Vehicle control method, control device, and vehicle

Abstract

The application discloses a vehicle control method, device and vehicle, and belongs to the technical field of vehicle control. The application obtains vehicle multi-source information for representing the operation behavior of a driver and the running state of a vehicle in a reference historical time period, determines the expected acceleration of the driver in a reference future time period according to the vehicle multi-source information in the reference historical time period, determines the target driving mode of the vehicle pre-switching according to the expected acceleration in the reference future time period, and controls the driving mode of the vehicle to switch to the target driving mode. In this way, the expected acceleration of the driver is determined in combination with the operation behavior of the driver and the running state of the vehicle in the historical time period, the driving intention of the driver can be accurately recognized, misjudgment caused by slight operation of the driver unconsciously or short-term bumping of a road surface can be avoided, and the accuracy of mode switching is ensured.

Description

Technical Field

[0001] This application relates to the field of vehicle control technology, and in particular to a vehicle control method, control device, and vehicle. Background Technology

[0002] With the continuous evolution of automotive intelligent technology, various driver assistance functions have been widely installed in mass-produced models and are developing towards multi-mode collaborative control and seamless switching between modes.

[0003] In related technologies, the mode control system uses simple driving action thresholds or steering wheel torque intervention signals as the basis for determining driving mode switching, and controls the switching of vehicle driving modes accordingly. However, it is easily affected by unintentional operation or instantaneous action of the driver, which can lead to erroneous switching of driving modes. Summary of the Invention

[0004] This application provides a vehicle control method, control device, and vehicle, which can be used to solve problems existing in related technologies. The technical solution is as follows: On one hand, embodiments of this application provide a vehicle control method, the method comprising: Obtain multi-source vehicle information for a reference historical time period, wherein the multi-source vehicle information is used to characterize the driver's operating behavior and the vehicle's operating status; Based on the multi-source vehicle information from the reference historical time period, determine the driver's expected acceleration in the reference future time period; Based on the desired acceleration, determine the target driving mode to be switched to by the vehicle; Control the vehicle's driving mode to switch to the target driving mode.

[0005] Optionally, determining the target driving mode to be switched by the vehicle based on the desired acceleration includes: If the expected acceleration in the reference future time period satisfies the first reference condition and the current driving mode of the vehicle is the automatic driving mode, the target driving mode is determined to be the manual driving mode. If the expected acceleration during the reference future time period satisfies the second reference condition and the current driving mode of the vehicle is manual driving mode, the target driving mode is determined to be automatic driving mode.

[0006] Optionally, the first reference condition includes: within the reference future time period, the duration for which the absolute value of the expected acceleration is continuously greater than a first threshold reaches a first duration threshold. The second reference condition includes: within the reference future time period, the duration for which the absolute value of the expected acceleration is continuously less than a second threshold reaches a second duration threshold.

[0007] Optionally, before determining that the target driving mode is an autonomous driving mode, the method further includes: Obtain the steering wheel grip state of the vehicle; Determining the target driving mode as an autonomous driving mode includes: When the steering wheel grip status of the vehicle indicates that the steering wheel is being gripped, the target driving mode is determined to be the automatic driving mode.

[0008] Optionally, switching the driving mode of the vehicle to the target driving mode includes: If the target driving mode is determined to be the same in multiple consecutive tests, the vehicle's driving mode is switched to the target driving mode.

[0009] Optionally, if the target driving mode is an automatic driving mode, the method further includes: Based on the reference vehicle speed, the vehicle's current actual speed, the relative distance and relative speed with the vehicle in front, the target acceleration of the vehicle is obtained using a model predictive control algorithm. The target execution torque of the vehicle is determined based on the target acceleration; The vehicle is controlled according to the target torque.

[0010] Optionally, determining the target execution torque of the vehicle based on the target acceleration includes: Based on a reference acceleration change rate threshold, the target acceleration is smoothed to obtain a smoothed target acceleration. The target execution torque of the vehicle is determined based on the target acceleration after the smoothing process.

[0011] Optionally, controlling the vehicle based on the target torque includes: When a switch between driving and braking states is detected, the torque weighting coefficient is obtained; The target braking torque and the target driving torque are determined based on the torque weighting coefficient and the target execution torque. The braking system of the vehicle is controlled according to the target braking torque, and the driving system of the vehicle is controlled according to the target driving torque.

[0012] On the other hand, a control device is provided, which includes a signal acquisition module and a control module; The signal acquisition module is used to acquire multi-source vehicle information over a reference historical time period. The multi-source vehicle information is used to characterize the driver's operating behavior and the vehicle's operating status. The control module is used to determine the driver's expected acceleration in a reference future time period based on the vehicle's multi-source information from the reference historical time period; determine the target driving mode to be switched to by the vehicle based on the expected acceleration; and control the vehicle's driving mode to switch to the target driving mode.

[0013] Optionally, the control module includes an intent recognition unit, which is used to determine the driver's expected acceleration in a reference future time period based on the vehicle multi-source information of the reference historical time period through an intent recognition model.

[0014] Optionally, the control module further includes a switching determination unit, which is configured to: determine the target driving mode as manual driving mode when the expected acceleration in the reference future time period meets a first reference condition and the current driving mode of the vehicle is automatic driving mode; and determine the target driving mode as automatic driving mode when the expected acceleration in the reference future time period meets a second reference condition and the current driving mode of the vehicle is manual driving mode.

[0015] Optionally, the first reference condition includes: within the reference future time period, the duration for which the absolute value of the expected acceleration is continuously greater than a first threshold reaches a first duration threshold. The second reference condition includes: within the reference future time period, the duration for which the absolute value of the expected acceleration is continuously less than a second threshold reaches a second duration threshold.

[0016] Optionally, before determining the target driving mode as an autonomous driving mode, the switching determination unit is further configured to: obtain the grip state of the vehicle's steering wheel; and determine the target driving mode as an autonomous driving mode when the grip state of the vehicle's steering wheel indicates that the vehicle's steering wheel is being gripped.

[0017] Optionally, the switching determination unit is used to: control the vehicle's driving mode to switch to the target driving mode when the target driving mode is determined to be the same in multiple consecutive determinations.

[0018] Optionally, the control device further includes an execution module, and the control module further includes a control decision unit. When the target driving mode is an automatic driving mode, the control decision unit is used to: obtain the target acceleration of the vehicle based on a model predictive control algorithm according to a reference vehicle speed, the current actual vehicle speed, the relative distance and relative speed with the vehicle in front; determine the target execution torque of the vehicle based on the target acceleration; and the execution module is used to control the vehicle according to the target execution torque.

[0019] Optionally, the control decision unit is configured to: smooth the target acceleration based on a reference acceleration change rate threshold to obtain a smoothed target acceleration; and determine the target execution torque of the vehicle based on the smoothed target acceleration.

[0020] Optionally, the execution module is configured to: acquire a torque weighting coefficient when it detects that the vehicle is switching between a driving state and a braking state; determine a target braking torque and a target driving torque based on the torque weighting coefficient and the target driving torque; control the vehicle's braking system based on the target braking torque; and control the vehicle's driving system based on the target driving torque.

[0021] Optionally, the control device further includes a prompting module, which outputs a switching prompt message before the vehicle's driving mode is switched to the target driving mode, to prompt the driver that the vehicle is about to switch modes.

[0022] Optionally, the control device further includes a storage control module and a data storage unit. The storage control module is used to store multi-source vehicle information, intent recognition results, and mode switching event information obtained during vehicle operation into the data storage unit.

[0023] On the other hand, a vehicle is provided, the vehicle including a control device and a vehicle body, the control device being configured to execute the operation steps of the method corresponding to any of the optional implementations of the first aspect, so as to control the vehicle body.

[0024] On the other hand, a computer device is provided, the computer device including a processor and a memory, the memory storing at least one computer program, the at least one computer program being loaded and executed by the processor to enable the computer device to implement the operation steps of the method corresponding to any of the optional implementations of the first aspect.

[0025] On the other hand, a computer-readable storage medium is also provided, wherein at least one computer program is stored in the computer-readable storage medium, the at least one computer program being loaded and executed by a processor to enable a computer to implement the operation steps of the method corresponding to any of the optional implementations of the first aspect.

[0026] On the other hand, a computer program product or computer program is also provided, the computer program product or computer program including computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, causing the computer device to perform the operational steps of the method corresponding to any of the optional implementations of the first aspect described above.

[0027] The technical solution provided in this application has at least the following beneficial effects: By combining driver behavior and vehicle operating status over historical time periods, the system determines the driver's desired acceleration, accurately identifying the driver's driving intentions and avoiding misjudgments caused by unintentional slight driver input or brief road disturbances. Furthermore, based on continuous and precise desired acceleration, the system determines the target driving mode to switch to, effectively filtering out transient interference signals, ensuring the accuracy of the target driving mode, and avoiding frequent and unnecessary switching between different modes, significantly improving the stability of mode switching. Attached Figure Description

[0028] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0029] Figure 1 This is a schematic diagram of the structure of the first type of control device provided in the embodiments of this application; Figure 2 This is a schematic flowchart of the first vehicle control method provided in the embodiments of this application; Figure 3 This is a schematic flowchart of the second vehicle control method provided in the embodiments of this application; Figure 4 This is a schematic diagram of the structure of the second type of control device provided in the embodiments of this application; Figure 5 This is a schematic diagram of a vehicle structure provided in an embodiment of this application. Detailed Implementation

[0030] To make the objectives, technical solutions, and advantages of this application clearer, the embodiments of this application will be described in further detail below with reference to the accompanying drawings.

[0031] Currently, driver assistance functions are developing towards multi-mode collaborative control and seamless switching between modes. Based on this trend, the mode control system needs to accurately identify the driver's intentions in order to achieve a smooth handover of control between the autonomous driving system and the driver.

[0032] However, most current mode control systems rely on simple driving action thresholds or steering wheel torque intervention signals as the basis for determining control switching. This makes it difficult to deeply identify the driver's intentions, and they are highly susceptible to interference from unintentional driver operations and momentary actions. This can lead to erroneous control switching or delayed switching responses, disrupting the continuity of the driving process and reducing trust in human-machine collaborative driving. Furthermore, most current mode control systems exhibit issues such as sudden changes in vehicle acceleration and driving shocks when performing control switching, affecting ride comfort and driving safety.

[0033] Based on the above, this application provides a vehicle control method that can acquire multi-source vehicle information representing the driver's operating behavior and the vehicle's operating state over a reference historical time period. Based on this multi-source vehicle information, the method determines the driver's expected acceleration over a reference future time period. Based on the expected acceleration over the reference future time period and the vehicle's current driving mode, it determines the target driving mode to be switched to and controls the vehicle to switch to the target driving mode. Thus, by combining the driver's operating behavior and the vehicle's operating state over a historical time period to determine the driver's expected acceleration, the method can accurately identify the driver's driving intentions and avoid misjudgments caused by the driver's unconscious slight operations or brief road bumps. Furthermore, determining the target driving mode based on continuous and accurate expected acceleration can effectively filter instantaneous interference signals, ensuring the accuracy of the target driving mode, and avoiding high-frequency, unnecessary switching between different modes, significantly improving the stability of mode switching.

[0034] In addition, the vehicle control method provided in this application embodiment can also smooth the target torque when controlling the vehicle, reduce the sudden acceleration change and impact during mode switching, and achieve smooth transition between different modes.

[0035] The vehicle control method provided in this application embodiment can be applied to... Figure 1 The control device 100 shown. Figure 1 The control device 100 shown is installed in the vehicle and can be used to switch the vehicle's driving modes. Figure 1As shown, the control device 100 may include a signal acquisition module 110, a control module 120, and an execution module 130. The signal acquisition module 110 is connected to the control module 120, and the control module 120 is connected to the execution module 130.

[0036] The signal acquisition module 110 is used to acquire multi-source vehicle information, which is used to characterize the driver's operating behavior and the vehicle's operating status. For example, the multi-source vehicle information includes driver operating behavior information and vehicle operating status information.

[0037] As an example, driver behavior information includes, but is not limited to, driver's foot input to the accelerator and brake pedals, and driver's hand input to the steering wheel. Vehicle operating status information includes, but is not limited to, actual vehicle speed, actual acceleration, relative distance to the vehicle in front, and relative speed. The vehicle in front refers to a vehicle in the same lane as this vehicle, traveling in the same direction ahead of it.

[0038] In some examples, the preceding vehicle can also include vehicles traveling in the same direction as this vehicle in the same lane, as well as vehicles traveling in the same direction in adjacent lanes.

[0039] As an example, the signal acquisition module 110 includes, but is not limited to, an angle sensor mounted on the accelerator pedal pivot, a displacement sensor mounted on the brake pedal bracket, a grip sensor integrated into the steering wheel rim, and a gateway interface for accessing the vehicle network via a Controller Area Network (CAN) bus and / or a FlexRay bus. The vehicle network is an on-board communication network system composed of an on-board electronic control unit and multiple on-board communication buses, used to support efficient interaction of perceived data, vehicle operating status, and control commands between various nodes.

[0040] An angle sensor is used to acquire information about the driver's foot operation on the accelerator pedal, such as the accelerator pedal opening percentage signal. As an example, the angle sensor uses a non-contact Hall effect sensor. The analog signal output by the angle sensor is converted from analog to digital to obtain a digital quantity with twelve bits of precision, used to represent the accelerator pedal opening percentage signal from 0% to 100%.

[0041] Displacement sensors are used to acquire information about the driver's foot action on the brake pedal, such as the percentage of brake pedal travel. As an example, a linear variable differential transformer is used as the displacement sensor. The analog voltage signal output by the displacement sensor is first filtered and amplified by a signal conditioning circuit, and then converted from analog to digital to obtain a digital quantity with ten digits of precision, used to represent the percentage of brake pedal travel from 0% to 100%.

[0042] A grip sensor is used to acquire information about how the steering wheel is held, such as the grip status signal. As an example, a grip sensor includes a capacitive sensing array embedded under the steering wheel cover and detection circuitry. The human body can act as an equivalent conductive electrode plate; when the driver's hand contacts the steering wheel, a coupling capacitance is formed between the electrodes of the capacitive sensing array and the human body, introducing additional parasitic capacitance and altering the inherent electric field distribution between the electrodes of the capacitive sensing array.

[0043] The detection circuit is connected to the capacitive sensing array and includes a signal detection unit and a microcontroller. The signal detection unit continuously acquires the raw capacitance signal corresponding to the capacitive sensing array, performs low-pass filtering, and outputs it to the microcontroller. The microcontroller samples the processed capacitance signal at a reference frequency (e.g., once every 200 milliseconds) to obtain the capacitance sample value and calculates the change in the capacitance sample value. If the change is greater than a reference grip threshold, the driver is determined to be holding the steering wheel, and a grip state signal representing the driver's grip state is output. If the change is less than or equal to the reference grip threshold, the driver is determined not to be holding the steering wheel, and a grip state signal indicating the driver is not holding the steering wheel is output.

[0044] The grip status signal can be a Boolean logic signal. When the driver's hands are gripping the steering wheel, the grip status signal is a high-level signal.

[0045] Optionally, the microcontroller calculates the change in the capacitance sample value by comparing the current capacitance sample value with the previous capacitance sample value, or by comparing the current capacitance sample value with a pre-calibrated reference capacitance value in a non-grip state, to obtain the change in the current capacitance sample value.

[0046] Optionally, the microcontroller can output a signal representing the driver's grip state on the steering wheel when the change in the current capacitance sample value is greater than a reference grip threshold. It can also output a signal representing the driver's grip state on the steering wheel when the changes in multiple consecutive capacitance sample values ​​are all greater than the reference grip threshold. No specific limitations are imposed.

[0047] The gateway interface is used to obtain the vehicle's actual speed, actual acceleration (such as longitudinal acceleration), relative distance to the vehicle in front, and relative speed from the vehicle network. As an example, the gateway interface receives and parses broadcast data frames from the vehicle network in real time via a CAN bus and / or FlexRay bus to obtain the vehicle's actual speed, actual acceleration, relative distance to the vehicle in front, and relative speed. Optionally, the speed and acceleration can be calculated by the chassis electronic control unit and sent to the vehicle network, while the relative distance and relative speed to the vehicle in front can be measured by the vehicle's millimeter-wave radar and transmitted to the vehicle network.

[0048] In this embodiment, the signal acquisition module 110 can continuously acquire vehicle multi-source information during vehicle operation and store the acquired vehicle multi-source information in a data buffer for subsequent modules to call.

[0049] For example, let's take the gateway interface as an example. The gateway interface can update the data buffer every 50 milliseconds, storing the parsed operating status information (such as actual vehicle speed, actual acceleration, relative distance to the vehicle in front, and relative speed) into the data buffer.

[0050] The control module 120 is used to acquire the driver's driving intention, and based on the driving intention, determines the target driving mode to be switched. Based on the target driving mode, it outputs the target torque to be executed.

[0051] As an example, the control module 120 may be an on-board central controller that employs a multi-core microprocessor architecture.

[0052] In this embodiment, the control module 120 may include an intent recognition unit 121, a switching judgment unit 122, and a control decision unit 123. The intent recognition unit 121 determines the driver's driving intent based on multi-source vehicle information. The switching judgment unit 122 determines the target driving mode to be switched to based on the driver's driving intent. The control decision unit 123 determines the target execution torque when the target driving mode is an automatic driving mode, and outputs the target execution torque to the execution module 130.

[0053] The intent recognition unit 121, the switching judgment unit 122, and the control decision unit 123 can be implemented in software or in hardware. When implemented in software, the intent recognition unit 121, the switching judgment unit 122, and the control decision unit 123 can be deployed on different processors or on the same processor.

[0054] When implemented in hardware, the intent recognition unit 121, the switching judgment unit 122, and the control decision unit 123 can be independent hardware circuits, such as application-specific integrated circuits (ASICs) or programmable logic devices (PLDs). The intent recognition unit 121, the switching judgment unit 122, and the control decision unit 123 can be integrated within the same processor, or they can be located on different processors, or they can be processor-independent hardware modules. No specific limitations are imposed.

[0055] In this embodiment, the execution module 130 is used to control the vehicle's braking system and / or drive system according to the target execution torque.

[0056] The vehicle's drive system includes, but is not limited to, an engine and an electric motor. The braking system includes an electronic stability control system. In some embodiments, the braking system may also include an electromechanical braking system. No specific limitations are imposed.

[0057] As an example, the execution module 130 includes a drive controller and a brake controller. When the target execution torque is positive, the drive controller adjusts the parameters of the engine or motor according to the target execution torque. When the target execution torque is negative, the brake controller controls the electronic stability control system to adjust the braking force according to the target execution torque.

[0058] The above is merely an example of the signal acquisition module 110, control module 120, and execution module 130 included in the control device 100 according to the embodiments of this application, and is not intended as a specific limitation.

[0059] In some embodiments of this application, the control device 100 may further include a prompting module. The prompting module can be used to output switching prompt information to remind the driver that the vehicle will switch modes.

[0060] In some embodiments of this application, the control device 100 may further include a storage control module and a data storage unit. The storage control module is used to store data during vehicle operation (such as multi-source vehicle information, intent recognition results, and information on mode switching events) into the data storage unit for event retrieval.

[0061] As an example, the data storage unit can employ an embedded multimedia card memory.

[0062] It is understood that the above is merely an illustrative example of the structure of the control device 100 in the embodiments of this application, and is not intended to be a specific limitation.

[0063] To facilitate understanding, the following will be combined with Figure 1 The control device shown describes the vehicle control method provided in the embodiments of this application.

[0064] Please refer to the following: Figure 2 , Figure 2 This is a flowchart illustrating the vehicle control method provided in an embodiment of this application. Figure 2 The method shown can be applied to Figure 1 The control device shown is composed of Figure 1 The control device in the middle performs the operation. For example... Figure 2 As shown, the vehicle control method provided in this application embodiment may include S201 to S204.

[0065] S201, The control device acquires multi-source vehicle information for a reference historical time period.

[0066] In this embodiment, vehicle multi-source information is used to characterize the driver's operating behavior and the vehicle's operating status. For example, the vehicle multi-source information includes driver operating behavior information and vehicle operating status information. A description of the driver's operating information and the vehicle's operating status information can be found above. Figure 1 The descriptions in the examples shown will not be repeated here.

[0067] In this embodiment, the reference historical time period can refer to a period of time tracing back from the current moment to the first reference duration. The control device can acquire multi-source vehicle information within the reference historical time period through the signal acquisition module.

[0068] For example, the signal acquisition module can continuously acquire vehicle multi-source information during vehicle operation and store the acquired vehicle multi-source information in a data buffer. In this way, the control device can obtain vehicle multi-source information for a reference historical time period from the cached vehicle multi-source information.

[0069] Optionally, the signal acquisition module can update the vehicle multi-source information in the data buffer so that the data buffer only retains the vehicle multi-source information acquired within a preset time period backward from the current time, thereby reducing storage overhead. The preset time period is greater than or equal to the first reference time period.

[0070] S202, the control device determines the driver's expected acceleration in a reference future time period based on multi-source vehicle information from a reference historical time period.

[0071] Among them, the driver's expected acceleration in the reference future time period represents the driver's driving intention. In other words, the driver's driving intention can be used to represent the driver's expected acceleration in the reference future time period.

[0072] The reference future time period can refer to a period of time two reference lengths after the current moment. The second reference length can be the same as or different from the first reference length.

[0073] In this embodiment, the control device can determine the driver's desired acceleration over a reference future time period through an intent recognition unit. For example, the intent recognition unit can process multi-source vehicle information within a reference historical time period using an intent recognition model to obtain the driver's desired acceleration over the reference future time period.

[0074] The intent recognition model can employ a neural network model. The input layer of the intent recognition model corresponds to multi-source vehicle information within a reference historical time period, while the output layer corresponds to the driver's expected acceleration within a reference future time period. In other words, the input layer can take in multi-source vehicle information from a historical time period, and the output layer can output the driver's expected acceleration within a future time period.

[0075] In this embodiment, the expected acceleration over the reference future time period is a sequence of accelerations that changes continuously over time. In other words, the output layer node of the intent recognition model can output a sequence of expected accelerations that changes dynamically and continuously over time within the reference future time period.

[0076] As an example, neural network models can include deep learning models based on time series analysis and network models based on multilayer perceptrons.

[0077] In this embodiment, the intent recognition model can employ a deep learning model based on a multi-layer long short-term memory (LSTM) network. The intent recognition model can be iteratively trained using a stochastic gradient descent algorithm with momentum, using mean squared error as the loss function; the training data used is a time-series data sequence collected from real vehicles, covering various driving scenarios.

[0078] As an example, the intent recognition model includes an input layer, three long short-term memory networks, a fully connected layer, and an output layer. The input layer of the intent recognition model can receive feature data for a preset number of time steps and pass this feature data to the first long short-term memory network. The feature data for each time step includes normalized throttle opening percentage signal, brake pedal travel percentage, steering wheel grip state signal, vehicle speed, acceleration, relative distance to the vehicle in front, and relative speed, among other features.

[0079] Optionally, three Long Short-Term Memory (LSTM) network layers are cascaded sequentially. The first LSM layer contains 128 hidden units, and its output sequence is directly passed to the second LSM layer. The second LSM layer contains 64 hidden units, and its output sequence is directly passed to the third LSM layer. The third LSM layer contains 32 hidden units, and its output is fed into a fully connected layer.

[0080] The fully connected layer contains 16 neurons and uses the rectified linear unit activation function. The output layer of the intent recognition model is a single neuron, and the hyperbolic tangent activation function is used to constrain the output to between -1 and +1, which can characterize the normalized expected acceleration range.

[0081] For example, the intent recognition model can obtain the desired acceleration through formula (1).

[0082] a_des_norm=tanh(W_o h_t+b_o)(1) Where a_des_norm is the normalized expected acceleration, with a value range of [-1, 1]. W_o is the weight matrix of the output layer. h_t is the output of the fully connected layer. b_o is the bias vector of the output layer. Equation (1) can map the high-dimensional features of the output of the Long Short-Term Memory network to physically meaningful expected acceleration values, ensuring that the output is stable and interpretable.

[0083] In this embodiment, the intent recognition model can extract and predict the temporal features of multi-source vehicle information in the input reference historical time period using a sliding time window. Thus, by sequentially updating the sliding time window, the intent recognition model can continuously output the expected acceleration sequence that represents the change of time in the reference future time period.

[0084] It is understood that the above is merely a specific description of the structure of the intent recognition model in the embodiments of this application, and is not intended to be a specific limitation. The intent recognition model can include more or fewer long short-term memory network layers, and the long short-term memory network layers can include more or fewer hidden units, etc. Fully connected layers can also include more or fewer neurons, and there is no specific limitation.

[0085] S203, the control device determines the target driving mode to be switched to by the vehicle based on the desired acceleration.

[0086] In this embodiment, the vehicle's driving mode may include an automated driving mode and a manual driving mode. The manual driving mode refers to a mode where the driver takes the lead in vehicle control, and the vehicle's driver assistance systems do not actively perform vehicle motion control. The automated driving mode refers to a mode where the driver assistance systems take the lead in vehicle motion control, such as adaptive cruise control.

[0087] In this embodiment of the application, the control device can determine the target driving mode to be switched to by the switching determination unit. For example, the switching determination unit can determine the target driving mode to be switched to by the vehicle based on the driver's driving intention, such as based on the driver's expected acceleration in a reference future time period.

[0088] Optionally, the switching determination unit can determine the vehicle's target driving mode based on pre-configured switching conditions and the expected acceleration over a future time period.

[0089] For example, the pre-configured switching conditions include switching conditions for both autonomous driving mode and manual driving mode. If the expected acceleration over a future time period meets the switching conditions for autonomous driving mode, the target driving mode is determined to be autonomous driving mode. If the expected acceleration over a future time period meets the switching conditions for manual driving mode, the target driving mode is determined to be manual driving mode. No specific restrictions are imposed.

[0090] S204, The control device controls the vehicle's driving mode to switch to the target driving mode.

[0091] In this embodiment of the application, after determining the target driving mode of the vehicle, it is possible to control the vehicle's driving mode to switch to the target driving mode.

[0092] The control module can switch driving modes via a switching judgment unit. For example, if the current driving mode is manual driving mode and the target driving mode is autonomous driving mode, the switching judgment unit can control the vehicle's driving mode to switch from manual driving mode to autonomous driving mode. Conversely, if the current driving mode is autonomous driving mode and the target driving mode is manual driving mode, the switching judgment unit can control the vehicle's driving mode to switch from autonomous driving mode to manual driving mode.

[0093] The vehicle control method of this application can acquire multi-source vehicle information representing the driver's operation behavior and the vehicle's operating state over a reference historical time period. Based on this information, it determines the driver's expected acceleration over a reference future time period. Then, based on this expected acceleration, it determines the target driving mode to be switched to and controls the vehicle to switch to the target driving mode. By combining the driver's operation behavior and the vehicle's operating state over a historical time period to determine the driver's expected acceleration, it can accurately identify the driver's driving intentions and avoid misjudgments caused by unintentional slight operations or brief road bumps. Furthermore, determining the target driving mode based on continuous and accurate expected acceleration effectively filters out transient interference signals, ensuring the accuracy of the target driving mode and avoiding high-frequency, unnecessary switching between different modes, significantly improving the stability of mode switching.

[0094] In addition, this application employs a deep learning model based on a multi-layer long short-term memory network. This model learns the complex nonlinear mapping relationship between driver operating habits and vehicle dynamic response through training, enabling it to infer the driver's true intention from historical multi-source information, rather than making judgments solely based on instantaneous operations. This achieves deep and continuous quantitative recognition of the driver's driving intentions, greatly improving the accuracy and generalization ability of intention recognition. It effectively avoids system misjudgments caused by unconscious slight operations by the driver or brief road bumps, providing a more reliable and accurate basis for driving mode switching decisions.

[0095] In some embodiments of this application, the control device, such as the switching determination unit of the control device, can combine the expected acceleration of a reference future time period and the current driving mode of the vehicle to determine the target driving mode to be switched according to the pre-configured switching conditions.

[0096] The pre-configured switching conditions include a first reference condition and a second reference condition. The first reference condition can be the switching condition corresponding to the manual driving mode, and the second reference condition can be the switching condition corresponding to the autonomous driving mode.

[0097] The first reference condition includes: within a reference future time period, the duration for which the absolute value of the expected acceleration is continuously greater than a first threshold reaches a first duration threshold. The second reference condition includes: within a reference future time period, the duration for which the absolute value of the expected acceleration is continuously less than a second threshold reaches a second duration threshold. The second threshold is less than the first threshold. The first duration threshold and the second duration threshold can be the same or different; there is no specific restriction.

[0098] For example, the switching determination unit can determine the target driving mode as manual driving mode when the expected acceleration in the reference future time period meets a first reference condition and the vehicle's current driving mode is autonomous driving mode. When the expected acceleration in the reference future time period meets a second reference condition and the vehicle's current driving mode is manual driving mode, it determines the target driving mode as autonomous driving mode.

[0099] In other words, within a future timeframe, if the absolute value of the expected acceleration is continuously greater than a first threshold for a duration exceeding the first threshold, and the vehicle is currently in autonomous driving mode, then it is determined that the vehicle needs to switch from autonomous driving mode to manual driving mode; that is, the target driving mode for the vehicle is manual driving mode. Conversely, if the absolute value of the expected acceleration is continuously less than a second threshold for a duration exceeding the second threshold, and the vehicle is currently in manual driving mode, then it is determined that the vehicle needs to switch from manual driving mode to autonomous driving mode; that is, the target driving mode for the vehicle is autonomous driving mode.

[0100] In this embodiment of the application, if the absolute value of the expected acceleration is continuously greater than the first threshold for a duration of time within a reference future time period, or if the absolute value of the expected acceleration is continuously less than the second threshold for a duration of time, it indicates that the driver's driving intention is valid, that is, the expected acceleration in the reference future time period is valid.

[0101] For example, the switching judgment unit can determine whether the expected acceleration for the reference future time period is valid by formula (2).

[0102] If |a_des|>a_thres1 for Δt≥T_min1, then T_hold=1, otherwise equals 0. Or, if |a_des|<a_thres2 for Δt≥T_min2, then T_hold=1, otherwise equals 0. (2) Where a_des is the expected acceleration over a reference future time period. a_thres1 is the first threshold. Δt is the duration. T_min1 is the first duration threshold. a_thres2 is the second threshold. Δt is the duration. T_min2 is the second duration threshold. T_hold is the intent flag; when T_hold=1, it indicates that the driver's driving intent is valid, i.e., the expected acceleration over the reference future time period is valid.

[0103] In this embodiment, by detecting the duration of the signal, mode switching can be triggered while ensuring the driver's driving intention is valid, effectively avoiding interference from instantaneous data and improving the accuracy of mode switching.

[0104] In some embodiments of this application, if the expected acceleration over a future time period satisfies a first reference condition and the vehicle's current driving mode is manual driving mode, the control device may not perform a mode switch. Alternatively, if the expected acceleration over a future time period satisfies a second reference condition and the vehicle's current driving mode is automatic driving mode, the control device may not perform a mode switch. If the expected acceleration over a future time period does not satisfy either the first or second reference condition, the control device may also not perform a mode switch. No specific limitations are imposed.

[0105] In this embodiment, the switching determination unit can also determine the target driving mode by considering the steering wheel grip state. For example, before determining the target driving mode as autonomous driving mode, the switching determination unit can also obtain the steering wheel grip state. If the steering wheel grip state indicates that the steering wheel is being gripped, the target driving mode is determined to be autonomous driving mode.

[0106] In other words, the target driving mode can only be determined to be the autonomous driving mode if the expected acceleration in the reference future time period meets the second reference condition, the vehicle's current driving mode is manual driving mode, and the vehicle's steering wheel is being held.

[0107] The switching judgment unit can receive the steering wheel grip state signal output by the grip sensor. Based on the grip state signal, the grip state of the vehicle's steering wheel is determined.

[0108] For example, if the grip state signal indicates that the driver is holding the steering wheel, the grip state of the vehicle's steering wheel is determined to be "held". If the grip state signal indicates that the driver is not holding the steering wheel, the grip state of the vehicle's steering wheel is determined to be "not held".

[0109] Optionally, the grip state signal output by the grip sensor can be input to the switching determination unit via an asynchronous interrupt. This allows for immediate triggering of an interrupt notification when the steering wheel grip state changes, eliminating the need for the switching determination unit to periodically poll and sample, thus achieving real-time response to grip state changes and reducing mode determination response latency.

[0110] This application embodiment ensures that the driver is ready to take over the vehicle at any time by switching to autonomous driving mode while the driver is holding the steering wheel, thus improving driving safety during the human-machine co-driving transition.

[0111] In some embodiments, the switching determination unit can also determine the steering wheel grip state based on the steering wheel grip state signal and the torque sensor signal provided by the vehicle's electric power steering system.

[0112] For example, the switching determination unit can perform a logical AND operation on the grip state signal and the torque sensor signal, and determine that the grip state of the vehicle's steering wheel is "being gripped" when both the grip state signal and the torque sensor signal are valid.

[0113] Specifically, a valid grip state signal indicates that the driver is holding the steering wheel. A valid torque sensor signal indicates that the driver is applying steering wheel torque. In other words, the switching determination unit can also determine that the steering wheel is being gripped when both the grip state signal and the torque sensor signal indicate that the driver is applying steering wheel torque. Conversely, if the grip state signal indicates that the driver is not holding the steering wheel and / or the torque sensor signal indicates that the driver is not applying steering wheel torque, the unit determines that the steering wheel is not being gripped.

[0114] This application embodiment, by simultaneously utilizing the grip state signal and the steering wheel torque signal for joint verification, can effectively reduce the risk of misjudgment caused by single sensor detection and improve the accuracy and reliability of autonomous driving mode verification results.

[0115] In this embodiment, a multi-condition collaborative switching logic is adopted that combines threshold comparison, duration determination, and steering wheel grip state verification. This logic not only determines the intensity of the driver's intention, but also requires the corresponding driving intention to be maintained for a specified duration to ensure the stability of the driving intention. At the same time, the steering wheel grip state is set as a necessary switching condition for switching to the autonomous driving mode, thereby ensuring the reliability of the driving mode switching.

[0116] In this embodiment of the application, in order to avoid false switching caused by momentary interference, the control device can continuously determine the target driving mode to be switched to by the vehicle multiple times, and when the target driving mode is determined to be the same multiple times, the control device controls the vehicle's driving mode to switch to the target driving mode.

[0117] Optionally, a single calculation cycle can determine the vehicle's target driving mode once. Within a single calculation cycle, the control device can complete one round of target driving mode determination. Based on this calculation cycle, the control device can periodically trigger the mode switching determination process. Each round of the mode switching determination process can determine the driver's expected acceleration in a reference future time period based on the vehicle's multi-source information from a reference historical time period obtained in the current calculation cycle, and determine the target driving mode for the current calculation cycle based on the expected acceleration. The length of the calculation cycle can be configured according to actual calculation requirements.

[0118] The control device will switch the vehicle's driving mode to the target driving mode only if the target driving mode obtained in multiple consecutive calculation cycles (such as 5 calculation cycles) is the same.

[0119] The embodiments of this application use a de-jittering algorithm based on consistency judgment over multiple consecutive calculation cycles to control driving mode switching. This ensures the stability of driving mode switching, achieves high robustness and high reliability in driving mode switching, effectively filters instantaneous interference signals, and prevents the control device from performing high-frequency and unnecessary oscillation switching between "manual driving" and "autonomous driving" modes. This significantly improves the stability of mode switching and the driver's confidence.

[0120] In this embodiment, after the vehicle's driving mode is switched to the target driving mode, the vehicle can respond to the control commands generated by the control logic corresponding to the target driving mode and complete driving control according to the control commands. For example, when the target driving mode is manual driving mode, the vehicle can respond to control commands generated based on the driver's operating behavior. When the target driving mode is automatic driving mode, the vehicle can respond to control commands output based on the control logic corresponding to the automatic driving mode. It is understood that the vehicle's response to control commands may include the vehicle's drive system and / or braking system responding to control commands, which may involve adjusting its own operating parameters according to the control commands.

[0121] In this embodiment, when the target driving mode is automatic driving mode, the control device can first determine the target acceleration of the vehicle and control the vehicle based on the target acceleration. Based on this, please refer to... Figure 3 The vehicle control method provided in this application embodiment also includes S301 to S303.

[0122] S301, when the target driving mode is automatic driving mode, the control device obtains the target acceleration of the vehicle based on the reference vehicle speed, the vehicle's current actual speed, the relative distance and relative speed with the vehicle in front, and the model predictive control algorithm.

[0123] The control device can determine the target acceleration of the vehicle through a control decision unit. For example, the control decision unit can receive a target driving mode sent by a switching judgment unit, and determine the target acceleration of the vehicle based on the target driving mode.

[0124] Optionally, the switching determination unit can instruct the control decision unit of the vehicle's target driving mode via a mode switching command. For example, after determining the target driving mode, the switching determination unit can send a mode switching command to the control decision unit based on the target driving mode.

[0125] Optionally, the switching judgment unit has a built-in state machine, and the switching judgment unit can output a mode switching instruction based on the flag bits of the state machine.

[0126] For example, the switching decision unit can first load the specific values ​​of the first threshold, the second threshold, the first duration threshold, and the second duration threshold from non-volatile memory. Then, upon receiving the expected acceleration for the reference future time period, it compares the expected acceleration for the reference future time period with the first threshold and the second threshold, and starts the duration counter.

[0127] If the absolute value of the desired acceleration is continuously greater than the first threshold for a duration that reaches the first duration threshold, and the vehicle is currently in autonomous driving mode, the state machine sets the flag to exit autonomous driving mode, and the switching judgment unit outputs a mode switching command to switch from autonomous driving mode to manual driving mode.

[0128] If the absolute value of the desired acceleration is continuously less than the second threshold for a duration that reaches the second duration threshold, and the vehicle is currently in manual driving mode and the steering wheel is being held, the state machine sets the flag to request entry into autonomous driving mode, and the switching judgment unit outputs a mode switching command to switch from manual driving mode to autonomous driving mode.

[0129] Understandably, the switching decision unit can generate the final mode switching command based on a de-jitter algorithm that determines consistency across consecutive cycles. For example, the switching decision unit can output a mode switching command to the control decision unit if the target driving mode is determined to be the same across multiple consecutive calculation cycles.

[0130] As an example, a mode switching command may include a mode field, allowing the control decision unit to determine the target driving mode based on this field. Different mode fields indicate different driving modes.

[0131] For example, after receiving a mode switching command, the control decision unit queries a pre-stored mode mapping table based on the mode field in the command to determine the target driving mode to be switched to. The mode mapping table stores the correspondence between different driving modes and the mode field.

[0132] In this embodiment of the application, after the control decision unit determines the target driving mode of the vehicle, if the target driving mode is an automatic driving mode, the control decision unit can determine the target acceleration of the vehicle according to the pre-configured acceleration solution strategy.

[0133] For example, when the target driving mode is autonomous driving mode, the control decision unit can use an acceleration solution strategy based on model predictive control (MPC) to determine the target acceleration of the vehicle.

[0134] Based on the MPC-based acceleration calculation strategy, the control decision unit can obtain the vehicle's target acceleration using a model predictive control algorithm, based on a reference vehicle speed, the vehicle's current actual speed, and the relative distance and speed to the vehicle ahead. In other words, when the target driving mode is autonomous driving, the control decision unit can obtain the vehicle's target acceleration based on a model predictive control algorithm, using a reference vehicle speed, the vehicle's current actual speed, and the relative distance and speed to the vehicle ahead. The reference vehicle speed can be the cruise speed stored in the vehicle and set by the driver.

[0135] For example, the control decision unit can use the reference vehicle speed, the vehicle's current actual vehicle speed, the relative distance and relative speed with the vehicle in front as input parameters to the model predictive control algorithm. The model predictive control algorithm then calculates the target acceleration of the vehicle based on the reference vehicle speed, the vehicle's current actual vehicle speed, and the relative distance and relative speed with the vehicle in front.

[0136] In some embodiments, to improve the accuracy of target acceleration calculation, the input parameters of the model predictive control algorithm may also include vehicle load and road slope. No specific limitations are imposed. The vehicle load can be obtained through a suspension height sensor, and the road slope can be obtained through an inertial measurement unit.

[0137] Understandably, the process of model predictive control algorithms solving for the target acceleration of a vehicle can be described with reference to conventional techniques, and will not be elaborated here.

[0138] Optionally, if the target driving mode is manual driving mode, the control decision unit does not participate in the vehicle driving control decision. For example, if the control decision unit does not need to calculate the target acceleration, the vehicle driving control is achieved by the driver's driving behavior.

[0139] This application embodiment determines the target acceleration based on the model predictive control algorithm in autonomous driving mode. On the one hand, it can ensure the safety and ride comfort of following other vehicles in autonomous driving mode, avoid sudden changes in acceleration characteristics before and after mode switching, and improve the driving experience.

[0140] In some embodiments of this application, after obtaining the target acceleration of the vehicle, the control decision unit can also smooth the target acceleration based on a reference acceleration change rate threshold.

[0141] For example, the control decision unit includes an acceleration rate of change limiting subunit and an acceleration calculation subunit. The acceleration calculation subunit is used to determine the target acceleration of the vehicle, and after determining the target acceleration, it inputs the target acceleration of the vehicle to the acceleration rate of change limiting subunit. The acceleration rate of change limiting subunit is used to smooth the target acceleration of the vehicle.

[0142] Optionally, the acceleration rate of change limiting subunit can smooth the target acceleration based on a reference acceleration rate of change threshold. The reference acceleration rate of change threshold can be set as needed and is not specifically limited.

[0143] For example, the acceleration rate of change limiting subunit can acquire the vehicle's current actual acceleration and calculate the acceleration deviation between the target acceleration and the current actual acceleration. Based on a reference acceleration rate of change threshold, it determines the maximum allowable acceleration adjustment step size within a single control cycle. If the absolute value of the acceleration deviation is greater than the maximum adjustment step size, the target acceleration is smoothed. If the absolute value of the acceleration deviation is less than or equal to the maximum adjustment step size, no smoothing of the target acceleration is required, and the target acceleration is used directly.

[0144] The control device can periodically send control commands based on a control cycle. Within a single control cycle, the control device can complete one acceleration calculation (including acceleration calculation and smoothing) and output control commands to the actuators (drive system and / or braking system) based on the calculation results. The duration of the control cycle can be configured according to the overall vehicle control requirements.

[0145] Specifically, when the absolute value of the acceleration deviation exceeds the maximum adjustment step size, the acceleration rate of change limiting subunit can adjust the maximum adjustment step size along the direction towards the target acceleration based on the actual acceleration, thus obtaining a smoothed target acceleration. For example, if the acceleration deviation is positive, the maximum adjustment step size is increased based on the actual acceleration to obtain the smoothed target acceleration. If the acceleration deviation is negative, the maximum adjustment step size is subtracted from the actual acceleration to obtain the smoothed target acceleration. Through iterative calculations cycle by cycle, the control device can gradually bring the output acceleration closer to the target acceleration. This effectively limits the change in acceleration per unit time during the transition to the target value, preventing abrupt changes in acceleration.

[0146] Optionally, the acceleration rate of change limiting subunit can also acquire the maximum jerk threshold set by the vehicle's longitudinal dynamics characteristics and perform smoothing filtering on the target acceleration based on this maximum jerk threshold. For example, the acceleration rate of change limiting subunit can input the target acceleration command into the jerk limiting filter, such as an S-curve planner or a tracking differentiator, to ensure that the rate of change (jerk) during the acceleration transition process never exceeds the set maximum jerk threshold. With the help of continuous domain filtering constraints, the acceleration curve can be made continuous and differentiable, fundamentally eliminating the mechanical shock and occupant discomfort caused by sudden acceleration changes. For details, please refer to the description of conventional techniques; further elaboration is not provided here.

[0147] This application embodiment, by smoothing the target acceleration, can eliminate the sudden acceleration change and impact during mode switching, greatly improving the vehicle's ride comfort and handling quality.

[0148] S302, the control device determines the target execution torque of the vehicle based on the target acceleration.

[0149] In the embodiments of this application, after obtaining the target acceleration of the vehicle (such as the smoothed target acceleration), the control device, such as the control decision unit of the control device, can also calculate the target execution torque of the vehicle based on the target acceleration. The vehicle is then controlled based on the target execution torque.

[0150] For example, the control decision unit can convert the vehicle's target acceleration into the vehicle's target torque based on the vehicle's longitudinal dynamics model. For instance, the control decision unit can use the target acceleration, load coefficient, road slope angle, and the vehicle's current actual speed as inputs to the vehicle's longitudinal dynamics model. The model then uses these parameters to determine the vehicle's target torque. The process of determining the vehicle's target torque using the longitudinal dynamics model can be described using conventional techniques and will not be elaborated upon here.

[0151] S303, the control device controls the vehicle according to the target torque.

[0152] In the embodiments of this application, after obtaining the target execution torque of the vehicle, the control device can control the vehicle according to the target execution torque.

[0153] In this embodiment, the control device can control the vehicle through the execution module. For example, after obtaining the target execution torque of the vehicle, the control decision unit can output the target execution torque to the execution module, so that the execution module controls the vehicle's braking system and / or drive system according to the target execution torque.

[0154] As described above, the vehicle's drive system includes, but is not limited to, an engine and an electric motor. The braking system includes an electronic stability control system.

[0155] As an example, the control decision unit can send a target torque request to the execution module via the CAN bus. This target torque request indicates the vehicle's target execution torque. Upon receiving the target torque request, the execution module can control the vehicle according to the target execution torque indicated in the request.

[0156] As an example, the execution module is configured with a powertrain control interface (such as a CAN bus) and a vehicle controller interface. The execution module includes a drive controller and a brake controller.

[0157] When the target torque is a positive torque, the drive controller, upon receiving the target torque, can map it to underlying control parameters by looking up a table. For example, the drive controller can map the target torque to a throttle opening parameter based on the engine's universal characteristic curve. Based on the throttle opening parameter, it sends a control command indicating the throttle opening to the engine control unit via the powertrain control interface to adjust the power output. As another example, the drive controller can map the target torque to a motor torque parameter using a motor efficiency graph. Based on the motor torque parameter, it sends a control command indicating the motor torque to the motor controller via the powertrain control interface to adjust the power output.

[0158] When the target torque is negative, the brake controller, upon receiving the target torque, maps it to a target braking torque based on the hydraulic characteristic model of the braking system and obtains the braking pressure increase slope. Then, via the vehicle controller interface, it sends a control message (control command) containing the target braking torque and the braking pressure increase slope to the electronic stability control system using a low-level communication protocol. Upon receiving the control message, the hydraulic control unit within the electronic stability control system can adjust the wheel cylinder pressure accordingly to achieve coordinated control of vehicle deceleration and braking. The braking pressure increase slope can be determined based on the target acceleration.

[0159] Understandably, when the target acceleration is positive, the target execution torque is positive, and when the target acceleration is negative, the target execution torque is negative.

[0160] In some embodiments of this application, considering that when a vehicle transitions between driving and braking states, a direct hard switch between the driving and braking systems can easily create a control vacuum, causing vehicle jerking and speed fluctuations. Therefore, embodiments of this application can enable a torque cross-gradient algorithm during vehicle state transitions, allowing the driving torque and braking torque to be adjusted in a gradually decreasing manner, thereby achieving a continuous and smooth transition between driving and braking states and eliminating longitudinal jerking and speed fluctuations during the state transition process.

[0161] Based on this, the execution module can control the vehicle according to the target torque, and can also obtain a torque weighting coefficient when it detects a switch between driving and braking states. Based on the torque weighting coefficient, the target braking torque and target driving torque are determined. The vehicle's braking system is controlled based on the target braking torque, and the vehicle's driving system is controlled based on the target driving torque.

[0162] Optionally, the execution module can set a reference torque range. This reference torque range can be the critical torque transition range between the driving state and the braking state. After obtaining the target execution torque, the execution module can compare the target execution torque with the reference torque range. If the target execution torque is within the reference torque range, it is determined that the vehicle is about to switch between the driving state and the braking state. If the target execution torque is not within the reference torque range, it is determined that the vehicle will not switch between the driving state and the braking state.

[0163] Optionally, the execution module can also acquire historical execution torque and obtain the torque change trend based on the historical execution torque and the target execution torque. If the torque change trend indicates that the execution torque sequence (including the sequences of historical execution torque and target execution torque) has a tendency to cross zero, it is determined that the vehicle is about to switch between driving and braking states.

[0164] If the torque change trend indicates that the executed torque sequence does not exhibit a tendency to cross zero points, it is determined that the vehicle will not switch between driving and braking states. Here, a tendency for the executed torque sequence to cross zero points means that the executed torque sequence exhibits a trend of continuously changing from positive to negative values, or vice versa. It can be understood that when the executed torque sequence shows a trend of continuously changing from positive to negative values, it corresponds to the vehicle transitioning from driving to braking states; conversely, when the executed torque sequence shows a trend of continuously changing from negative to positive values, it corresponds to the vehicle transitioning from braking to driving states.

[0165] Optionally, the execution module can also determine that the vehicle is about to switch between driving and braking states if the target execution torque is within the reference torque range and the torque change trend indicates that the execution torque sequence has a tendency to cross zero. If the target execution torque is not within the reference torque range, and / or the torque change trend indicates that the execution torque sequence does not have a tendency to cross zero, the module can determine that the vehicle will not switch between driving and braking states.

[0166] Once the execution module determines that the vehicle is about to switch between driving and braking states, it can obtain the torque weighting coefficient and determine the target braking torque and target driving torque based on the torque weighting coefficient.

[0167] In this embodiment, the target execution torque represents the total longitudinal torque requirement of the vehicle. However, since the drive system and the braking system are two independent actuators, they cannot work together directly based on a single total torque command. Therefore, this embodiment can decompose the target execution torque in real time according to the torque weighting coefficient to obtain the driving torque and braking torque.

[0168] The torque weighting coefficient is a time-varying coefficient, which the execution module can obtain based on time. During the transition from driving to braking, the torque weighting coefficient smoothly decreases from 1 to 0 over time, thus gradually attenuating the driving torque while simultaneously increasing the braking torque. Conversely, during the transition from braking to driving, the torque weighting coefficient smoothly increases from 0 to 1 over time, thus gradually decreasing the braking torque while simultaneously and smoothly increasing the driving torque.

[0169] Optionally, the execution module can obtain the torque weighting coefficient by looking up a table. For example, when controlling the vehicle according to the target execution torque, the execution module can start a timer and retrieve the torque weighting coefficient corresponding to the current time from a pre-configured coefficient table. The coefficient table stores the torque weighting coefficients for different times.

[0170] Optionally, the execution module can also obtain the torque weighting coefficient corresponding to the current road slope and current timing from a pre-configured coefficient table. The coefficient table stores the torque weighting coefficients at different times and under different road slopes.

[0171] It is understood that the above is merely an illustrative example of how to obtain the torque weighting coefficient in the embodiments of this application, and is not intended as a specific limitation.

[0172] After obtaining the torque weighting coefficient, the execution module can determine the target braking drive and target driving torque based on the torque weighting coefficient and the target execution torque.

[0173] Optionally, the execution module can determine the target execution torque and the target drive torque according to formula (3).

[0174] T_drive=T_final×a, T_brake=T_final×(1-a)(3) Where T_final is the target execution torque, T_drive is the target driving torque, T_brake is the target braking torque, and a is the torque weighting coefficient.

[0175] Taking the transition from driving to braking state as an example, where the execution module obtains the torque weighting coefficient based on the timing interval, the following explanation is provided. When the execution module controls the vehicle based on the target execution torque, it starts a timer. At the initial moment, the torque weighting coefficient is 1; at this time, the target driving torque and the target execution torque are the same, and the target braking torque is 0. When the timer's timing interval is T1, the execution module looks up the torque weighting coefficient in the table based on T1, and it is 0.5. At this time, the target driving torque and the target braking torque are equal in magnitude and are half of the target execution torque. When the timer's timing interval is T2, the execution module looks up the torque weighting coefficient in the table based on T2, and it is 0. At this time, the target braking torque and the target execution torque are the same, and the target driving torque is 0.

[0176] After obtaining the target driving torque and target braking torque, the execution module can control the vehicle's braking system based on the target braking torque, and control the vehicle's drive system based on the target driving torque. The process of controlling the vehicle's braking system based on the target braking torque can be referenced to the process of controlling the braking system when the target driving torque is negative, and will not be repeated here. The process of controlling the vehicle's drive system based on the target driving torque can be referenced to the process of controlling the drive system when the target driving torque is positive.

[0177] In this embodiment of the application, when the vehicle is transitioning between states, the torque cross-gradient algorithm (torque weight coefficient) can smoothly transition the driving torque and braking torque, achieving a smooth connection between different control modes and between the driving / braking systems, which greatly improves the ride comfort of the vehicle.

[0178] In some embodiments of this application, before switching the vehicle's driving mode to the target driving mode, the control device can also output a switching prompt message to remind the driver that the vehicle is about to switch modes.

[0179] The control device can output switching prompts via a prompting module. For example, the prompting module can receive a mode switching command from the switching judgment unit and output switching prompts in response to receiving the mode switching command.

[0180] As an example, the prompting module can output a switching prompt message after the switching judgment unit outputs the mode switching command but before the actual mode switching is executed. In this way, the prompting module can inform the driver in advance of the impending transfer of control before the actual mode switching, facilitating timely intervention by the driver and improving the safety performance of mode switching.

[0181] For example, the prompt module can output a switching prompt message within the first time period after the switching judgment unit generates the mode switching command, and continue to output it until the second time period before the actual mode switching occurs, thereby providing the driver with sufficient reaction time and effectively improving the safety of human-machine operation switching.

[0182] Taking a first duration of 200 milliseconds and a second duration of 1.5 seconds as an example, the prompt module outputs a switching prompt message within 200ms after the switching judgment unit 122 generates the mode switching command, and the switching prompt message continues to be output until the distance execution module starts the control according to the target execution torque, that is, 1.5 seconds before the actual mode switching takes effect.

[0183] As an example, the prompt module can output toggle prompts through visual, auditory, and / or tactile means. No specific restrictions are imposed.

[0184] Let's take visually displaying switching prompts as an example. The prompt module can output switching prompts through a display screen on the vehicle, such as a prompt icon on the instrument cluster. For example, after receiving a mode switching command, the prompt module controls the color of the prompt icon to switch from a first color (such as green) to a second color (such as red), and controls the prompt icon to flash continuously. In essence, controlling the color change and flashing of the prompt icon constitutes outputting the switching prompt information.

[0185] Let's take the auditory output of switching prompts as an example. The prompt module can emit a preset number of prompts at a preset frequency through the vehicle's audio playback module, such as the audio channel of the in-vehicle infotainment system. For example, the prompt module can emit three prompts through the audio channel of the in-vehicle infotainment system, and the frequency of the prompts is 800 Hz.

[0186] Let's take the tactile output of switching prompts as an example. The prompt module applies a pulse torque to the steering wheel through the electric power steering system. This pulse torque has a preset frequency and a preset amplitude. For example, the preset frequency is 5 Hz, and the preset amplitude is 0.3 N·m. In essence, the process of applying pulse torque to the steering wheel through the electric power steering system is the output of the switching prompt information.

[0187] It is understood that the above is merely an example of output switching prompt information in the embodiments of this application and is not intended to be a specific limitation.

[0188] In some embodiments of this application, the control device can also store data such as multi-source vehicle information, intent recognition results, and mode switching events during vehicle driving, and can incrementally learn the intent recognition model based on the stored data to optimize the recognition accuracy of the intent recognition model.

[0189] The control device can store data through a storage control module. For example, the storage control module can store multi-source vehicle information, intent recognition results, and mode switching event information obtained during vehicle operation into the data storage unit.

[0190] The intent recognition result may include the acceleration sequence (i.e. the expected acceleration) obtained from each intent recognition.

[0191] The mode switching events include a first mode switching event and a second mode switching event. The first mode switching event refers to the mode switching event triggered by the switching determination unit. The second mode switching event refers to the mode switching event triggered by the driver. The information for each mode switching event includes the time of the event, the driving mode before the switch, the driving mode after the switch, and the source of the event trigger (such as the driver or the switching determination unit).

[0192] Optionally, the control decision unit can receive mode switching commands from the driver or from the mode switching judgment unit. The driver can send mode switching commands via voice, touch, or buttons, etc. No specific restrictions are imposed.

[0193] After receiving a mode switching command, the control decision unit can instruct the storage control module to record information about the mode switching event triggered by the mode switching command.

[0194] As an example, the storage control module writes data to the data storage unit using a circular buffer. For instance, the storage control module can cache data in the data storage unit for a preset duration (such as one month) using a circular buffer. When the time corresponding to the cached data reaches the preset duration, newly written data overwrites the oldest historical data, ensuring that the storage unit always retains the latest data that does not exceed the preset duration.

[0195] In some embodiments of this application, the control device is capable of incrementally learning the intent recognition model through the intent recognition unit.

[0196] Optionally, the intent recognition unit can trigger incremental learning of the intent recognition model when the vehicle is off and connected to a charging device.

[0197] Specifically, the intent recognition unit employs an online sequence learning algorithm for incremental learning of the intent recognition model. The intent recognition unit can acquire training data from the data storage unit and perform incremental learning of the intent recognition model based on this training data.

[0198] Optionally, the intent recognition unit can acquire information about mode switching events from the data storage unit. Based on the information about the mode switching events, it can detect whether a second mode switching event has occurred within a reference time period following the first mode switching event.

[0199] For example, the intent recognition unit can detect whether a second mode switching event occurs within a reference time period after the first mode switching event, based on the event trigger source and the time of event occurrence.

[0200] After detecting a second mode switching event within a reference time period following a first mode switching event, the system checks whether the driving modes before and after the first mode switching event are the same as those before and after the second mode switching event. If the driving modes are different, the intent recognition unit retrieves training data from the data storage unit based on the event occurrence time of the first mode switching event. For example, the intent recognition unit can retrieve multi-source vehicle information for a third reference time period before the event occurrence time and the actual acceleration for a fourth reference time period after the event occurrence time from the data storage unit. Training data is then obtained based on the multi-source vehicle information for the third reference time period before the event occurrence time and the actual acceleration for the fourth reference time period after the event occurrence time.

[0201] The training data includes training sample data and label data. The multi-source vehicle information during the third reference time before the event occurs can be used as training sample data, while the actual acceleration during the fourth reference time after the event occurs is used as label data.

[0202] Optionally, the third reference duration may be the same as or different from the first reference duration, and the fourth reference duration may be the same as or different from the second reference duration.

[0203] The intent recognition unit inputs training sample data into the intent recognition model to obtain predicted data and determine the error between the predicted and labeled data. It calculates the gradient using backpropagation and the error, and updates the weight parameters of the intent recognition model based on the gradient and the learning rate. The learning rate is a hyperparameter used to limit the magnitude of each parameter update.

[0204] Optionally, the intent recognition unit can update the weight parameters of the intent recognition model using formula (4).

[0205] W_new=W_old+n L(4) Where W_new is the updated weight, W_old is the old weight, and n is the learning rate; L is the gradient.

[0206] Alternatively, when updating the weight parameters of the intent recognition model, only the weight parameters of the fully connected layers can be updated, thus avoiding the destruction of the temporal features already learned by the long short-term memory network layers.

[0207] This application embodiment records system operation data through a built-in data storage unit and designs an incremental learning process based on an online sequence learning algorithm to be triggered under specific conditions, periodically optimizing the parameters of the fully connected layer in the intent recognition model. In this way, the intent recognition model can be continuously optimized using actual user data, giving it the ability to continuously evolve and personalize. Furthermore, over time, the intent recognition model can increasingly accurately adapt to the unique operating styles and habits of different drivers, mitigating the impact of behavioral differences between drivers on device performance, thereby improving the device's personal adaptability and long-term user satisfaction.

[0208] Based on the same inventive concept, embodiments of this application also provide a control device. Please refer to the following: Figure 4 The control device may include a signal acquisition module and a control module. The signal acquisition module is used to acquire multi-source vehicle information, which represents the driver's operating behavior and the vehicle's operating status. The control module is used to determine the driver's expected acceleration in a reference future time period based on multi-source vehicle information from a reference historical time period; determine the target driving mode to be switched to based on the expected acceleration; and control the vehicle to switch its driving mode to the target driving mode.

[0209] Optionally, the control module includes a switching judgment unit, which is used to: determine the target driving mode as manual driving mode when the expected acceleration in the reference future time period meets the first reference condition and the current driving mode of the vehicle is automatic driving mode; and determine the target driving mode as automatic driving mode when the expected acceleration in the reference future time period meets the second reference condition and the current driving mode of the vehicle is manual driving mode.

[0210] Optionally, the first reference condition includes: within a reference future time period, the duration for which the absolute value of the expected acceleration is continuously greater than a first threshold reaches a first duration threshold. The second reference condition includes: within a reference future time period, the duration for which the absolute value of the expected acceleration is continuously less than the second threshold reaches the second duration threshold.

[0211] Optionally, before determining the target driving mode as autonomous driving mode, the switching judgment unit is further configured to: obtain the steering wheel grip state of the vehicle; and determine the target driving mode as autonomous driving mode when the steering wheel grip state indicates that the steering wheel is being gripped.

[0212] Optionally, the switching judgment unit is used to control the vehicle's driving mode to switch to the target driving mode when the target driving mode is determined to be the same multiple times in a row.

[0213] Optionally, the control device further includes an execution module, which in turn includes a control decision unit. When the target driving mode is an autonomous driving mode, the control decision unit is used to: obtain the target acceleration of the vehicle based on a model predictive control algorithm according to the reference vehicle speed, the vehicle's current actual speed, the relative distance and relative speed with the vehicle in front; determine the target execution torque of the vehicle based on the target acceleration; and the execution module is used to control the vehicle according to the target execution torque.

[0214] Optionally, the control decision unit is configured to: smooth the target acceleration based on a reference acceleration change rate threshold to obtain a smoothed target acceleration; and determine the target execution torque of the vehicle based on the smoothed target acceleration.

[0215] Optionally, the execution module is configured to: acquire a torque weighting coefficient when a switch between driving and braking states is detected; determine a target braking torque and a target driving torque based on the torque weighting coefficient and the target driving torque; control the vehicle's braking system based on the target braking torque; and control the vehicle's driving system based on the target driving torque.

[0216] It should be noted that the apparatus provided in the above embodiments is only illustrated by the division of the above functional modules. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. In addition, the apparatus and method embodiments provided in the above embodiments belong to the same concept, and their specific implementation process can be found in the method embodiments, which will not be repeated here.

[0217] It is understandable that, since the control device has essentially the same technical effect as the vehicle control method described above, the technical effect of the control device will not be described again here for the sake of brevity.

[0218] In addition, please refer to Figure 5 This application also provides a vehicle, which may include a control device and a vehicle body, the control device being capable of performing the above-described actions. Figure 2 and Figure 3 The method shown is used to control the vehicle locally.

[0219] In this embodiment, the vehicle may further include a braking system and a drive system. The control device can be used to control the braking system and the drive system. Specific control procedures can be found in [reference needed]. Figure 2 and Figure 3 An example of the method is shown. Further details will not be provided here.

[0220] It is understandable that, since the vehicle and the vehicle control method described above have essentially the same technical effects, for the sake of brevity, the technical effects of the vehicle will not be described again here.

[0221] Furthermore, this application embodiment also provides a computer device, which includes a processor and a memory, wherein the memory stores at least one computer program. The at least one computer program is loaded and executed by one or more processors to enable the computer device to perform the functions or steps in the above-described method.

[0222] Furthermore, in this embodiment of the application, a computer-readable storage medium is provided, which stores at least one computer program, which is loaded and executed by the processor of a computer device to enable the computer to perform the functions or steps in the above-described method.

[0223] In one possible implementation, the aforementioned computer-readable storage medium may be a read-only memory (ROM), a random access memory (RAM), a compact disc read-only memory (CD-ROM), magnetic tape, floppy disk, and optical data storage device, etc.

[0224] Furthermore, this application embodiment also provides a computer program product or computer program, which includes computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the functions or steps in the above-described method.

[0225] It is understood that the steps of the methods or algorithms described in conjunction with the embodiments of this application can be implemented in hardware or by a processor executing software instructions. The software instructions can consist of corresponding software modules, which can be stored in random access memory, flash memory, erasable programmable read-only memory, electrically erasable programmable read-only memory, registers, hard disks, portable hard disks, read-only optical disks, or any other type of storage medium. An exemplary embodiment involves a storage medium coupled to a processor, enabling the processor to read information from and write information to the storage medium. Of course, the storage medium can also be a component of the processor. The processor and storage medium can reside in an application-specific integrated circuit (ASIC). Alternatively, the ASIC can reside within an electronic device. Of course, the processor and storage medium can also exist as discrete components within an electronic device.

[0226] In an alternative approach, when implemented using software, it can be implemented entirely or partially in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are implemented. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., digital video disk (DVD)), or a semiconductor medium (e.g., solid-state disk (SSD)).

[0227] It should be noted that all information (including but not limited to user device information, user personal information, etc.), data (including but not limited to data used for analysis, stored data, displayed data, etc.), and signals involved in this application have been authorized by the user or fully authorized by all parties, and the collection, use, and processing of related data must comply with the relevant laws, regulations, and standards of the relevant countries and regions. For example, the vehicle multi-source information and mode switching event information involved in this application were obtained with full authorization.

[0228] It should be noted that the terms "first," "second," etc. (if applicable) in the specification and claims of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0229] It should be understood that "multiple" as used in this article refers to two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone. The character " / " generally indicates that the preceding and following related objects have an "or" relationship.

[0230] The above description is merely a specific implementation of the embodiments of this application, but the protection scope of the embodiments of this application is not limited thereto. Any changes or substitutions within the technical scope disclosed in the embodiments of this application should be covered within the protection scope of the embodiments of this application. Therefore, the protection scope of the embodiments of this application should be determined by the protection scope of the claims.

Claims

1. A vehicle control method, characterized in that, The method includes: Obtain multi-source vehicle information for a reference historical time period, wherein the multi-source vehicle information is used to characterize the driver's operating behavior and the vehicle's operating status; Based on the multi-source vehicle information from the reference historical time period, determine the driver's expected acceleration in the reference future time period; Based on the desired acceleration, determine the target driving mode to be switched to by the vehicle; Control the vehicle's driving mode to switch to the target driving mode.

2. The method according to claim 1, characterized in that, Determining the target driving mode to be switched to by the vehicle based on the desired acceleration includes: If the expected acceleration in the reference future time period satisfies the first reference condition and the current driving mode of the vehicle is the automatic driving mode, the target driving mode is determined to be the manual driving mode. If the expected acceleration during the reference future time period satisfies the second reference condition and the current driving mode of the vehicle is manual driving mode, the target driving mode is determined to be automatic driving mode.

3. The method according to claim 2, characterized in that, The first reference condition includes: within the reference future time period, the duration for which the absolute value of the expected acceleration is continuously greater than a first threshold reaches a first duration threshold. The second reference condition includes: within the reference future time period, the duration for which the absolute value of the expected acceleration is continuously less than a second threshold reaches a second duration threshold.

4. The method according to claim 2, characterized in that, Before determining that the target driving mode is an autonomous driving mode, the method further includes: Obtain the steering wheel grip state of the vehicle; Determining the target driving mode as an autonomous driving mode includes: When the steering wheel grip status of the vehicle indicates that the steering wheel is being gripped, the target driving mode is determined to be the automatic driving mode.

5. The method according to claim 1, characterized in that, Switching the vehicle's driving mode to the target driving mode includes: If the target driving mode is determined to be the same in multiple consecutive tests, the vehicle's driving mode is switched to the target driving mode.

6. The method according to any one of claims 1-5, characterized in that, When the target driving mode is autonomous driving mode, the method further includes: Based on the reference vehicle speed, the vehicle's current actual speed, the relative distance and relative speed with the vehicle in front, the target acceleration of the vehicle is obtained using a model predictive control algorithm. The target execution torque of the vehicle is determined based on the target acceleration; The vehicle is controlled according to the target torque.

7. The method according to claim 6, characterized in that, Determining the target execution torque of the vehicle based on the target acceleration includes: Based on a reference acceleration change rate threshold, the target acceleration is smoothed to obtain a smoothed target acceleration. The target execution torque of the vehicle is determined based on the target acceleration after the smoothing process.

8. The method according to claim 6, characterized in that, The step of controlling the vehicle according to the target torque includes: When a switch between driving and braking states is detected, the torque weighting coefficient is obtained; The target braking torque and the target driving torque are determined based on the torque weighting coefficient and the target execution torque. The braking system of the vehicle is controlled according to the target braking torque, and the driving system of the vehicle is controlled according to the target driving torque.

9. A control device, characterized in that, The control device includes a signal acquisition module and a control module; The signal acquisition module is used to acquire multi-source vehicle information over a reference historical time period. The multi-source vehicle information is used to characterize the driver's operating behavior and the vehicle's operating status. The control module is used to determine the driver's expected acceleration in a reference future time period based on the vehicle multi-source information of the reference historical time period. Based on the desired acceleration, determine the target driving mode to be switched to by the vehicle; control the vehicle's driving mode to switch to the target driving mode.

10. A vehicle, characterized in that, The vehicle includes a control device and a vehicle body, the control device being configured to perform the method according to any one of claims 1-8 to control the vehicle body.

Images