Vehicle control method, device and system

Abstract

The invention provides a vehicle control method, device and system, and the method comprises the steps: obtaining the current driving data and driving environment data of a vehicle, carrying out the driving style prediction of the current driving data and driving environment data through a parameter prediction model pre-issued by a cloud end, and obtaining a driving style prediction result; the parameter prediction model is obtained by training the cloud according to the driving data of the vehicle, the driving environment data and the user behavior, and in response to a received parameter adjustment request of a user, the driving parameter initial value is adjusted according to the parameter adjustment request, and the current driving style and the driving parameter initial value corresponding to the current driving style are obtained. And adjusting the driving parameter value to obtain an adjusted driving parameter value, performing safety verification on the adjusted driving parameter value to obtain a driving parameter target value meeting a parameter safety boundary, and controlling the vehicle to drive according to the driving parameter target value. Through the cloud training parameter prediction model, the driving style and the driving parameter initial value of the user are predicted, and the requirement of the user for adjusting the driving parameters is met.

Description

Technical Field

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

[0002] With the rapid development of vehicle technology, driving mode settings have gradually become an important means of enhancing the driving experience. Currently, vehicles on the market offer driving modes such as Comfort, Sport, and Eco. These driving modes are typically bound to fixed parameters; that is, the chassis, powertrain, steering system, and other driving parameters are set together for each mode (Eco, Comfort, Sport, etc.). For example, Comfort mode usually sets a softer chassis and lighter steering wheel to provide a smooth driving experience; Sport mode sets a stiffer chassis and heavier steering wheel to enhance handling; and Eco mode improves fuel efficiency by reducing power output and enhancing regenerative braking.

[0003] However, while these driving modes meet users' driving needs to a certain extent, the fixed binding of driving parameters to driving modes means that users cannot adjust certain driving parameters individually based on their driving habits, real-time road conditions, and vehicle status when selecting a particular driving mode, thus making it difficult to meet users' personalized driving needs. Summary of the Invention

[0004] In view of this, the present invention aims to propose a vehicle control method, device and system to solve the problem that the driving parameters are fixed in the current driving mode, and the driving parameters cannot be adjusted individually, making it difficult to meet the personalized needs of users driving vehicles.

[0005] According to a first aspect of the present invention, a vehicle control method is provided, applied to a vehicle, wherein the vehicle communicates with a cloud, the method comprising: Acquire the vehicle's current driving data and driving environment data; Using the parameter prediction model pre-deployed from the cloud, driving style prediction is performed on the current driving data and the driving environment data to obtain the current driving style and the initial values ​​of the driving parameters corresponding to the current driving style; wherein, the parameter prediction model is trained by the cloud based on the vehicle's driving data, driving environment data, and user behavior; In response to receiving a user's parameter adjustment request, the initial value of the driving parameter is adjusted according to the parameter adjustment request to obtain the adjusted driving parameter value; The adjusted driving parameter values ​​are verified by using a predetermined parameter safety boundary to obtain target driving parameter values ​​that satisfy the parameter safety boundary. The vehicle is controlled to drive using the target values ​​of the driving parameters.

[0006] Optionally, the step of using the parameter prediction model pre-deployed from the cloud to predict the driving style of the current driving data and the driving environment data, and obtaining the current driving style and the initial values ​​of the driving parameters corresponding to the current driving style, includes: Receive the parameter prediction model sent from the cloud in advance, and deploy the parameter prediction model to the vehicle's local device; Feature extraction is performed on the current driving data and the driving environment data to obtain feature vectors corresponding to the current driving data and the driving environment data; The feature vector is input into the parameter prediction model to predict the driving style, and the current driving style and the initial values ​​of the driving parameters corresponding to the current driving style are obtained from the output of the parameter prediction model.

[0007] Optionally, the step of responding to receiving a user's parameter tuning request and adjusting the initial value of the driving parameters according to the parameter tuning request to obtain the adjusted driving parameters includes: In response to receiving a user's parameter adjustment request, the parameter adjustment request is identified, and the driving parameters to be adjusted and the parameter adjustment amount in the parameter adjustment request are determined; The initial value of the driving parameter to be adjusted is adjusted using the parameter adjustment amount to obtain the adjusted driving parameter value.

[0008] Optionally, the step of performing a safety check on the adjusted driving parameter values ​​using a pre-determined parameter safety boundary to obtain target driving parameter values ​​that satisfy the parameter safety boundary includes: The parameter safety boundary is received in advance from the cloud; wherein the parameter safety boundary is the safety threshold of each driving parameter determined by the cloud under different driving styles. The adjusted driving parameter values ​​are verified using the parameter safety boundary to determine whether the adjusted driving parameter values ​​exceed the parameter safety boundary. If the adjusted driving parameter value exceeds the parameter safety boundary, the adjusted driving parameter value is truncated to obtain the target driving parameter value that satisfies the parameter safety boundary.

[0009] Optionally, after the controlled vehicle drives using the target values ​​of the driving parameters, the method further includes: Obtain user feedback on the target values ​​of the driving parameters; The parameter prediction model deployed locally on the vehicle is fine-tuned based on the feedback information to obtain the fine-tuned parameter prediction model. The current driving data, driving environment data, feedback information, and the fine-tuned parameter prediction model are uploaded to the cloud. The cloud is used to iteratively update the parameter prediction model based on the current driving data, driving environment data, feedback information, and the fine-tuned parameter prediction model.

[0010] According to a second aspect of the present invention, another vehicle control method is provided, applied in a cloud environment, wherein the cloud environment communicates with a vehicle, the method comprising: The system collects vehicle driving data, driving environment data, and user behavior; wherein, the user behavior includes user parameter adjustment requests and feedback information. The driving data, driving environment data, and user behavior are fused to obtain a fusion result. The preset model is trained based on the fusion result to obtain a trained parameter prediction model; wherein, the parameter prediction model is used to output the driving style and the initial values ​​of the driving parameters corresponding to the driving style; The trained parameter prediction model is then sent to the vehicle.

[0011] Optionally, after the trained parameter prediction model is sent to the vehicle, the method further includes: Receives current driving data, driving environment data, feedback information, and fine-tuned parameter prediction model uploaded by the vehicle terminal; Based on the current driving data, driving environment data, feedback information, and the fine-tuned parameter prediction model, the parameter prediction model is iteratively updated.

[0012] According to a third aspect of the present invention, a vehicle control device is provided, applied to a vehicle, the vehicle communicating with a cloud, the device comprising: The data acquisition module is used to acquire the vehicle's current driving data and driving environment data; The parameter prediction module is used to predict the driving style of the current driving data and the driving environment data using a parameter prediction model pre-deployed from the cloud, and obtain the current driving style and the initial values ​​of the driving parameters corresponding to the current driving style; wherein, the parameter prediction model is trained by the cloud based on the vehicle's driving data, driving environment data and user behavior; The parameter adjustment module is used to respond to a user's parameter adjustment request, adjust the initial value of the driving parameter according to the parameter adjustment request, and obtain the adjusted driving parameter value. The parameter verification module is used to perform safety verification on the adjusted driving parameter values ​​through a pre-determined parameter safety boundary, so as to obtain the target driving parameter values ​​that meet the parameter safety boundary. The control module is used to control the vehicle to drive using the target values ​​of the driving parameters.

[0013] According to a fourth aspect of the present invention, another vehicle control device is provided, applied in a cloud environment, wherein the cloud environment communicates with a vehicle, the device comprising: The data acquisition module is used to collect vehicle driving data, driving environment data, and user behavior; wherein, the user behavior includes user parameter adjustment requests and feedback information; The feature processing module is used to fuse the driving data, driving environment data, and user behavior features to obtain a fusion result; The model training module is used to train a preset model based on the fusion result to obtain a trained parameter prediction model; wherein, the parameter prediction model is used to output the driving style and the initial values ​​of the driving parameters corresponding to the driving style; The model distribution module is used to distribute the trained parameter prediction model to the vehicle terminal.

[0014] According to a fifth aspect of the present invention, a vehicle control system is provided, the vehicle control system including a cloud and a vehicle terminal, the vehicle terminal communicating with the cloud, the system comprising: The vehicle-side system is used to acquire the vehicle's current driving data and driving environment data. Using a parameter prediction model pre-deployed from the cloud, it predicts the driving style based on the current driving data and the driving environment data to obtain the current driving style and the initial values ​​of the driving parameters corresponding to the current driving style. In response to receiving a user's parameter adjustment request, it adjusts the initial values ​​of the driving parameters according to the parameter adjustment request to obtain adjusted driving parameter values. The adjusted driving parameter values ​​are then subjected to safety verification through a pre-determined parameter safety boundary to obtain target values ​​of the driving parameters that satisfy the parameter safety boundary. Finally, the system controls the vehicle to drive using the target values ​​of the driving parameters. The cloud platform is used to collect vehicle driving data, driving environment data, and user behavior. The user behavior includes user parameter adjustment requests and feedback information. The driving data, driving environment data, and user behavior are fused to obtain a fusion result. Based on the fusion result, a preset model is trained to obtain a trained parameter prediction model. The parameter prediction model is used to output the driving style and the initial values ​​of the driving parameters corresponding to the driving style. The trained parameter prediction model is then sent to the vehicle.

[0015] The vehicle control method provided in this invention involves the vehicle acquiring current driving data and driving environment data, and using a parameter prediction model pre-deployed in the cloud to predict driving style based on the current driving data and driving environment data. This yields the current driving style and the initial values ​​of the corresponding driving parameters. The parameter prediction model is trained in the cloud based on the vehicle's driving data, driving environment data, and user behavior. In response to a user's parameter tuning request, the initial values ​​of the driving parameters are adjusted according to the request to obtain adjusted driving parameter values. The adjusted driving parameter values ​​are then verified against a pre-determined parameter safety boundary to obtain target values ​​of the driving parameters that satisfy the parameter safety boundary. Finally, the vehicle is controlled to drive using the target values ​​of the driving parameters. This invention utilizes a cloud-trained parameter prediction model, combined with current driving data and environmental data, to predict the user's driving style and generate corresponding initial values ​​for driving parameters. Users can interact with the vehicle to output parameter adjustment requests to fine-tune the driving parameters, enabling flexible adjustment. Through safety verification and dynamic adjustment, driving parameter values ​​that meet the parameter safety boundaries are obtained, ensuring vehicle safety and driving according to user needs. This avoids loss of vehicle control or dangerous situations caused by improper parameter settings, improves the safety of driving parameter regulation, and meets the personalized driving needs of users.

[0016] The above description is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention and to implement it in accordance with the contents of the specification, and in order to make the above and other objects, features and advantages of the present invention more apparent and understandable, specific embodiments of the present invention are described below. Attached Figure Description

[0017] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings: Figure 1 This is a flowchart of the steps of a vehicle control method provided in an embodiment of the present invention; Figure 2 This is a flowchart of another vehicle control method provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of the structure of a vehicle control device provided in an embodiment of the present invention; Figure 4 This is a schematic diagram of another vehicle control device provided in an embodiment of the present invention; Figure 5 This is a schematic diagram of the structure of a vehicle control system provided in an embodiment of the present invention. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the various embodiments of the present invention will be described in detail below with reference to the accompanying drawings. However, those skilled in the art will understand that many technical details are presented in the various embodiments of the present invention to facilitate a better understanding of the invention. However, the technical solutions claimed in the present invention can be implemented even without these technical details and with various changes and modifications based on the following embodiments. The division of the various embodiments below is for ease of description and should not constitute any limitation on the specific implementation of the present invention. The various embodiments can be combined with and referenced by each other without contradiction.

[0019] Reference Figure 1 This diagram illustrates a flowchart of a vehicle control method according to an embodiment of the present invention, applied to a vehicle-side device that communicates with a cloud platform. The method includes: Step 101: Obtain the vehicle's current driving data and driving environment data.

[0020] In this embodiment, the vehicle acquires real-time driving data and driving environment data through onboard sensors and vehicle controllers. The vehicle reads the current driving data from the CAN bus, which may include signals such as vehicle speed, acceleration, braking force, steering angle, suspension travel, vehicle posture, wheel speed, and tire pressure. The vehicle acquires driving environment data through sensors such as cameras, lidar, and GPS, which includes data such as road type, road surface smoothness, weather information, traffic density, and light intensity. Road types include urban, highway, rural, and mountain roads, and weather information includes rainfall, temperature, and humidity.

[0021] Step 102: Using the parameter prediction model pre-deployed in the cloud, predict the driving style based on the current driving data and driving environment data to obtain the current driving style and the initial values ​​of the driving parameters corresponding to the current driving style; wherein, the parameter prediction model is trained in the cloud based on the vehicle's driving data, driving environment data and user behavior.

[0022] In this embodiment of the invention, the vehicle uses a parameter prediction model pre-deployed from the cloud to predict the driving style based on the current driving data and driving environment data, thereby obtaining the current driving style and the initial values ​​of the driving parameters corresponding to the current driving style. It should be noted that this embodiment is for adjusting the driving parameters of the vehicle. The adjustable driving parameters in the vehicle mainly include adjustable driving data such as chassis stiffness, suspension damping, steering wheel power assist, power response sensitivity, electronic differential lock locking logic, ESC (Electronic Stability Control) response threshold, transmission shift time, and torque distribution strategy. Therefore, the current driving style and the initial values ​​of the driving parameters corresponding to the current driving style output by the parameter prediction model are the initial recommended values ​​of the adjustable driving parameters.

[0023] In this embodiment, the vehicle receives the parameter prediction model pre-downloaded from the cloud and deploys it locally. This parameter prediction model is trained by the cloud based on vehicle driving data, driving environment data, and user behavior, which will not be elaborated upon here. Specifically, the cloud uses OTA (Over-The-Air) technology to distribute the trained parameter prediction model to the vehicle. The vehicle receives the pre-trained parameter prediction model from the cloud and stores it locally. The parameter prediction model is typically encapsulated to accommodate the computing power and storage limitations of the vehicle's embedded devices. The vehicle system then loads the parameter prediction model into the inference engine to ensure the model can run in real-time on the vehicle.

[0024] Specifically, the current driving data and driving environment data are preprocessed to extract feature vectors such as time-domain statistical features and frequency-domain features. The extracted feature vectors are then input into the parameter prediction model to obtain the probability distribution of the current driving style and the corresponding initial values ​​of driving parameters. The initial values ​​of driving parameters include the predicted initial values ​​of adjustable parameters such as chassis stiffness, suspension damping, and steering wheel power assist. In this embodiment, the driving style is defined based on features such as acceleration changes, steering angle changes, suspension system preferences, and shift logic in the driving data of different users in various driving environments. Driving styles include different user experiences, such as comfort, fuel efficiency, and sportiness; different driving scenarios, such as nighttime and rainy days; and different user preferences, such as aggressive, off-road, and stable driving styles. In this embodiment, the driving style is learned and defined based on historical driving data, historical driving environment data, and user driving behavior preferences of various vehicle models, road conditions, climates, and various user driving habits. The specific definition is not limited.

[0025] Step 103: In response to receiving the user's parameter adjustment request, adjust the initial values ​​of the driving parameters according to the parameter adjustment request to obtain the adjusted driving parameter values.

[0026] In this embodiment of the invention, the vehicle monitors the user's parameter adjustment request in real time. The parameter adjustment request includes interactive forms such as touch interface and voice command. Upon receiving the user's parameter adjustment request, the vehicle identifies the parameter adjustment request and determines the driving parameters to be adjusted and the parameter adjustment amount in the parameter adjustment request. The driving parameters to be adjusted are the driving parameters that need to be adjusted in the user's parameter adjustment request, and the parameter adjustment amount is the parameter increase or decrease amount determined based on the parameter adjustment request.

[0027] In this embodiment, after obtaining the current driving style and the corresponding initial values ​​of driving parameters, the user will adjust the initial values ​​of driving parameters. The vehicle responds by receiving the user's parameter adjustment request, identifying the parameter adjustment request, determining the driving parameter to be adjusted and the parameter adjustment amount in the parameter adjustment request, and adjusting the initial value of the driving parameter to be adjusted using the parameter adjustment amount to obtain the adjusted driving parameter value.

[0028] Specifically, users can issue parameter adjustment requests through the vehicle's touchscreen interface or voice commands (such as "steering wheel assist +30%)". The vehicle parses the parameter adjustment request and converts it into specific parameter increments. The vehicle system recognizes the direction and percentage of movement of the slider or knob in the touchscreen interface to determine the name of the driving parameter to be adjusted and the amount to be adjusted. It should be noted that the cloud pre-generates a mapping relationship between voice intent and parameter adjustment amount based on user behavior and stores it as a voice mapping table. This table can be sent to the vehicle along with the parameter prediction model or sent to the vehicle separately. When a user issues a parameter adjustment command via voice, such as "reduce suspension damping by 20%", the vehicle performs voice recognition, converts the voice into text, and parses the name of the parameter to be adjusted, "suspension damping", and the parameter adjustment amount, "-20%", through the voice mapping table.

[0029] Step 104: Perform a safety check on the adjusted driving parameter values ​​using a pre-determined parameter safety boundary to obtain the target driving parameter values ​​that meet the parameter safety boundary.

[0030] In this embodiment of the invention, the vehicle receives parameter safety boundaries from the cloud in advance. The parameter safety boundaries are safety thresholds for each driving parameter under different driving styles determined by the cloud. The adjusted driving parameter values ​​are then checked against the pre-determined parameter safety boundaries to determine whether the adjusted driving parameter values ​​exceed the parameter safety boundaries. Boundary restrictions are then applied to the adjusted driving parameter values ​​to obtain target values ​​for driving parameters that meet the parameter safety boundaries. In other words, if the adjusted driving parameter values ​​exceed the parameter safety boundaries, the adjusted driving parameter values ​​are truncated.

[0031] In this embodiment, the vehicle receives parameter safety boundaries from the cloud in advance. These safety boundaries are safety thresholds set by the cloud for each driving parameter under different driving styles, based on a large amount of driving data and user feedback, combined with vehicle hardware limitations and driving safety regulations. The parameter safety boundaries include the driving parameter name, driving style, and the safety threshold for the driving parameter under the driving style. The safety thresholds include minimum and maximum values. The parameter safety boundaries ensure that the adjustment of driving parameters does not exceed the safety range of the vehicle hardware, avoiding driving risks caused by improper parameter settings.

[0032] In this embodiment, a safety boundary is used to perform a safety check on the adjusted driving parameter values ​​to determine whether the adjusted driving parameter values ​​exceed the safety boundary. If the adjusted driving parameter values ​​exceed the safety boundary, the adjusted driving parameter values ​​are truncated to obtain the target driving parameter values ​​that satisfy the safety boundary, i.e., the truncated values ​​are taken as the boundary values ​​of the safety boundary. If the parameter value is less than the lower boundary of the safety boundary, it is adjusted to the minimum value of the safety boundary. If the parameter value is greater than the upper boundary of the safety boundary, it is adjusted to the maximum value of the safety boundary. The target driving parameter values ​​that satisfy the safety boundary are obtained and the truncated target driving parameter values ​​are displayed in real time on the vehicle interface or in voice prompts to ensure that users can understand the adjustment results in a timely manner.

[0033] Step 105: Control the vehicle to drive using the target values ​​of driving parameters.

[0034] In this embodiment of the invention, after determining the target values ​​of the driving parameters, the vehicle sends these target values ​​to various control units in the vehicle to control the vehicle to drive using the target values. Specifically, the safety-verified target values ​​of the driving parameters are sent to vehicle controllers (such as chassis control and steering control) via protocols such as CAN / CAN-FD to control the vehicle to drive using new parameters. For example, adjustable parameters such as chassis stiffness, suspension damping, and steering assist are sent to the corresponding controllers, such as ECU (Electronic Control Unit), MDC (Motor Drive Controller), VDC (Vehicle Dynamics Control), CDC (Continuous Damping Control), etc., so that the vehicle drives according to the target values ​​of the driving parameters.

[0035] It should be noted that the vehicle's driving characteristics (such as chassis stiffness, suspension damping, etc.) are adjusted according to the received target values ​​of driving parameters, and the driving parameter status is confirmed to the user through the vehicle interface or voice terminal. Through the vehicle interface and voice feedback, the user can understand the current driving parameter settings in real time, forming a complete interactive loop.

[0036] The vehicle control method provided in this invention involves the vehicle acquiring current driving data and driving environment data, and using a parameter prediction model pre-deployed in the cloud to predict driving style based on the current driving data and driving environment data. This yields the current driving style and the initial values ​​of the corresponding driving parameters. The parameter prediction model is trained in the cloud based on the vehicle's driving data, driving environment data, and user behavior. In response to a user's parameter tuning request, the initial values ​​of the driving parameters are adjusted according to the request to obtain adjusted driving parameter values. The adjusted driving parameter values ​​are then verified against a pre-determined parameter safety boundary to obtain target values ​​of the driving parameters that satisfy the parameter safety boundary. Finally, the vehicle is controlled to drive using the target values ​​of the driving parameters. This invention utilizes a cloud-trained parameter prediction model, combined with current driving data and environmental data, to predict the user's driving style and generate corresponding initial values ​​for driving parameters. Users can interact with the vehicle to output parameter adjustment requests to fine-tune the driving parameters, enabling flexible adjustment. Through safety verification and dynamic adjustment, driving parameter values ​​that meet the parameter safety boundaries are obtained, ensuring vehicle safety and driving according to user needs. This avoids loss of vehicle control or dangerous situations caused by improper parameter settings, improves the safety of driving parameter regulation, and meets the personalized driving needs of users.

[0037] In some embodiments, step 102 uses a parameter prediction model pre-deployed from the cloud to predict the driving style based on the current driving data and driving environment data, obtaining the current driving style and the initial values ​​of the driving parameters corresponding to the current driving style. Specifically, this may include the following steps: S11 receives the parameter prediction model from the cloud in advance and deploys the parameter prediction model to the vehicle's local terminal. S12, extract features from the current driving data and driving environment data to obtain the feature vectors corresponding to the current driving data and driving environment data; S13, input the feature vector into the parameter prediction model to predict the driving style, and obtain the current driving style and the initial values ​​of the driving parameters corresponding to the current driving style output by the parameter prediction model.

[0038] In this embodiment of the invention, the vehicle receives a parameter prediction model pre-deployed from the cloud and deploys it locally. Specifically, the cloud distributes the trained parameter prediction model to the vehicle via OTA (Over-The-Air). The vehicle receives the pre-trained parameter prediction model from the cloud and stores it locally. The parameter prediction model is typically encapsulated to accommodate the computing power and storage limitations of the embedded vehicle device. The vehicle system loads the parameter prediction model into the inference engine to ensure the model can run in real-time on the vehicle. During deployment, the vehicle system can also directly load the parameter prediction model's configuration file, including information such as feature dimensions and safety boundaries. It should be noted that the parameter prediction model is trained in the cloud based on a large amount of driving data, driving environment data, and user behavior data. This embodiment will not elaborate on the cloud-trained parameter prediction model.

[0039] In this embodiment, the vehicle-side performs feature dimension processing on the current driving data and driving environment data to meet the model input requirements. Specifically, feature extraction is performed on the current driving data and driving environment data to obtain feature vectors corresponding to the current driving data and driving environment data. The current driving data may include signals such as vehicle speed, acceleration, braking force, steering angle, suspension travel, vehicle posture, wheel speed, and tire pressure. The driving environment data includes data such as road type, road surface smoothness, weather information, traffic density, and light intensity. Feature extraction is performed on the collected current driving data and driving environment data to ensure that the feature vector is consistent with the feature processing used when training the parameter prediction model in the cloud, so as to meet the model input requirements. Specifically, temporal feature extraction is performed. Sliding window processing is applied to temporal data such as vehicle speed, acceleration, and steering angle to extract temporal statistical features such as mean, variance, extreme values, and zero crossover rate. Category data such as road type and weather information are encoded or numerically processed. High-frequency sensor data (such as suspension travel and vehicle posture) are downsampled to reduce the amount of data and reduce computing power consumption. The extracted features are combined into a multi-dimensional feature vector, which is used as the input of the parameter prediction model.

[0040] In this embodiment, the feature vector is input into the parameter prediction model to predict the driving style, obtaining the current driving style and the initial values ​​of the corresponding driving parameters output by the parameter prediction model. Specifically, the feature vectors corresponding to the extracted current driving data and driving environment data are input into the parameter prediction model deployed on the vehicle for forward inference. The model inference process includes driving style prediction and determining the initial values ​​of driving parameters. The model outputs the probability distribution of the current driving style and determines the current driving style based on the probability distribution. Based on the current driving style, the model outputs the corresponding initial values ​​of driving parameters, such as chassis stiffness, suspension damping, steering wheel power assist, power response, and shift logic. It should be noted that in this embodiment, the driving style is defined by learning from historical driving data, historical driving environment data, and user driving behavior preferences, and is not specifically limited to any particular style.

[0041] In some embodiments, the feature vector is input into the parameter prediction model to predict the driving style. After obtaining the current driving style and the initial values ​​of the driving parameters corresponding to the current driving style output by the parameter prediction model, the current driving style and the initial values ​​of the parameters are displayed in real time on the vehicle interface or voice prompts to ensure that users can understand the results recommended by the system in a timely manner.

[0042] This invention utilizes a cloud-based parameter prediction model and vehicle-side feature extraction of driving data. By employing the cloud-based parameter prediction model, driving style prediction is accurately and efficiently obtained, and initial values ​​of driving parameters are determined.

[0043] In some embodiments, step 103, in response to receiving a user's parameter tuning request, adjusts the initial values ​​of the driving parameters according to the parameter tuning request to obtain adjusted driving parameter values, may specifically include the following steps: S21, in response to receiving the user's parameter adjustment request, identify the parameter adjustment request and determine the driving parameters to be adjusted and the parameter adjustment amount in the parameter adjustment request; S22, the initial value of the driving parameter to be adjusted is used to adjust the driving parameter value to obtain the adjusted driving parameter value.

[0044] In this embodiment of the invention, after obtaining the current driving style and the corresponding initial values ​​of driving parameters, the user will adjust the initial values ​​of driving parameters. Therefore, the vehicle responds to the user's parameter adjustment request by identifying the parameter adjustment request, determining the driving parameters to be adjusted and the parameter adjustment amount in the parameter adjustment request, and adjusting the initial values ​​of the driving parameters to be adjusted using the parameter adjustment amount to obtain the adjusted driving parameter values.

[0045] Specifically, the vehicle-mounted system can receive user parameter adjustment requests through various interaction methods. These requests are typically explicit inputs, including user touchscreen controls and voice control. Users can adjust driving parameters using sliders, knobs, or buttons on the vehicle's screen. For example, a user dragging the "steering wheel assist" slider 30% to the right indicates a desire to increase steering wheel assist by 30%. Alternatively, a user can issue a parameter adjustment command via voice, such as "reduce suspension damping by 20%." The vehicle will then perform voice recognition, convert the speech to text, and use natural language processing to extract the names and adjustment amounts of the parameters to be adjusted, namely "suspension damping" and "-20%".

[0046] In this embodiment, the vehicle-side system parses the user's parameter adjustment request, identifying the driving parameters to be adjusted and the adjustment amounts. For the vehicle's touchscreen interface, the system identifies the direction and percentage of movement of the slider or knob to determine the parameter name and adjustment amount. For voice control, the system uses voice recognition to parse the name of the parameter to be adjusted and the adjustment amount slot. The parsed parameter name and adjustment amount are then mapped to specific driving parameters and values. For example, "reduce suspension damping by 20%" is mapped to a 20% reduction in the initial value of the suspension damping parameter. Users can adjust parameters through various interaction methods such as vehicle touchscreen and voice control. The vehicle-side system can accurately identify the user's parameter adjustment request, ensuring the correct parsing of parameter names and adjustment amounts and avoiding misoperation.

[0047] In this embodiment, the initial values ​​of the driving parameters to be adjusted are adjusted using parameter adjustment amounts to obtain the adjusted driving parameter values. The vehicle-side system, based on the initial values ​​of the driving parameters corresponding to the current driving style, adjusts the initial values ​​of the driving parameters according to the parsed parameter adjustment amounts. The adjusted driving parameter values ​​are displayed in real-time on the vehicle interface or through voice prompts, ensuring that the user can promptly understand the adjustment results. For example, if the current driving style is "Comfort," the corresponding initial suspension damping value is 50%, and the initial steering wheel assist value is 40%. If the user requests "suspension damping -20%," the initial suspension damping value is reduced by 20% from 50%, resulting in an adjusted suspension damping value of 30%. If the user requests "steering wheel assist +30%," the initial steering wheel assist value is increased by 30% from 40%, resulting in an adjusted steering wheel assist value of 70%. Of course, the above are only specific examples, and this embodiment does not specifically limit the adjustment amount of the adjustable driving parameters.

[0048] This invention provides a variety of interactive methods, allowing users to personalize driving parameters according to their needs. It supports independent adjustment of multiple driving parameters, and users can flexibly select the parameters they need to adjust, thus realizing personalized adjustment of driving parameters.

[0049] In some embodiments, step 104 performs a safety check on the adjusted driving parameter values ​​using a predetermined parameter safety boundary to obtain target driving parameter values ​​that satisfy the parameter safety boundary. This may specifically include the following steps: S31, receive the parameter safety boundary sent by the cloud in advance; wherein, the parameter safety boundary is the safety threshold of each driving parameter determined by the cloud under different driving styles; S32, The adjusted driving parameter value is verified using the parameter safety boundary to determine whether the adjusted driving parameter value exceeds the parameter safety boundary; S33, if the adjusted driving parameter value exceeds the parameter safety boundary, the adjusted driving parameter value is truncated to obtain the driving parameter target value that satisfies the parameter safety boundary.

[0050] In this embodiment of the invention, the vehicle receives parameter safety boundaries from the cloud in advance. These safety boundaries are safety thresholds set by the cloud for each driving parameter under different driving styles, based on extensive driving data and user feedback, combined with vehicle hardware limitations and driving safety regulations. They are used to verify whether the adjusted driving parameter values ​​meet safety requirements. The cloud can send these safety boundaries to the vehicle via OTA (Over-The-Air) technology, and the vehicle system receives and stores them. The safety boundaries include the driving parameter name, driving style, and the safety threshold for the driving parameter under that driving style. The safety thresholds include minimum and maximum values. The vehicle system stores the received safety boundaries in a local database, categorized by driving parameter and driving style, for quick subsequent retrieval and use. The safety boundaries ensure that adjustments to driving parameters do not exceed the vehicle hardware's safety limits, avoiding driving risks caused by improper parameter settings. The safety boundaries can be dynamically adjusted according to different driving styles to adapt to driving needs in different scenarios. Through OTA technology, the cloud can update the safety boundaries in real time, ensuring that the vehicle system always uses the latest safety thresholds.

[0051] In this embodiment, a parameter safety boundary is used to perform a safety check on the adjusted driving parameter values ​​to determine whether the adjusted driving parameter values ​​exceed the parameter safety boundary. Specifically, based on the current driving style, the corresponding parameter safety boundary is queried from the local database, and the adjusted driving parameter values ​​are checked to determine whether they exceed the safety boundary. For each adjusted driving parameter value, if the parameter value is less than the minimum value of the safety boundary, it is marked as "exceeding the lower boundary"; if the parameter value is greater than the maximum value of the safety boundary, it is marked as "exceeding the upper boundary"; if the parameter value is within the safety boundary range, it is marked as "meeting the safety boundary". If the adjusted driving parameter value exceeds the parameter safety boundary, the adjusted driving parameter value is truncated to obtain the target driving parameter value that meets the parameter safety boundary. That is, if the parameter is marked as exceeding the upper boundary or exceeding the lower boundary, the adjusted driving parameter value exceeds the safety boundary and is truncated to the boundary value of the safety boundary. If the parameter value is less than the lower boundary of the safety boundary, it is adjusted to the minimum value of the safety boundary; if the parameter value is greater than the upper boundary of the safety boundary, it is adjusted to the maximum value of the safety boundary. The captured driving parameter values ​​are recorded as driving parameter target values. The driving parameter target values ​​that meet the parameter safety boundaries are obtained and displayed in real time on the vehicle interface or voice prompts to ensure that users can understand the adjustment results in a timely manner.

[0052] This invention ensures that the target values ​​of driving parameters are always within a safe range by verifying and intercepting the safety boundaries of parameters, thus avoiding driving risks caused by improper parameter settings, improving the safety and real-time performance of driving parameter control, and further enhancing the user experience.

[0053] In some embodiments, after step 105 controls the vehicle to drive using target values ​​of driving parameters, it may further include: Obtain user feedback on the target values ​​of the driving parameters; The parameter prediction model deployed locally on the vehicle is fine-tuned based on the feedback information to obtain the fine-tuned parameter prediction model. The current driving data, driving environment data, feedback information, and the fine-tuned parameter prediction model are uploaded to the cloud. The cloud is used to iteratively update the parameter prediction model based on the current driving data, driving environment data, feedback information, and the fine-tuned parameter prediction model.

[0054] In this embodiment, after the vehicle completes the adjustment of the target values ​​of the driving parameters, the user's feedback information on the current driving experience is obtained. The feedback information is the user's experience feedback based on the target values ​​of the driving parameters. A feedback interface can be displayed on the vehicle's infotainment system. The user can express their feelings about the adjusted driving parameters by clicking the "Satisfied", "Neutral", or "Dissatisfied" buttons. The user can also express feedback directly by voice, such as saying "Too stiff", "Too soft", "Not enough power", "The steering wheel is a bit heavy", etc., which contain personal preferences. The vehicle recognizes the user's feedback information and records the user's feedback information (such as satisfaction rating, voice feedback text, etc.) locally for subsequent fine-tuning of the model and data uploading.

[0055] Specifically, the vehicle-side system fine-tunes the locally deployed parameter prediction model based on user feedback, resulting in a refined parameter prediction model. User feedback is processed to extract key features and labels. For example, if a user reports "the suspension is too stiff," then "suspension damping" is extracted as a key parameter and marked as "unsatisfactory." The feedback is then used to fine-tune the locally deployed parameter prediction model. This feedback is combined with current driving data and driving environment data to generate new training samples, which are then used to further refine the parameter prediction model, such as adjusting model weights. This fine-tuning process is lightweight and efficient, completed quickly, ensuring the system's real-time responsiveness. Through continuous fine-tuning based on user feedback, the model can continuously evolve, improving prediction accuracy and user experience.

[0056] In this embodiment, the vehicle uploads current driving data, driving environment data, feedback information, and a fine-tuned parameter prediction model to the cloud. Current driving data includes vehicle speed, acceleration, steering angle, and suspension travel; driving environment data includes road type and weather information; and feedback information includes user satisfaction ratings and voice feedback text. The vehicle also uploads the locally fine-tuned model parameters to the cloud for iterative model updates. Upon receiving the data from the vehicle, the cloud uses the current driving data, driving environment data, uploaded feedback information, and fine-tuned model parameters to iteratively update the global parameter prediction model.

[0057] This invention forms a closed-loop feedback mechanism through user feedback, local fine-tuning, and cloud iteration, continuously optimizing the parameter prediction model and providing data support for the global optimization of the cloud model. It can combine local fine-tuning and global data to iteratively update the model, improving the personalization and accuracy of driving parameter control.

[0058] Reference Figure 2 This diagram illustrates a flowchart of another vehicle control method provided by an embodiment of the present invention, applied in a cloud environment, wherein the cloud environment communicates with the vehicle. The method may include: Step 201: Collect vehicle driving data, driving environment data, and user behavior; among which, user behavior includes user parameter adjustment requests and feedback information.

[0059] In this embodiment of the invention, the cloud, through communication with the vehicle, trains a parameter prediction model based on massive and diverse vehicle driving and user feedback data. This model can output multi-dimensional adjustable driving parameters under various driving styles according to multimodal inputs and outputs, covering the driving needs of most vehicle configurations and user parameter tuning. Specifically, the cloud can collect vehicle driving data, driving environment data, and user behavior, i.e., raw data of various vehicle models, road conditions, climates, and various user driving habits. Driving data includes real-time signals uploaded from the vehicle, such as vehicle speed, acceleration, braking force, steering angle, suspension travel, vehicle posture, wheel speed, and tire pressure. Driving environment data includes data uploaded from the vehicle, such as road type (city, highway, mountain road), road surface smoothness, weather information (rainfall, temperature, humidity), traffic density, and light intensity. User behavior includes user parameter tuning requests and feedback information, such as user parameter tuning requests like "steering wheel power assist +30%", or feedback information like "slight vibration in the rear seats" or "steering too heavy", as well as satisfaction ratings after the drive.

[0060] It should be noted that the cloud collects multi-dimensional driving data, environmental data, and user behavior to comprehensively reflect the vehicle's driving status and user preferences, providing a high-quality data foundation for subsequent feature fusion and model training. Furthermore, the real-time collection of current driving data and environmental data, combined with historical data, forms time-series data, which helps the model learn the user's long-term driving habits and preferences.

[0061] Step 202: Feature fusion of driving data, driving environment data, and user behavior is performed to obtain the fusion result.

[0062] In this embodiment of the invention, the collected driving data, driving environment data, and user behavior are fused to form a unified feature vector. Specifically, the driving data is subjected to temporal statistical feature extraction, such as mean speed, acceleration variance, and steering angle range; the driving environment data is encoded, such as road type encoding and weather condition numericalization; and the user behavior is parsed, such as voice commands being parsed into specific parameter increments and user feedback ratings being converted into preference labels. Multimodal fusion technology (such as Transformer encoder or cross-attention mechanism) is used to fuse temporal features, environmental features, and user behavior features to form a unified feature representation, thus obtaining the fusion result of multimodal data.

[0063] In some embodiments, the cloud performs preprocessing on driving data, driving environment data, and user behavior, including alignment, denoising, and missing data filling. The preprocessed driving data, driving environment data, and user behavior are then fused using features. Specifically, the driving data, driving environment data, and user behavior are spatiotemporally aligned to obtain multimodal data samples at the same timestamp. For numerical signals such as speed, steering angle, and suspension displacement, median filtering or Kalman filtering is used to eliminate sudden pulses. For user manual / touch / voice input, a confidence threshold is added after text recognition, discarding invalid commands with excessively low confidence, thus completing anomaly removal and denoising. For short-term missing data caused by sensor frame loss, forward or backward linear interpolation is used for completion; long-term missing data is marked as incomplete data samples and discarded or separately categorized. This embodiment will not elaborate on these aspects.

[0064] Specifically, within each time window (e.g., 1s, 5s, 10s), the mean, variance, range, and zero-crossing rate of speed / acceleration / steering angle / suspension travel are calculated in the time domain. Short-time Fourier transform (STFT) is performed on the suspension travel or vehicle vibration signal to extract the energy spectrum of the bump frequency band, obtaining frequency domain features used to determine road smoothness or bump level. User behavior features can be based on the number of breakpoints and gradient change frequencies of acceleration / braking / steering, statistically analyzing "number of rapid accelerations," "number of rapid brakings," and "number of sharp turns" to quantify driving "aggression" or "smoothness." NLP segmentation and slot filling are performed on historical user voice commands, parsing phrases like "please soften the chassis" into specific numerical increments, such as "chassis damping -10%." Road type (city, highway, mountain road) is encoded; weather conditions (sunny / rainy, temperature, humidity) are used as numerical or categorical features; and time periods (day / night) are binarized to obtain driving environment features.

[0065] Step 203: Train the preset model based on the fusion result to obtain the trained parameter prediction model; wherein, the parameter prediction model is used to output the driving style and the initial values ​​of the driving parameters corresponding to the driving style.

[0066] In this embodiment of the invention, the cloud trains a preset model based on the fusion results of multimodal data to obtain a trained parameter prediction model. The preset model includes a temporal modeling layer, a multimodal fusion layer, and a fully connected layer. The temporal modeling layer (such as LSTM / GRU) is used to perform temporal dependency learning on the sequential temporal and frequency domain features. The multimodal fusion layer (such as a Transformer encoder or cross-attention mechanism) is used to fuse the temporal output with environmental / voice intent features to form a unified expression. The fully connected layer is used to feed the fusion result into several fully connected network layers to output the driving style probability distribution and the initial prediction values ​​of each adjustable driving parameter.

[0067] It should be noted that, based on a large amount of driving data and user feedback, combined with vehicle hardware limitations and driving safety regulations, the cloud sets safety thresholds for each driving parameter under different driving styles, resulting in parameter safety boundaries. These boundaries are used by the vehicle to verify whether the adjusted driving parameter values ​​meet safety requirements. The parameter safety boundaries include the driving parameter name, driving style, and the safety threshold for the driving parameter under that driving style. The safety thresholds include minimum and maximum values. The cloud can send the parameter safety boundaries to the vehicle via OTA technology. The parameter safety boundaries ensure that the adjustment of driving parameters does not exceed the safety range of the vehicle hardware, avoiding driving risks caused by improper parameter settings. The parameter safety boundaries can be dynamically adjusted according to different driving styles to adapt to driving needs in different scenarios. The cloud can update the parameter safety boundaries in real time to ensure that the vehicle system always uses the latest safety thresholds.

[0068] In this embodiment, cross-entropy loss can be used to optimize driving style probability prediction, or mean squared error (MSE) or Huber loss can be used to optimize the prediction of initial driving parameter values. MSE is a loss function that measures the difference between predicted and true values, while Huber loss is a loss function that is robust to outliers. The weights of classification loss and regression loss are determined through grid search or Bayesian optimization to balance the accuracy of driving style prediction and parameter prediction. Furthermore, to optimize the parameter prediction model, K-fold cross-validation is performed on different scenarios (city, highway, mountain road, rainy day) and different user groups (stable, aggressive, economical). K-fold cross-validation involves dividing the sample set into k parts, using k-1 parts as the training dataset and the remaining part as the validation dataset. The validation dataset is used to validate the model output. This process typically requires k iterations until all k parts of data have been selected, ensuring the model's stability under various conditions and verifying whether the generated parameter combinations always fall within the safe boundary. If the boundary is exceeded, a boundary penalty term is added during the training phase; no specific limitations are specified here.

[0069] Step 204: Send the trained parameter prediction model to the vehicle.

[0070] In this embodiment of the invention, the cloud distributes the trained parameter prediction model to the vehicle via OTA (Over-The-Air). Specifically, the weights, network structure configuration, and safety boundaries of the parameter prediction model can be packaged into a lightweight inference engine or script that can run on the vehicle. Through the communication channel between the cloud and the vehicle, the packaged model is distributed to the vehicle, ensuring that the vehicle can load and use the latest model in real time. A version number can be assigned to each distributed model, and the updated content and optimization points of the model can be recorded for easy vehicle management and rollback.

[0071] The vehicle control method provided in this invention collects vehicle driving data, driving environment data, and user behavior in the cloud. User behavior includes user parameter adjustment requests and feedback information. The driving data, driving environment data, and user behavior are fused to obtain a fusion result. Based on the fusion result, a preset model is trained to obtain a trained parameter prediction model. The parameter prediction model outputs driving style and corresponding initial values ​​of driving parameters. The trained parameter prediction model is then sent to the vehicle. This invention, through a cloud-trained parameter prediction model combined with current driving data and environmental data, predicts the user's driving style and generates corresponding initial values ​​of driving parameters. The user can interact with the vehicle to output parameter adjustment requests to fine-tune the driving parameters, enabling flexible adjustment. Through safety checks and dynamic adjustments, driving parameter values ​​that meet safety boundaries are obtained, ensuring vehicle safety and driving according to user needs. This avoids loss of control or dangerous situations caused by improper parameter settings, improves the safety of driving parameter control, and meets the personalized driving needs of users.

[0072] In some embodiments, after step 204, which involves sending the trained parameter prediction model to the vehicle, the method may further include: Receives current driving data, driving environment data, feedback information, and fine-tuned parameter prediction model uploaded by the vehicle terminal; Based on the current driving data, driving environment data, feedback information, and the fine-tuned parameter prediction model, the parameter prediction model is iteratively updated.

[0073] In this embodiment of the invention, the cloud receives current driving data, driving environment data, feedback information, and a fine-tuned parameter prediction model uploaded from the vehicle. Based on these data, the cloud iteratively updates the parameter prediction model. Specifically, the cloud can fuse the received driving data, environment data, and feedback information with historical data stored in the cloud to generate a larger-scale training dataset. This fused dataset is then used to iteratively update the global parameter prediction model. Deep learning algorithms (such as Transformer or LSTM) are used to train the fused data, optimizing the model's prediction performance. The model's performance is evaluated on a validation set to ensure that it has not overfitted or degraded. Finally, the iteratively updated parameter prediction model is deployed to the vehicle via OTA (Over-The-Air) technology for use by the vehicle's system.

[0074] Compared with the prior art, the embodiments of the present invention, based on achieving the beneficial effects of the first embodiment, collect vehicle driving data, environmental data and user behavior, and combine feature fusion and model training to obtain a parameter prediction model that can output driving style and driving parameter predictions, effectively improving the accuracy of driving style and driving parameter predictions, and sending new parameter prediction models to the vehicle in real time to provide personalized driving parameter recommendations, thereby enhancing the user's personalized and intelligent driving experience.

[0075] Reference Figure 3 This diagram illustrates a vehicle control device according to an embodiment of the present invention, applied to a vehicle, wherein the vehicle communicates with a cloud, and the device includes: The data acquisition module 301 is used to acquire the vehicle's current driving data and driving environment data; The parameter prediction module 302 is used to predict the driving style of the current driving data and the driving environment data using the parameter prediction model pre-deployed by the cloud, so as to obtain the current driving style and the initial values ​​of the driving parameters corresponding to the current driving style; wherein, the parameter prediction model is trained by the cloud based on the vehicle's driving data, driving environment data and user behavior; The parameter adjustment module 303 is used to respond to a user's parameter adjustment request, adjust the initial value of the driving parameter according to the parameter adjustment request, and obtain the adjusted driving parameter value. The parameter verification module 304 is used to perform safety verification on the adjusted driving parameter values ​​through a predetermined parameter safety boundary to obtain a driving parameter target value that satisfies the parameter safety boundary. The control module 305 is used to control the vehicle to drive using the target values ​​of the driving parameters.

[0076] Optionally, the parameter prediction module 302 includes: The first receiving submodule is used to receive the parameter prediction model sent from the cloud in advance and deploy the parameter prediction model to the local machine; The processing submodule is used to extract features from the current driving data and the driving environment data to obtain feature vectors corresponding to the current driving data and the driving environment data; The prediction submodule is used to input the feature vector into the parameter prediction model to predict the driving style, and obtain the current driving style output by the parameter prediction model and the initial values ​​of the driving parameters corresponding to the current driving style.

[0077] Optionally, the parameter adjustment module 303 includes: The identification submodule is used to identify the parameter adjustment request received from the user and determine the driving parameter to be adjusted and the parameter adjustment amount in the parameter adjustment request. The adjustment submodule is used to adjust the initial value of the driving parameter to be adjusted using the parameter adjustment amount, so as to obtain the adjusted driving parameter value.

[0078] Optionally, the parameter verification module 304 includes: The second receiving submodule is used to receive the parameter safety boundary sent by the cloud in advance; wherein, the parameter safety boundary is the safety threshold of each driving parameter determined by the cloud under different driving styles; The verification submodule is used to perform a safety verification on the adjusted driving parameter value using the parameter safety boundary, and to determine whether the adjusted driving parameter value exceeds the parameter safety boundary; The interception submodule is used to intercept the adjusted driving parameter value if it exceeds the parameter safety boundary, so as to obtain the target driving parameter value that satisfies the parameter safety boundary.

[0079] Optionally, the device further includes: The feedback acquisition module is used to acquire user feedback information on the target values ​​of the driving parameters; The model fine-tuning module is used to fine-tune the locally deployed parameter prediction model based on the feedback information to obtain the fine-tuned parameter prediction model. The data upload module is used to upload the current driving data, driving environment data, feedback information, and the fine-tuned parameter prediction model to the cloud. The cloud is used to iteratively update the parameter prediction model based on the current driving data, driving environment data, feedback information, and the fine-tuned parameter prediction model.

[0080] As the device embodiment is basically similar to the method embodiment, the description is relatively simple, and relevant parts can be found in the description of the method embodiment.

[0081] The vehicle control device provided in this embodiment of the invention acquires the vehicle's current driving data and driving environment data at the vehicle end. Using a parameter prediction model pre-deployed from the cloud, it predicts the driving style based on the current driving data and driving environment data, obtaining the current driving style and the initial values ​​of the corresponding driving parameters. The parameter prediction model is trained in the cloud based on the vehicle's driving data, driving environment data, and user behavior. In response to a user's parameter tuning request, the device adjusts the initial values ​​of the driving parameters according to the request, obtaining adjusted driving parameter values. The adjusted driving parameter values ​​are then verified against a pre-determined parameter safety boundary to obtain target values ​​of the driving parameters that satisfy the parameter safety boundary. The device then controls the vehicle to drive using the target values ​​of the driving parameters. This invention utilizes a cloud-trained parameter prediction model, combined with current driving data and environmental data, to predict the user's driving style and generate corresponding initial values ​​for driving parameters. Users can interact with the vehicle to output parameter adjustment requests to fine-tune the driving parameters, enabling flexible adjustment. Through safety verification and dynamic adjustment, driving parameter values ​​that meet the parameter safety boundaries are obtained, ensuring vehicle safety and driving according to user needs. This avoids loss of vehicle control or dangerous situations caused by improper parameter settings, improves the safety of driving parameter regulation, and meets the personalized driving needs of users.

[0082] Reference Figure 4 This diagram illustrates another vehicle control device provided by an embodiment of the present invention, applied in a cloud environment, wherein the cloud environment communicates with the vehicle. The device includes: The data acquisition module 401 is used to collect vehicle driving data, driving environment data, and user behavior; wherein, the user behavior includes user parameter adjustment requests and feedback information; Feature processing module 402 is used to perform feature fusion on the driving data, driving environment data and user behavior to obtain a fusion result; The model training module 403 is used to train a preset model based on the fusion result to obtain a trained parameter prediction model; wherein, the parameter prediction model is used to output the driving style and the initial values ​​of the driving parameters corresponding to the driving style. The model distribution module 404 is used to distribute the trained parameter prediction model to the vehicle terminal.

[0083] Optionally, the device further includes: The data receiving module is used to receive the current driving data, driving environment data, feedback information, and fine-tuned parameter prediction model uploaded by the vehicle terminal. The model update module is used to iteratively update the parameter prediction model based on the current driving data, driving environment data, feedback information, and the fine-tuned parameter prediction model.

[0084] As the device embodiment is basically similar to the method embodiment, the description is relatively simple, and relevant parts can be found in the description of the method embodiment.

[0085] The vehicle control device provided in this invention collects vehicle driving data, driving environment data, and user behavior in the cloud. User behavior includes user parameter adjustment requests and feedback information. The device performs feature fusion of the driving data, driving environment data, and user behavior to obtain a fusion result. Based on the fusion result, a preset model is trained to obtain a trained parameter prediction model. The parameter prediction model outputs driving style and corresponding initial values ​​of driving parameters. The trained parameter prediction model is then sent to the vehicle. This invention, through a cloud-trained parameter prediction model combined with current driving data and environmental data, predicts the user's driving style and generates corresponding initial values ​​of driving parameters. The user can interact with the vehicle to output parameter adjustment requests to fine-tune the driving parameters, enabling flexible adjustment. Through safety checks and dynamic adjustments, driving parameter values ​​that meet safety boundaries are obtained, ensuring vehicle safety and driving according to user needs. This avoids loss of control or dangerous situations caused by improper parameter settings, improves the safety of driving parameter control, and meets the personalized driving needs of users.

[0086] Reference Figure 5 This illustration shows a vehicle control system provided by an embodiment of the present invention. The vehicle control system includes a cloud and a vehicle terminal, the vehicle terminal communicating with the cloud, and the system includes: The vehicle-side system is used to acquire the vehicle's current driving data and driving environment data. Using a parameter prediction model pre-deployed from the cloud, it predicts the driving style based on the current driving data and the driving environment data to obtain the current driving style and the initial values ​​of the driving parameters corresponding to the current driving style. In response to receiving a user's parameter adjustment request, it adjusts the initial values ​​of the driving parameters according to the parameter adjustment request to obtain adjusted driving parameter values. The adjusted driving parameter values ​​are then subjected to safety verification through a pre-determined parameter safety boundary to obtain target values ​​of the driving parameters that satisfy the parameter safety boundary. Finally, the system controls the vehicle to drive using the target values ​​of the driving parameters. The cloud platform is used to collect vehicle driving data, driving environment data, and user behavior. The user behavior includes user parameter adjustment requests and feedback information. The driving data, driving environment data, and user behavior are fused to obtain a fusion result. Based on the fusion result, a preset model is trained to obtain a trained parameter prediction model. The parameter prediction model is used to output the driving style and the initial values ​​of the driving parameters corresponding to the driving style. The trained parameter prediction model is then sent to the vehicle.

[0087] As the system implementation is basically similar to the method implementation, it is described in a relatively simple way. For relevant details, please refer to the description of the method implementation.

[0088] The vehicle control system provided in this invention uses a parameter prediction model trained in the cloud, combined with current driving data and environmental data, to predict the user's driving style and generate corresponding initial values ​​for driving parameters. The user can interact with the vehicle to output parameter adjustment requests to fine-tune the driving parameters, enabling flexible adjustment of driving parameters. Through safety verification and dynamic adjustment, driving parameter values ​​that meet the parameter safety boundaries are obtained, ensuring vehicle safety and driving according to user needs. This avoids loss of vehicle control or dangerous situations caused by improper parameter settings, improves the safety of driving parameter regulation, and meets the personalized needs of users driving vehicles.

[0089] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented in software, it can be implemented, in whole or in part, as 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 the present invention are generated. 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 website, computer, server, or data center 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 accessible to a computer 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., DVD), or a semiconductor medium (e.g., a solid-state disk (SSD)).

[0090] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0091] The various embodiments in this specification are described in a related manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0092] The above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention are included within the scope of protection of the present invention.

Claims

1. A vehicle control method, characterized in that, Applied to a vehicle-side application, wherein the vehicle-side communicates with the cloud, the method includes: Acquire the vehicle's current driving data and driving environment data; Using the parameter prediction model pre-deployed from the cloud, driving style prediction is performed on the current driving data and the driving environment data to obtain the current driving style and the initial values ​​of the driving parameters corresponding to the current driving style; wherein, the parameter prediction model is trained by the cloud based on the vehicle's driving data, driving environment data, and user behavior; In response to receiving a user's parameter adjustment request, the initial value of the driving parameter is adjusted according to the parameter adjustment request to obtain the adjusted driving parameter value; The adjusted driving parameter values ​​are verified by using a predetermined parameter safety boundary to obtain target driving parameter values ​​that satisfy the parameter safety boundary. The vehicle is controlled to drive using the target values ​​of the driving parameters.

2. The method according to claim 1, characterized in that, The step of using the parameter prediction model pre-deployed from the cloud to predict driving style based on the current driving data and the driving environment data, and obtaining the current driving style and the initial values ​​of the driving parameters corresponding to the current driving style, includes: Receive the parameter prediction model sent from the cloud in advance, and deploy the parameter prediction model to the vehicle's local device; Feature extraction is performed on the current driving data and the driving environment data to obtain feature vectors corresponding to the current driving data and the driving environment data; The feature vector is input into the parameter prediction model to predict the driving style, and the current driving style and the initial values ​​of the driving parameters corresponding to the current driving style are obtained from the output of the parameter prediction model.

3. The method according to claim 1, characterized in that, The step of responding to a user's parameter adjustment request by adjusting the initial values ​​of the driving parameters according to the parameter adjustment request to obtain the adjusted driving parameters includes: In response to receiving a user's parameter adjustment request, the parameter adjustment request is identified, and the driving parameters to be adjusted and the parameter adjustment amount in the parameter adjustment request are determined; The initial value of the driving parameter to be adjusted is adjusted using the parameter adjustment amount to obtain the adjusted driving parameter value.

4. The method according to claim 1, characterized in that, The step of performing a safety check on the adjusted driving parameter values ​​using a pre-determined parameter safety boundary to obtain target driving parameter values ​​that satisfy the parameter safety boundary includes: The parameter safety boundary is received in advance from the cloud; wherein the parameter safety boundary is the safety threshold of each driving parameter determined by the cloud under different driving styles. The adjusted driving parameter values ​​are verified using the parameter safety boundary to determine whether the adjusted driving parameter values ​​exceed the parameter safety boundary. If the adjusted driving parameter value exceeds the parameter safety boundary, the adjusted driving parameter value is truncated to obtain the target driving parameter value that satisfies the parameter safety boundary.

5. The method according to claim 1 or 2, characterized in that, After the controlled vehicle is driven using the target values ​​of the driving parameters, the method further includes: Obtain user feedback on the target values ​​of the driving parameters; The locally deployed parameter prediction model is fine-tuned based on the feedback information to obtain the fine-tuned parameter prediction model. The current driving data, driving environment data, feedback information, and the fine-tuned parameter prediction model are uploaded to the cloud. The cloud is used to iteratively update the parameter prediction model based on the current driving data, driving environment data, feedback information, and the fine-tuned parameter prediction model.

6. A vehicle control method, characterized in that, The method is applied in the cloud, where the cloud communicates with the vehicle, and includes: The system collects vehicle driving data, driving environment data, and user behavior; wherein, the user behavior includes user parameter adjustment requests and feedback information. The driving data, driving environment data, and user behavior are fused to obtain a fusion result. The preset model is trained based on the fusion result to obtain a trained parameter prediction model; wherein, the parameter prediction model is used to output the driving style and the initial values ​​of the driving parameters corresponding to the driving style; The trained parameter prediction model is then sent to the vehicle.

7. The method according to claim 6, characterized in that, After the trained parameter prediction model is sent to the vehicle, the process further includes: Receives current driving data, driving environment data, feedback information, and fine-tuned parameter prediction model uploaded by the vehicle terminal; Based on the current driving data, driving environment data, feedback information, and the fine-tuned parameter prediction model, the parameter prediction model is iteratively updated.

8. A vehicle control device, characterized in that, Applied to a vehicle-mounted system that communicates with the cloud, the device includes: The data acquisition module is used to acquire the vehicle's current driving data and driving environment data; The parameter prediction module is used to predict the driving style of the current driving data and the driving environment data using a parameter prediction model pre-deployed from the cloud, and obtain the current driving style and the initial values ​​of the driving parameters corresponding to the current driving style; wherein, the parameter prediction model is trained by the cloud based on the vehicle's driving data, driving environment data and user behavior; The parameter adjustment module is used to respond to a user's parameter adjustment request, adjust the initial value of the driving parameter according to the parameter adjustment request, and obtain the adjusted driving parameter value. The parameter verification module is used to perform safety verification on the adjusted driving parameter values ​​through a pre-determined parameter safety boundary, so as to obtain the target driving parameter values ​​that meet the parameter safety boundary. The control module is used to control the vehicle to drive using the target values ​​of the driving parameters.

9. A vehicle control device, characterized in that, The device is applied in the cloud, whereby the cloud communicates with the vehicle, and includes: The data acquisition module is used to collect vehicle driving data, driving environment data, and user behavior; wherein, the user behavior includes user parameter adjustment requests and feedback information; The feature processing module is used to fuse the driving data, driving environment data, and user behavior features to obtain a fusion result; The model training module is used to train a preset model based on the fusion result to obtain a trained parameter prediction model; wherein, the parameter prediction model is used to output the driving style and the initial values ​​of the driving parameters corresponding to the driving style; The model distribution module is used to distribute the trained parameter prediction model to the vehicle terminal.

10. A vehicle control system, characterized in that, The vehicle control system includes a cloud platform and a vehicle terminal, wherein the vehicle terminal communicates with the cloud platform, and the system includes: The vehicle-side system is used to acquire the vehicle's current driving data and driving environment data. Using a parameter prediction model pre-deployed from the cloud, it predicts the driving style based on the current driving data and the driving environment data to obtain the current driving style and the initial values ​​of the driving parameters corresponding to the current driving style. In response to receiving a user's parameter adjustment request, it adjusts the initial values ​​of the driving parameters according to the parameter adjustment request to obtain adjusted driving parameter values. The adjusted driving parameter values ​​are then subjected to safety verification through a pre-determined parameter safety boundary to obtain target values ​​of the driving parameters that satisfy the parameter safety boundary. Finally, the system controls the vehicle to drive using the target values ​​of the driving parameters. The cloud platform is used to collect vehicle driving data, driving environment data, and user behavior. The user behavior includes user parameter adjustment requests and feedback information. The driving data, driving environment data, and user behavior are fused to obtain a fusion result. Based on the fusion result, a preset model is trained to obtain a trained parameter prediction model. The parameter prediction model is used to output the driving style and the initial values ​​of the driving parameters corresponding to the driving style. The trained parameter prediction model is then sent to the vehicle.

Images