Vehicle control method, device and equipment in multiple working condition states and medium

Abstract

The invention provides a vehicle control method, device and equipment in a multi-working-condition state and a medium, and belongs to the field of vehicle control. The method comprises the steps that multiple state parameters of a target vehicle at the current moment are obtained; determining a matching degree weight coefficient of the plurality of state parameters corresponding to the working condition type of the target vehicle; constructing a system state space mapping relation; the optimal control quantity is obtained through the system state space mapping relation; the system state space mapping relationship represents a mapping relationship between a plurality of state parameters of the target vehicle and the control quantity; performing weighted fusion on the matching degree weight coefficient and the system state space mapping relation, and determining the optimal control quantity of the target vehicle at the next moment under different working conditions; and the target vehicle is controlled according to the optimal control quantity, so that the boundary between different working conditions is weakened, the optimization effect of vehicle control is improved, the current working condition of the vehicle can be effectively identified based on the current state parameter of the target vehicle, and optimal control of the current working condition is realized.

Description

Technical Field

[0001] The present invention belongs to the field of vehicle control, and in particular relates to a vehicle control method, device, equipment and medium under multiple working conditions. Background Art

[0002] In the innovation and development of science and technology, in addition to the pursuit of automobile intelligence, people also pay attention to the safety performance of automobiles. As an important part of the development of intelligent automobiles, vehicle stability has attracted much attention from major companies. Although various vehicle auxiliary systems and advanced technologies have promoted the development of vehicle stability, it is still easy to become unstable under extreme conditions such as high speed and low adhesion road surface due to the nonlinear dynamic characteristics of tires, threatening the safety of drivers and passengers. From the perspective of automobile stability and safety performance, if the vehicle stability status can be accurately and effectively obtained in advance and intervened, the vehicle's handling stability and active safety will be effectively improved.

[0003] In terms of vehicle stability control, Model Predictive Control (MPC) has the characteristics of multi-step prediction and rolling optimization. Its prediction ability can accurately solve the problems with multiple constraints and multiple outputs in nonlinear systems. It can not only control the current state, but also predict the state of the system after several cycles, so that the stability of the entire system control is greatly improved. Therefore, it becomes a favorable guarantee for solving vehicle stability problems. However, the system must have stable and unstable states. If MPC control is used throughout the process, on the one hand, the online computing amount of the system will increase and the real-time performance will be reduced. On the other hand, MPC has poor generalization under different working conditions. There may be inaccurate matching of MPC working conditions under different working conditions, resulting in problems in vehicle control.

[0004] Therefore, in the prior art, there are still safety hazards in the stability control of the vehicle. Summary of the invention

[0005] In order to solve the problem of potential safety hazards in vehicle control, the present invention provides a vehicle control method, device, equipment and medium under multiple working conditions.

[0006] In order to achieve the above object, the present invention provides the following technical solutions:

[0007] First, a vehicle control method under multiple working conditions is provided, the method comprising:

[0008] Obtain multiple state parameters of the target vehicle at the current moment;

[0009] The plurality of state parameters are processed by a self-attention mechanism to obtain a plurality of vectorized state parameters; the vectorized state parameters are then subjected to dimensionality reduction processing to determine matching weight coefficients of the plurality of state parameters corresponding to the operating condition type of the target vehicle;

[0010] Based on the stability, energy consumption and maximum motor torque of the target vehicle, a system state space mapping relationship of the target vehicle at the next moment under different working conditions is constructed; the control amount of the target vehicle at the next moment under different working conditions is subjected to rolling optimization through the system state space mapping relationship to obtain the optimal control amount; the system state space mapping relationship represents the mapping relationship between multiple state parameters of the target vehicle and the control amount;

[0011] Performing weighted fusion on the matching degree weight coefficient and the system state space mapping relationship to determine the optimal control amount of the target vehicle at the next moment under different working conditions;

[0012] The target vehicle is controlled according to the optimal control amount.

[0013] Optionally, the operating conditions of the target vehicle include longitudinal operating conditions and lateral operating conditions; the longitudinal operating conditions include acceleration conditions and deceleration conditions; the lateral operating conditions include: low-speed and small-angle conditions, low-speed and large-angle conditions, high-speed and small-angle conditions, and high-speed and large-angle conditions; based on the longitudinal operating conditions and lateral operating conditions, eight operating condition types of the target vehicle are determined.

[0014] Optionally, the multiple state parameters are processed by a self-attention mechanism to obtain multiple vectorized state parameters including:

[0015] Determine the correlation coefficient between different state parameters through the inner product of multiple state parameters;

[0016] Determine the attention score between the state parameter itself and itself and between the state parameter and other state parameters according to the correlation coefficient;

[0017] Normalizing the attention scores to determine an attention weight matrix;

[0018] The row vectors of the inner product matrix of the attention weight matrix and multiple state parameters are connected end to end to obtain multiple vectorized state parameters.

[0019] Optionally, constructing a system state space mapping relationship of the target vehicle at the next moment under different working conditions based on the stability, energy consumption and maximum motor torque of the target vehicle includes:

[0020] Based on the stability, energy consumption and maximum motor torque of the target vehicle, a multi-model MPC controller for the target vehicle under different working conditions is constructed;

[0021] According to the multiple state parameters at the current moment, the operating state of the target vehicle at the next moment is predicted by the multi-model MPC controller;

[0022] According to the multiple state parameters at the current moment and the running state of the target vehicle at the next moment, a system state space mapping relationship of the target vehicle at the next moment is determined.

[0023] Optionally, weighted fusion of the matching degree weight coefficient and the system state space mapping relationship to determine the optimal control amount of the target vehicle at the next moment under different working conditions includes:

[0024] Obtain the product of the matching weight coefficient and the optimal control amount corresponding to multiple state parameters at the current moment;

[0025] The optimal control amount of the target vehicle at the next moment under different working conditions is determined according to the product.

[0026] Secondly, a vehicle control device under multiple working conditions is provided, the device comprising:

[0027] An acquisition module is used to acquire multiple state parameters and operating condition types of the target vehicle at the current moment;

[0028] A weight determination module is used to perform self-attention mechanism processing on the multiple state parameters to obtain multiple vectorized state parameters; then perform dimensionality reduction processing on the vectorized state parameters to determine the matching weight coefficients of the multiple state parameters corresponding to the working condition type;

[0029] A control quantity determination module is used to construct a system state space mapping relationship of the target vehicle at the next moment under different working conditions based on the stability, energy consumption and maximum motor torque of the target vehicle; and to perform rolling optimization on the control quantity of the target vehicle at the next moment under different working conditions through the system state space mapping relationship to obtain the optimal control quantity;

[0030] A fusion module, used for weighted fusion of the matching degree weight coefficient and the system state space mapping relationship to determine the optimal control amount of the target vehicle at the next moment under different working conditions;

[0031] A control module is used to control the target vehicle according to the optimal control amount.

[0032] In addition, a computer-readable storage medium is provided, wherein the storage medium stores a computer program, and when the computer program is executed by a processor, the above-mentioned vehicle control method under multiple operating conditions is implemented.

[0033] Finally, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the vehicle control method under multiple working conditions is implemented.

[0034] The vehicle control method under multiple working conditions provided by the present invention has the following beneficial effects:

[0035] Firstly, the multiple state parameters of the target vehicle at the current moment are processed by the self-attention mechanism, and the correlation between the multiple state parameters is determined, which is beneficial to improve the accuracy of the correspondence between the vehicle state parameters and the working condition type; secondly, by constructing the system state space mapping relationship of the target vehicle under different working conditions, the optimal control quantity of the vehicle is obtained, which is beneficial to improve the optimization effect of vehicle control; and further combining the matching weight coefficient and the optimal control quantity, the boundary between different working conditions is weakened, and the current working condition of the vehicle can be effectively identified based on the current state parameters of the target vehicle, and the optimal control can be achieved according to the current working condition. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the embodiment of the present invention and its design scheme, the following briefly introduces the drawings required for this embodiment. The drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0037] Figure 1 The present invention is a flow chart of a vehicle control method under multiple operating conditions provided by an exemplary embodiment of the present invention.

[0038] Figure 2 A schematic diagram of a self-attention mechanism process provided by the present invention according to an exemplary embodiment.

[0039] Figure 3 A schematic diagram of a neural network structure provided according to an exemplary embodiment of the present invention.

[0040] Figure 4 It is a structural schematic diagram of a weighted fusion process provided by the present invention according to an exemplary embodiment.

[0041] Figure 5 A block diagram of a vehicle control device under multiple operating conditions provided by the present invention according to an exemplary embodiment. DETAILED DESCRIPTION

[0042] In order to enable those skilled in the art to better understand the technical solution of the present invention and implement it, the present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and cannot be used to limit the scope of protection of the present invention.

[0043] The technical solutions provided by various embodiments of the present invention are described in detail below in conjunction with the accompanying drawings.

[0044] First, the present invention provides a vehicle control method under multiple working conditions, specifically as follows: Figure 1 As shown, the following steps are included:

[0045] S101. Acquire multiple state parameters of the target vehicle at the current moment.

[0046] Among them, the multiple state parameters include the longitudinal speed v x , lateral velocity v y , yaw rate γ, wheel slip rate λ i 、Wheel slip angle α i and the optimal slip ratio λ of the four wheels opt .

[0047] S102: Determine matching weight coefficients of multiple state parameters corresponding to the operating condition types.

[0048] Specifically, the multiple state parameters are processed by a self-attention mechanism to obtain multiple vectorized state parameters; the vectorized state parameters are then subjected to dimensionality reduction processing to determine matching weight coefficients of the multiple state parameters corresponding to the working condition types.

[0049] First, with regard to the operating condition type of the target vehicle, the present invention stipulates that the operating condition of the target vehicle includes a longitudinal operating condition and a lateral operating condition; the longitudinal operating condition includes an acceleration condition and a deceleration condition; the lateral operating condition includes: a low-speed and small-angle operating condition, a low-speed and large-angle operating condition, a high-speed and small-angle operating condition, and a high-speed and large-angle operating condition; based on the longitudinal operating condition and the lateral operating condition, eight operating condition types of the target vehicle can be determined.

[0050] In one embodiment, according to the vehicle sub-model state value, the longitudinal operation state of the distributed drive electric vehicle is first divided into three typical operation conditions:

[0051] Acceleration condition: When the slip rate change rates of the four wheels become smaller, the vehicle is in the acceleration condition; Uniform speed condition: When the absolute values ​​of the errors between the slip rates of the four wheels and the optimal slip rate are less than 0.05, the vehicle is in the uniform speed condition; Deceleration condition: When the slip rate change rates of the four wheels become larger, the vehicle is in the deceleration condition.

[0052] The state equation of the distributed drive electric vehicle in the longitudinal operation mode is established, in which the state equations of the vehicle acceleration condition and the uniform speed condition are the same. Therefore, the acceleration condition is combined with the uniform speed condition, so the operation mode of the distributed drive electric vehicle is only divided into the acceleration condition and the deceleration condition. The state equations of the acceleration condition and the deceleration condition are shown in the following formulas:

[0053]

[0054]

[0055] In the formula: i = fl, fr, rl, rr.

[0056] The lateral operation state of distributed drive electric vehicles is divided into four typical operating conditions:

[0057] Low speed and small deflection angle conditions: v x ∈[0,45]km / h and α fl , α fr , α rl , α rr ∈[-0.01,0.01]rad, the distributed drive electric vehicle works under small side slip angle and low speed conditions;

[0058] Low speed and large deflection angle conditions: v x ∈[0,45]km / h and α fl , α fr , α rl , α rr ∈[-0.02,0.01)∪(0.01,0.02]rad, the distributed drive electric vehicle works under large side slip angle and low speed conditions;

[0059] High speed and small deflection angle conditions: v x ∈(45,90]km / h and α fl , α fr , α rl , α rr ∈[-0.01,0.01]rad, the distributed drive electric vehicle works under small side slip angle and high speed conditions;

[0060] High speed and large deflection angle conditions: v x ∈(45,90]km / h and α fl , α fr , α rl , α rr ∈[-0.02,0.01)∪(0.01,0.02]rad, the distributed drive electric vehicle works under large side slip angle and high speed conditions.

[0061] Tire lateral force F under different lateral operation modes y There are different expressions. By calculating the lateral force F in the Magic Formula tire model yi By performing Taylor expansion, we can obtain a cubic polynomial, namely the following formula:

[0062]

[0063] Where: i are fitting coefficients, i=fl,fr,rl,rr.

[0064] When the wheel slip angle α i When the system is in a large deflection angle condition, it is in a nonlinear region. The above formula is consistent with the lateral force F of the MagicFormula tire model. yi Very close; therefore, under large slip angle conditions, the tire lateral force F yi It is expressed using the above formula.

[0065] When the wheel slip angle α i When the system is in a small slip angle condition, it is in the linear region. The following formula is consistent with the lateral force F of the MagicFormula tire model: yi Very close. Thus, the Magic Formula tire model is simplified to obtain the lateral force formula:

[0066] F yi =-2C i α i .

[0067] Under small slip angle conditions, the tire lateral force F yi Use the above formula to express. Combining the above two formulas, we have:

[0068]

[0069] Based on the two longitudinal operating conditions and four lateral operating modes, eight typical operating modes can be obtained, namely: deceleration, low speed and small deflection angle condition; deceleration, low speed and large deflection angle condition; deceleration, high speed and small deflection angle condition; deceleration, high speed and large deflection angle condition; acceleration, low speed and small deflection angle condition; acceleration, low speed and large deflection angle condition; acceleration, high speed and small deflection angle condition; acceleration, high speed and large deflection angle condition.

[0070] In one embodiment, the correlation coefficient between different state parameters is determined by the inner product of multiple state parameters; the attention score between the state parameter itself and between itself and other state parameters is determined based on the correlation coefficient; the attention score is normalized to determine the attention weight matrix; the row vectors of the attention weight matrix and the inner product matrix of the multiple state parameters are connected end to end to obtain multiple vectorized state parameters.

[0071] For example, the vehicle status under various working conditions is classified, and the current working condition of the vehicle can be determined through the real-time data of the vehicle. However, the actual driving process of the vehicle does not completely match the classified working conditions; in addition, only using piecewise functions to achieve multi-working condition classification will cause the control amount of the vehicle to be discontinuous when switching working conditions, that is, overfitting. Therefore, this step will use neural networks for multi-working condition matching to improve versatility and weaken the boundaries of multiple working conditions, and improve the stability when switching working conditions.

[0072] Specifically, for the vehicle state, each parameter is strongly related and not independent. At the same time, the role of each parameter in each working condition is also different. To address this problem, the self-attention mechanism is introduced to calculate the relative importance of each input matrix element, so as to adaptively capture the internal dependencies between parameters. Figure 2 shown.

[0073] By inputting the matrix x' input The inner product of is used to calculate the correlation coefficient between each element to obtain the attention score between the input matrix element itself and itself and between itself and other elements, recorded as matrix x score ; Through the normalization function processing, the attention score is normalized to between 0 and 1, so as to obtain the attention weight matrix; the attention weight matrix and the input matrix are calculated by inner product again to obtain the matrix x after the self-attention mechanism processing attention , then the matrix x attention is a [12×32] matrix. input In addition to expanding the number of features, matrixization also takes into account the use of self-attention mechanism. Matrix input is more convenient for self-attention mechanism processing. However, matrix input requires convolution to extract features. input The matrix is ​​just a different way of expressing the data, so in x' input The internal elements of each row vector in have no direct connection with the internal elements of other row vectors. It is not reasonable to use convolution to extract features, so the subsequent neural network structure is still transmitted in the form of vectors. Finally, the matrix x attention The row vectors of attention .

[0074] The process of self-attention mechanism is described by formula:

[0075] x attention =softmax(x' input (x' input ) T )x' input

[0076] Construct a neural network such as Figure 3 As shown in the figure, the first layer is the input layer, which preprocesses the input data; the second layer is the self-attention mechanism layer, which obtains the attention features of the input matrix and converts the processed matrix x attention Vectorization; the third and fourth layers are hidden layers, which are two fully connected layers. The third fully connected layer is to transform the vector X with a dimension of [1×384] attention The fourth fully connected layer is used to convert the vector Y1 with a dimension of [1×180] into a vector Y2 with a dimension of [1×8]. After each fully connected layer, the ReLU function is used for activation; the fifth layer is the output layer. After normalization, the weight vector Y with a dimension of [1×8] can be obtained. ω .Y ω That is, the matching weight coefficient corresponding to eight typical operating models.

[0077] S103: Determine the optimal control amount of the target vehicle at the next moment under different working conditions.

[0078] Specifically, based on the stability, energy consumption and maximum motor torque of the target vehicle, a system state space mapping relationship of the target vehicle at the next moment under different working conditions is constructed; the control quantity of the target vehicle at the next moment under different working conditions is rolled optimized through the system state space mapping relationship to obtain the optimal control quantity; the system state space mapping relationship represents the mapping relationship between multiple state parameters of the target vehicle and the control quantity.

[0079] For example, based on the stability, energy consumption and maximum motor torque of the target vehicle, a multi-model MPC controller for the target vehicle under different operating conditions is constructed; based on multiple state parameters at the current moment, the operating state of the target vehicle at the next moment is predicted by the multi-model MPC controller; based on the multiple state parameters at the current moment and the operating state of the target vehicle at the next moment, the system state space mapping relationship of the target vehicle at the next moment is determined.

[0080] In one embodiment, the system state space model is first discretized. The state variable x, control signal u and output variable y are defined as:

[0081]

[0082] Using the Euler method to discretize the system state space equation, at time k we have:

[0083]

[0084] Where: f k represents the gradient of the state variable at time k; C M is the identity matrix.

[0085] The difference equation of the discrete state space model is obtained as:

[0086]

[0087] y i (k) = C M x i (k).

[0088] In the formula: i=1,2,3,4; j=fl,fr,rl,rr.

[0089] Define the prediction time domain as T p , the control time domain is T C Then at time k, the predicted output sequence Y(k+1) and the optimal control sequence U(k) are:

[0090]

[0091] Set the optimal slip ratio λ of the four wheels on the road surface opti is the reference input, namely:

[0092]

[0093] The change of the control signal is Δu(k)=u(k)-u(k-1), so the change sequence of the control signal is defined as:

[0094]

[0095] The objective function of MPC is designed in discrete form. To ensure the longitudinal stability of the vehicle, the slip rate λ i It should be stable to prevent longitudinal slippage, that is:

[0096]

[0097] The energy consumption of distributed drive electric vehicles depends on the driving torque of the motor, so the torque of the four motors should be small while achieving the control task, that is:

[0098]

[0099] In order to ensure the stability of distributed drive electric vehicles, the control quantity of the motor should be changed more smoothly. Therefore, the objective function is designed to reduce the change rate of the control signal, that is:

[0100]

[0101] In addition, in actual engineering, it is necessary to prevent excessive drive torque from damaging the motor, so the control amount should not exceed the maximum torque of a single motor. Therefore, the torque control amount of a single motor Tcmdi It must be limited to the maximum torque of the motor, that is:

[0102] 0≤T cmdi (k+j|k)≤T max .

[0103] Formula (46)-Formula (49) are integrated into the objective function of multi-model MPC, that is:

[0104]

[0105] Where: i = fl, fr, rl, rr; j = 0, 1, ..., N c -1; P, Q and R are weight matrices with appropriate dimensions.

[0106] By solving the above formula as an optimization problem, the optimal control sequence in the prediction time domain at time k can be obtained, and the first item u(k|k) in the sequence is input into the drive system as the control target. At time k+1, the optimal control sequence in the prediction time domain at time k+1 is recalculated, and the first item u(k|k+1) in the sequence is input into the drive system as the control target, thus realizing rolling optimization.

[0107] S104, performing weighted fusion on the matching degree weight coefficient and the system state space mapping relationship to determine the optimal control amount of the target vehicle at the next moment under different working conditions.

[0108] In the actual driving process of the vehicle, the driving condition at the next moment is always unknown, but the state equation of the vehicle has been described in differential form, and the driving state of the vehicle at the next moment will be affected by the current driving state of the vehicle. Therefore, the control quantity with a lower matching weight coefficient value cannot be directly discarded.

[0109] Based on this problem, the control quantities under eight typical operating modes are fused through the direct weighted fusion method to maximize the use of data information and effectively achieve the goal of stability when switching between working conditions. Figure 4 As shown, the control quantity under different working conditions is calculated by multi-model MPC:

[0110]

[0111] Where: u i (k|k) is the optimal control quantity of the i-th typical operation mode at time k.

[0112] Obtain the matching weight coefficient Y between the vehicle condition at time k and eight typical motion conditions through the matching neural network ω (k); The optimal control quantity is calculated by weighted fusion:

[0113]

[0114] S105. Control the target vehicle according to the optimal control amount.

[0115] By adopting the above method, firstly, the multiple state parameters of the target vehicle at the current moment are processed by the self-attention mechanism, and the correlation between the multiple state parameters is determined, which is beneficial to improving the accuracy of the correspondence between the vehicle state parameters and the working condition type; secondly, by constructing the system state space mapping relationship of the target vehicle under different working conditions, the optimal control quantity of the vehicle is obtained, which is beneficial to improving the optimization effect of vehicle control; and further combining the matching weight coefficient and the optimal control quantity, the boundary between different working conditions is weakened, and the current working condition of the vehicle can be effectively identified based on the current state parameters of the target vehicle, and the optimal control can be achieved according to the current working condition.

[0116] Secondly, the present invention also provides a vehicle control device under multiple working conditions, such as Figure 5 As shown, including:

[0117] The acquisition module 501 is used to acquire multiple state parameters and operating condition types of the target vehicle at the current moment.

[0118] The weight determination module 502 is used to perform self-attention mechanism processing on the multiple state parameters to obtain multiple vectorized state parameters; then perform dimensionality reduction processing on the vectorized state parameters to determine the matching weight coefficients of the multiple state parameters corresponding to the working condition types.

[0119] The control quantity determination module 503 is used to construct a system state space mapping relationship of the target vehicle at the next moment under different working conditions based on the stability, energy consumption and maximum motor torque of the target vehicle; and to perform rolling optimization on the control quantity of the target vehicle at the next moment under different working conditions through the system state space mapping relationship to obtain the optimal control quantity.

[0120] The fusion module 504 is used to perform weighted fusion on the matching degree weight coefficient and the system state space mapping relationship to determine the optimal control amount of the target vehicle at the next moment under different working conditions.

[0121] The control module 505 is used to control the target vehicle according to the optimal control amount.

[0122] By using the above device, firstly, a self-attention mechanism is used to process multiple state parameters of the target vehicle at the current moment, and the correlation between the multiple state parameters is determined, which is beneficial to improving the accuracy of the correspondence between the vehicle state parameters and the operating condition type; secondly, by constructing a system state space mapping relationship of the target vehicle under different operating conditions, the optimal control amount of the vehicle is obtained, which is beneficial to improving the optimization effect of vehicle control; and further combining the matching weight coefficient and the optimal control amount, the boundary between different operating conditions is weakened, and the current operating condition of the vehicle can be effectively identified based on the current state parameters of the target vehicle, and optimal control can be achieved according to the current operating condition.

[0123] The present invention also provides a computer-readable storage medium, which stores a computer program, which can be used to execute the above Figure 1 The steps of a vehicle control method under multiple operating conditions are provided.

[0124] The present invention also provides a computer device. At the hardware level, the computer device includes a processor, an internal bus, a network interface, a memory, and a non-volatile memory. Of course, it may also include hardware required for other services. The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs it to achieve the above Figure 1 The steps of a vehicle control method under multiple operating conditions are provided.

[0125] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0126] The present invention is described with reference to flowcharts and / or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as a combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0127] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0128] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0129] It should be noted that the above specific implementation method can enable those skilled in the art to understand the invention more comprehensively, but does not limit the invention in any way. Therefore, although the present invention has been described in detail in this specification, those skilled in the art should understand that the invention can still be modified or replaced by equivalents; and all technical solutions and improvements that do not deviate from the spirit and scope of the invention are included in the protection scope of the patent for the invention. Any figure mark in the claims should not be regarded as limiting the claims involved.

Claims

1. A vehicle control method under multiple working conditions, characterized in that: The method comprises: Obtain multiple state parameters of the target vehicle at the current moment; The plurality of state parameters are processed by a self-attention mechanism to obtain a plurality of vectorized state parameters; the vectorized state parameters are then subjected to dimensionality reduction processing to determine matching weight coefficients of the plurality of state parameters corresponding to the operating condition type of the target vehicle; Based on the stability, energy consumption and maximum motor torque of the target vehicle, a system state space mapping relationship of the target vehicle at the next moment under different working conditions is constructed; the control amount of the target vehicle at the next moment under different working conditions is subjected to rolling optimization through the system state space mapping relationship to obtain the optimal control amount; the system state space mapping relationship represents the mapping relationship between multiple state parameters of the target vehicle and the control amount; Performing weighted fusion on the matching degree weight coefficient and the system state space mapping relationship to determine the optimal control amount of the target vehicle at the next moment under different working conditions; The target vehicle is controlled according to the optimal control amount.

2. The vehicle control method under multiple working conditions according to claim 1, characterized in that: The operating conditions of the target vehicle include longitudinal operating conditions and lateral operating conditions; The longitudinal operating condition includes an acceleration condition and a deceleration condition; The lateral operating conditions include: low speed and small deflection angle conditions, low speed and large deflection angle conditions, high speed and small deflection angle conditions, and high speed and large deflection angle conditions; Based on the longitudinal operating condition and the lateral operating condition, eight operating condition types of the target vehicle are determined.

3. The vehicle control method under multiple working conditions according to claim 1, characterized in that: The multiple state parameters are processed by a self-attention mechanism to obtain multiple vectorized state parameters including: Determine the correlation coefficient between different state parameters through the inner product of multiple state parameters; Determine the attention score between the state parameter itself and itself and between the state parameter and other state parameters according to the correlation coefficient; Normalizing the attention scores to determine an attention weight matrix; The row vectors of the inner product matrix of the attention weight matrix and multiple state parameters are connected end to end to obtain multiple vectorized state parameters.

4. The vehicle control method under multiple operating conditions according to claim 1, characterized in that: The system state space mapping relationship of the target vehicle at the next moment under different working conditions is constructed based on the stability, energy consumption and maximum motor torque of the target vehicle, including: Based on the stability, energy consumption and maximum motor torque of the target vehicle, a multi-model MPC controller for the target vehicle under different working conditions is constructed; According to the multiple state parameters at the current moment, the operating state of the target vehicle at the next moment is predicted by the multi-model MPC controller; According to the multiple state parameters at the current moment and the running state of the target vehicle at the next moment, a system state space mapping relationship of the target vehicle at the next moment is determined.

5. The vehicle control method under multiple working conditions according to claim 1, characterized in that: The matching degree weight coefficient and the system state space mapping relationship are weightedly integrated to determine the optimal control amount of the target vehicle at the next moment under different working conditions, including: Obtain the product of the matching weight coefficient and the optimal control amount corresponding to multiple state parameters at the current moment; The optimal control amount of the target vehicle at the next moment under different working conditions is determined according to the product.

6. A vehicle control device under multiple working conditions, characterized in that: The device comprises: An acquisition module is used to acquire multiple state parameters and operating condition types of the target vehicle at the current moment; A weight determination module is used to perform self-attention mechanism processing on the multiple state parameters to obtain multiple vectorized state parameters; then perform dimensionality reduction processing on the vectorized state parameters to determine the matching weight coefficients of the multiple state parameters corresponding to the working condition type; A control quantity determination module is used to construct a system state space mapping relationship of the target vehicle at the next moment under different working conditions based on the stability, energy consumption and maximum motor torque of the target vehicle; and to perform rolling optimization on the control quantity of the target vehicle at the next moment under different working conditions through the system state space mapping relationship to obtain the optimal control quantity; A fusion module, used for weighted fusion of the matching degree weight coefficient and the system state space mapping relationship to determine the optimal control amount of the target vehicle at the next moment under different working conditions; A control module is used to control the target vehicle according to the optimal control amount.

7. A computer-readable storage medium, characterized in that: The storage medium stores a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 5 is implemented.

8. A computer device, characterized in that: The method comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor implements the method according to any one of claims 1 to 5 when executing the program.

Images