratic solver.

[0141] Figure 4 It is a flowchart of a vehicle control method according to another embodiment of the present application. Figure 4 Relative to Figure 3 The technical solution of this application is described in more detail.

[0142] See also Figure 4 , a vehicle control method, comprising:

[0143] In S410 , the vehicle state information at the current moment is obtained, a target reference point corresponding to the current moment in the vehicle reference trajectory is determined, and the vehicle state information of the target reference point is obtained.

[0144] This step can be referred to the description of S210 and will not be repeated here.

[0145] In S420 , a reference feedback matrix of the vehicle is obtained through a linear quadratic solver according to the vehicle state information at the current moment, the vehicle state information at the target reference point, and the vehicle dynamics model.

[0146] This step can be referred to the description of S220 and S310 and will not be repeated here.

[0147] In S430, a feedback matrix sequence including a plurality of feedback matrices is determined according to a second setting rule based on the reference feedback matrix.

[0148] In one embodiment, corresponding parameter values ​​are selected for each parameter value of the reference feedback matrix according to a second set adjustment step size, and a parameter value series of each parameter value of the reference feedback matrix is ​​determined. Based on the parameter value series of each parameter value of the reference feedback matrix, multiple feedback matrices are determined by permutations and combinations, and the multiple feedback matrices form a feedback matrix sequence.

[0149] In one embodiment, a reference feedback matrix K[k0, k1, k2, k3] is obtained through LQR, and the parameter values ​​of the reference feedback matrix K are k0, k1, k2, k3, respectively. A second setting adjustment step size N is used to select corresponding parameter values ​​for the parameter values ​​k0, k1, k2, and k3, respectively, to form a k0 sequence, a k1 sequence, a k2 sequence, and a k3 sequence, respectively. Different parameter values ​​of the k0 sequence include k0-2N, k0-N, k0, k0+N, and k0+2N. Different parameter values ​​of the k1 sequence include k1-2N, k1-N, k1, k1+N, and k1+2N. Different parameter values ​​of the k2 sequence include k2-2N, k2-N, k2, k2+N, and k2+2N. Different parameter values ​​of the k3 sequence include k3-2N, k3-N, k3, k3+N, and k3+2N. Based on different parameter values ​​in the k0 sequence, different parameter values ​​in the k1 sequence, different parameter values ​​in the k2 sequence, and different parameter values ​​in the k3 sequence, a plurality of feedback matrices including different parameter values ​​of the k0 sequence, the k1 sequence, the k2 sequence, and the k3 sequence are obtained by permutation and combination, and a feedback matrix sequence is formed by the plurality of feedback matrices.

[0150] For example, the reference feedback matrix K[k0, k1, k2, k3] obtained by LQR, the parameter value k0 is 0.01, the parameter value k1 is 0.02, the parameter value k2 is 0.03, the parameter value k3 is 0.04, the second setting adjustment step N is 0.001, the parameter value 0.01 is selected according to the second setting adjustment step 0.001, then the k0 sequence is [0.008, 0.009, 0.01, 0.011, 0.012]; the parameter value 0.02 is selected according to the second setting adjustment step 0. For parameter values ​​of 0.001, the k1 sequence is [0.018, 0.019, 0.02, 0.021, 0.022]. For parameter value 0.03, the corresponding parameter value is selected according to the second setting adjustment step size of 0.001, and the k2 sequence is [0.028, 0.029, 0.03, 0.031, 0.032]. For parameter value 0.04, the corresponding parameter value is selected according to the second setting adjustment step size of 0.001, and the k3 sequence is [0.038, 0.039, 0.04, 0.041, 0.042]. Based on the different parameter values ​​in the k0 sequence, the different parameter values ​​in the k1 sequence, the different parameter values ​​in the k2 sequence, and the different values ​​of the k3 sequence, 625 different feedback matrices Kn are obtained through permutations and combinations, and the feedback matrix sequence is composed of 625 different feedback matrices. For example, the feedback matrices in the feedback matrix sequence include matrices such as K[0.008, 0.018, 0.028, 0.038], K[0.008, 0.018, 0.028, 0.039], K[0.008, 0.02, 0.028, 0.042], K[0.008, 0.022, 0.028, 0.038], and K[0.008, 0.018, 0.032, 0.038].

[0151] In S440 , a plurality of prediction moments in the prediction time domain are determined.

[0152] This step can be referred to the description of S240 and will not be repeated here.

[0153] In S450 , the lateral error and heading error of the vehicle at each prediction moment in the prediction time domain are obtained according to each feedback matrix in the feedback matrix sequence and the vehicle dynamics model.

[0154] In one embodiment, for each feedback matrix in the feedback matrix sequence, the lateral error and heading error of the vehicle at each prediction time in the prediction time domain can be obtained based on the vehicle dynamics model. For a feedback matrix Kn in the feedback matrix sequence, the lateral error and heading error of the vehicle at the prediction time tj in the prediction time domain can be obtained by the following method:

[0155] According to the vehicle's current state information and the feedback matrix Kn in the feedback matrix sequence, the predicted state information of the vehicle at the predicted time tj is obtained through the vehicle dynamics model;

[0156] Determine the target reference point corresponding to the predicted time tj in the vehicle reference trajectory, and obtain the vehicle state information of the target reference point;

[0157] Based on the predicted state information of the vehicle at the predicted time tj and the vehicle state information of the target reference point corresponding to the predicted time tj, the lateral error and heading error of the vehicle at the predicted time tj are calculated.

[0158] In some embodiments, the lateral error lat_error of the vehicle at the predicted time tj is the error between the predicted position of the vehicle at the predicted time tj and the vehicle position of the target reference point corresponding to the predicted time tj in the L direction on the reference trajectory in the frent coordinate system. The heading error yaw_error of the vehicle at the predicted time tj is the error between the predicted heading of the vehicle at the predicted time tj and the vehicle heading of the target reference point corresponding to the predicted time tj in the frent coordinate system on the reference trajectory.

[0159] In S460 , a second loss function value corresponding to the prediction time domain of each feedback matrix in the feedback matrix sequence is obtained according to the lateral error and the heading error of the vehicle at each prediction moment in the prediction time domain.

[0160] In one embodiment, for each feedback matrix Kn in the feedback matrix sequence, a second loss function value corresponding to the prediction time domain of the feedback matrix Kn is obtained based on the lateral error and heading error of the vehicle at each prediction moment in the prediction time domain. The second loss function value is the sum of the prediction errors of the lateral error and heading error at each prediction moment in the prediction time domain.

[0161] According to the lateral error and heading error of the vehicle's predicted state, the loss function ∑cost is designed:

[0162] cost=Q×lat_error^2+R×yaw_error^2.

[0163] Where cost is the prediction error at the prediction time tj, lat_error is the lateral error at the prediction time tj, yaw_error is the heading error at the prediction time tj, Q is the coefficient of the lateral error, and R is the coefficient of the heading error. Q and R can be adjusted according to the control accuracy. The higher the accuracy requirement for the lateral error and heading error, the larger the coefficient of the lateral error and the coefficient of the heading error can be adjusted accordingly.

[0164] For example, the feedback matrix K1[k0, k1, k2, k3] in the feedback matrix sequence, the prediction time domain starts from the current time t0, includes the prediction times t0, t1, t2, t3, t4, and the time length between each prediction time is the prediction step length.

[0165] In one embodiment, based on the vehicle state information V0 at the current time t0 and the feedback matrix K1, the vehicle state information at the predicted time t1, t2, t3, and t4 is predicted through a vehicle dynamics model or a kinematics model; based on the vehicle state information at the time t0, t1, t2, t3, and t4, the lateral error lat_error at each predicted time t0, t1, t2, t3, and t4 is calculated respectively. K1-t0 , lat_error K1-t1 , lat_error K1-t2 , lat_error K1-t3 , lat_error K1-t4 , calculate the heading error yaw_error at each prediction time t0, t1, t2, t3, and t4 respectively K1-t0 , yaw_error K1-t1 , yaw_error K1-t2 , yaw_error K1-t3 , yaw_error K1-t4 ; According to the lateral error and heading error at each time t0, t1, t2, t3, t4, the prediction error cost at time t0, t1, t2, t3, t4 is calculated according to the loss function K1-t0 、cost K1-t1 、cost K1-t2 、cost K1-t3 、cost K1-t4 ,in:

[0166] cost K1-t0 =Q×(lat_error K1-t0 )^2+R×(yaw_error K1-t0 )^2,

[0167] cost K1-t1 =Q×(lat_error K1-t1 )^2+R×(yaw_error K1-t1 )^2,

[0168] cost K1-t2 =Q×(lat_error K1-t2 )^2+R×(yaw_error K1-t2 )^2,

[0169] cost K1-t3=Q×(lat_error K1-t3 )^2+R×(yaw_error K1-t3 )^2,

[0170] cost K1-t4 =Q×(lat_error K1-t4 )^2+R×(yaw_error K1-t4 )^2,

[0171] Based on the prediction error cost K1-t0 、cost K1-t1 、cost K1-t2 、cost K1-t3 、cost K1-t4 , calculate the feedback matrix K1 corresponding to the error and ∑cost of the prediction time domain K1 =cost K1-t0 +cost K1-t1 +cost K1-t2 +

[0172] cost K1-t3 +cost K1-t4 .

[0173] In S470 , based on the second loss function value of each feedback matrix in the feedback matrix sequence corresponding to the prediction time domain, a feedback matrix corresponding to the second loss function value that meets the preset conditions is selected as the vehicle feedback matrix.

[0174] In one embodiment, for each feedback matrix in the feedback matrix sequence, an error and Σcost corresponding to the prediction time domain may be obtained, and the feedback matrix corresponding to the smallest value of the error and Σcost may be selected as the vehicle feedback matrix.

[0175] In S480, a control variable of a preset vehicle control parameter is obtained through a linear quadratic solver according to the vehicle feedback matrix and the vehicle state matrix.

[0176] In one embodiment, the control amount of a preset vehicle control parameter is obtained by LQR calculation based on the vehicle feedback matrix and the vehicle state matrix. The preset vehicle control parameter may be, for example, a front wheel angle or a throttle opening.

[0177] In some embodiments, the vehicle feedback matrix is ​​multiplied by the vehicle state matrix to obtain the control amount of the preset vehicle control parameter.

[0178] In S490, the vehicle is controlled according to a control amount of a preset vehicle control parameter.

[0179] Corresponding to the aforementioned application function implementation method embodiment, the present application also provides a vehicle control device, a vehicle and corresponding embodiments.

[0180] Figure 5 It is a structural diagram of a vehicle control device according to an embodiment of the present application.

[0181] See also Figure 5 A vehicle control device 500 includes a first calculation module 501, a first determination module 502, a second calculation module 503, a selection module 504, and a control module 505.

[0182] The first calculation module 501 is used to obtain a control reference value of a preset vehicle control parameter through a linear quadratic solver according to the vehicle feedback matrix and the vehicle state matrix.

[0183] The first determining module 502 is configured to determine a control sequence including a plurality of control quantities according to a first setting rule based on the control reference quantity obtained by the first calculating module 501 .

[0184] The second calculation module 503 is used to traverse the control variables in the control sequence determined by the first determination module 502 and obtain the first loss function value of each control variable corresponding to the prediction time domain.

[0185] The selection module 504 is used to select the control quantity corresponding to the first loss function value corresponding to the predicted time domain of each control quantity in the control sequence obtained by the second calculation module 503, and select the control quantity corresponding to the first loss function value that meets the preset conditions as the control determination quantity of the preset vehicle control parameter.

[0186] The control module 505 is configured to control the vehicle according to the control determination amount of the preset vehicle control parameter selected by the selection module 504 .

[0187] In one embodiment, the first determining module 502 is further configured to select a plurality of control variables according to the control reference variable and a first set adjustment step size, and the plurality of control variables form a control sequence.

[0188] In one embodiment, the second calculation module 503 is also used to determine multiple prediction moments in the prediction time domain; based on each control quantity in the control sequence and the vehicle dynamics model, the lateral error and heading error of the vehicle at each prediction moment in the prediction time domain are obtained respectively; based on the lateral error and heading error of the vehicle at each prediction moment in the prediction time domain, the first loss function value of each control quantity corresponding to the prediction time domain is obtained respectively.

[0189] In one embodiment, the vehicle control device 500 further includes a first information acquisition module 506 and a matrix acquisition module 507 .

[0190] The first information acquisition module 506 is used to obtain the vehicle's current state information, determine the target reference point corresponding to the current moment in the vehicle reference trajectory, and obtain the vehicle state information of the target reference point.

[0191] The matrix acquisition module 507 is used to obtain the vehicle feedback matrix through a linear quadratic solver based on the vehicle state information at the current moment, the vehicle state information at the target reference point, and the vehicle dynamics model obtained by the first information acquisition module 506.

[0192] Figure 6 It is a structural schematic diagram of a vehicle control device according to another embodiment of the present application.

[0193] See also Figure 6 A vehicle control device 600 includes a matrix calculation module 601, a matrix sequence module 602, a loss calculation module 603, a matrix selection module 604, a control amount calculation module 605, and a vehicle control module 606.

[0194] The matrix calculation module 601 is used to obtain the vehicle's reference feedback matrix through a linear quadratic solver.

[0195] The matrix sequence module 602 is configured to determine a feedback matrix sequence including multiple feedback matrices in the prediction time domain according to a second setting rule based on the reference feedback matrix calculated by the matrix calculation module 601 .

[0196] The error loss calculation module 603 is used to traverse the feedback matrices in the feedback matrix sequence determined by the matrix sequence module 602, and calculate the error and second loss function value corresponding to each feedback matrix in each feedback matrix sequence in the prediction time domain according to the set function.

[0197] The matrix determination selection module 604 is used to determine the optimal feedback matrix as the vehicle feedback matrix based on the error and second loss function value corresponding to each feedback matrix in the feedback matrix sequence calculated by the error loss calculation module 603 in the prediction time domain.

[0198] The control amount calculation module 605 is used to calculate the control amount of the vehicle front wheel angle of the preset vehicle control parameters through a linear quadratic solver based on the vehicle optimal feedback matrix and vehicle state matrix determined by the matrix determination selection module 604.

[0199] The vehicle control module 606 is used to control the vehicle according to the control amount of the preset vehicle control parameters obtained by the control amount calculation module 605.

[0200] In one embodiment, the matrix sequence module 602 is further configured to select corresponding parameter values ​​for each parameter value of the reference feedback matrix obtained by the matrix calculation module 601 according to a second set adjustment step size, and respectively determine a parameter value series for each parameter value of the reference feedback matrix; and determine a plurality of feedback matrices by permutations and combinations based on the parameter value series for each parameter value of the reference feedback matrix, and form a feedback matrix sequence from the plurality of feedback matrices.

[0201] The loss calculation module 603 is also used to determine multiple prediction moments in the prediction time domain; based on each feedback matrix in the feedback matrix sequence and the vehicle dynamics model, the lateral error and heading error of the vehicle at each prediction moment in the prediction time domain are obtained respectively; based on the lateral error and heading error of the vehicle at each prediction moment in the prediction time domain, the second loss function value of each feedback matrix in the feedback matrix sequence corresponding to the prediction time domain is obtained respectively.

[0202] In one embodiment, the vehicle control device 600 further includes a second information acquisition module 607 .

[0203] The second information acquisition module 607 is used to obtain the vehicle's current state information, determine the target reference point corresponding to the current moment in the vehicle reference trajectory, and obtain the vehicle state information of the target reference point.

[0204] The matrix calculation module 601 is also used to obtain the vehicle's reference feedback matrix through a linear quadratic solver based on the vehicle's current state information obtained by the second information acquisition module 607, the vehicle state information of the target reference point, and the vehicle dynamics model.

[0205] Regarding the apparatus in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated again here.

[0206] Figure 7 It is a structural schematic diagram of a vehicle according to an embodiment of the present application.

[0207] See also Figure 7 , vehicle 700 includes a memory 701 and a processor 702 .

[0208] The processor 702 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.

[0209] The memory 701 may include various types of storage units, such as system memory, read-only memory (ROM), and permanent storage devices. ROM may store static data or instructions required by the processor 702 or other modules of the computer. The permanent storage device may be a readable and writable storage device. The permanent storage device may be a non-volatile storage device that does not lose stored instructions and data even after the computer loses power. In some embodiments, the permanent storage device uses a large-capacity storage device (such as a magnetic or optical disk, flash memory) as the permanent storage device. In other embodiments, the permanent storage device may be a removable storage device (such as a floppy disk, optical drive). The system memory may be a readable and writable storage device or a volatile readable and writable storage device, such as dynamic random access memory. The system memory may store some or all instructions and data required by the processor during operation. In addition, the memory 701 may include any combination of computer-readable storage media, including various types of semiconductor memory chips (such as DRAM, SRAM, SDRAM, flash memory, programmable read-only memory), and magnetic disks and / or optical disks may also be used. In some embodiments, the memory 701 may include a readable and / or writable removable storage device, such as a compact disc (CD), a read-only digital versatile disc (e.g., DVD-ROM, double-layer DVD-ROM), a read-only Blu-ray disc, an ultra-density optical disc, a flash memory card (e.g., SD card, mini SD card, Micro-SD card, etc.), a magnetic floppy disk, etc. Computer-readable storage media do not include carrier waves and transient electronic signals transmitted wirelessly or wired.

[0210] The memory 701 stores executable codes. When the executable codes are processed by the processor 702 , the processor 702 may execute part or all of the above-mentioned methods.

[0211] In addition, the method according to the present application may also be implemented as a computer program or a computer program product, which includes computer program code instructions for executing some or all of the steps in the above method of the present application.

[0212] Alternatively, the present application can also be implemented as a computer-readable storage medium (or non-transitory machine-readable storage medium or machine-readable storage medium) on which executable code (or computer program or computer instruction code) is stored. When the executable code (or computer program or computer instruction code) is executed by the processor of a vehicle (or server, etc.), the processor executes part or all of the steps of the above-mentioned method according to the present application.

[0213] The embodiments of the present application have been described above. The above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is selected to best explain the principles of the embodiments, their practical applications, or improvements to the technology in the market, or to enable other persons skilled in the art to understand the embodiments disclosed herein.

Claims

1. A vehicle control method, characterized in that: include: According to the vehicle feedback matrix and the vehicle state matrix, a control reference quantity of the preset vehicle control parameters is obtained through a linear quadratic solver; Determining a control sequence including a plurality of control quantities according to a first setting rule based on the control reference quantity; Traversing the control quantities in the control sequence, and obtaining a first loss function value of each control quantity corresponding to a prediction time domain; According to the first loss function value of each control variable in the control sequence corresponding to the prediction time domain, selecting the control variable corresponding to the first loss function value that meets the preset condition as the control determination variable of the preset vehicle control parameter; The vehicle is controlled according to the control determination amount of the preset vehicle control parameter.

2. The method according to claim 1, characterized in that The determining of a control sequence including a plurality of control quantities according to a first setting rule based on the control reference quantity includes: According to the control reference amount, a plurality of control amounts are selected according to a first set adjustment step, and the control sequence is composed of the plurality of control amounts.

3. The method according to claim 1, characterized in that The traversing the control quantities in the control sequence to obtain the first loss function value of each control quantity corresponding to the prediction time domain includes: Determining a plurality of prediction moments in the prediction time domain; According to each control variable in the control sequence and the vehicle dynamics model, respectively obtaining the lateral error and the heading error of the vehicle at each prediction moment in the prediction time domain; According to the lateral error and the heading error of the vehicle at each prediction moment in the prediction time domain, a first loss function value of each control variable corresponding to the prediction time domain is obtained respectively.

4. The method according to claim 1, wherein Before obtaining the control reference amount of the preset vehicle control parameter by the linear quadratic solver based on the vehicle feedback matrix and the vehicle state matrix, the method includes: obtaining the state information of the vehicle at the current moment, determining the target reference point corresponding to the current moment in the vehicle reference trajectory, and obtaining the vehicle state information of the target reference point; A vehicle feedback matrix is ​​obtained through a linear quadratic solver according to the vehicle's current state information, the vehicle state information of the target reference point, and the vehicle dynamics model.

5. A vehicle control method, characterized in that: include: Calculate the vehicle's reference feedback matrix using a linear quadratic solver; Determining a feedback matrix sequence including a plurality of feedback matrices according to a second setting rule based on the reference feedback matrix; Traversing the feedback matrices in the feedback matrix sequence, and obtaining a second loss function value corresponding to a prediction time domain for each feedback matrix; According to the second loss function value of each feedback matrix in the feedback matrix sequence corresponding to the prediction time domain, selecting the feedback matrix corresponding to the second loss function value that meets the preset conditions as the vehicle feedback matrix; According to the vehicle feedback matrix and the vehicle state matrix, the control amount of the preset vehicle control parameters is obtained by the linear quadratic solver; The vehicle is controlled according to the control amount of the preset vehicle control parameters.

6. The method according to claim 5, characterized in that The step of determining a feedback matrix sequence including a plurality of feedback matrices according to a second setting rule based on the reference feedback matrix includes: Selecting corresponding parameter values ​​for each parameter value of the reference feedback matrix according to a second set adjustment step size, and determining a parameter value series for each parameter value of the reference feedback matrix; According to the parameter value series of each parameter value of the reference feedback matrix, a plurality of feedback matrices are determined by permutation and combination, and the plurality of feedback matrices constitute the feedback matrix sequence.

7. The method according to claim 5, characterized in that The traversing the feedback matrices in the feedback matrix sequence to obtain the second loss function value of each feedback matrix corresponding to the prediction time domain includes: determining a plurality of prediction moments in the prediction time domain; Obtaining, according to each feedback matrix in the feedback matrix sequence and a vehicle dynamics model, a lateral error and a heading error of the vehicle at each prediction moment in a prediction time domain; According to the lateral error and the heading error of the vehicle at each prediction moment in the prediction time domain, a second loss function value of each feedback matrix in the feedback matrix sequence corresponding to the prediction time domain is obtained respectively.

8. A vehicle control device, characterized in that: include: A first calculation module is used to obtain a control reference value of a preset vehicle control parameter through a linear quadratic solver according to a vehicle feedback matrix and a vehicle state matrix; a first determining module, configured to determine a control sequence comprising a plurality of control quantities according to a first setting rule based on the control reference quantity obtained by the first calculating module; a second calculation module, configured to traverse the control variables in the control sequence determined by the first determination module, and obtain a first loss function value of each control variable corresponding to a prediction time domain; a selection module configured to select, based on the first loss function value corresponding to each control variable in the control sequence obtained by the second calculation module and corresponding to the prediction time domain, a control variable corresponding to the first loss function value that meets a preset condition as the control determination variable of the preset vehicle control parameter; A control module is used to control the vehicle according to the control determination amount of the preset vehicle control parameter selected by the selection module.

9. A vehicle control device, characterized in that: include: A matrix calculation module, used to obtain a reference feedback matrix of the vehicle through a linear quadratic solver; a matrix sequence module, configured to determine a feedback matrix sequence comprising a plurality of feedback matrices according to a second setting rule based on the reference feedback matrix obtained by the matrix calculation module; a loss calculation module, configured to traverse the feedback matrices in the feedback matrix sequence determined by the matrix sequence module, and obtain a second loss function value of each feedback matrix corresponding to a prediction time domain; A matrix selection module, configured to select, based on the second loss function value of each feedback matrix in the feedback matrix sequence obtained by the loss calculation module corresponding to the prediction time domain, a feedback matrix corresponding to the second loss function value that meets a preset condition as the vehicle feedback matrix; a control quantity calculation module, configured to obtain a control quantity of a preset vehicle control parameter through the linear quadratic solver based on the vehicle feedback matrix and the vehicle state matrix selected by the matrix selection module; The vehicle control module is used to control the vehicle according to the control amount of the preset vehicle control parameters obtained by the control amount calculation module.

10. A vehicle, characterized in that: include: processor; as well as A memory having executable codes stored thereon, which, when executed by the processor, causes the processor to execute the method according to any one of claims 1 to 7.

11. A computer-readable storage medium, characterized in that Executable codes are stored thereon, and when the executable codes are executed by a processor, the processor is caused to execute the method according to any one of claims 1 to 7.

Images