Driving parameter switching method, model training method and device and mapping relation generating method and device

Abstract

The invention provides a driving parameter switching method, a model training method and device and a mapping relation generation method and device. The driving parameter switching method comprises the following steps: monitoring real-time working condition data when a vehicle runs; identifying a scene type of the real-time working condition data based on a pre-trained scene classification model; based on the scene type of the real-time working condition data, a switching operation corresponding to the real-time working condition data is searched from the mapping relation between the locally cached scene and the switching operation, or a query request containing the scene type of the real-time working condition data is sent to a cloud end so as to query the switching operation corresponding to the real-time working condition data; and switching operation corresponding to the searched or received real-time working condition data is executed. Accurate scene recognition is achieved, the driving parameters are automatically switched through the recognized scene, operation of a driver is simplified, and the driving experience is improved.

Description

Technical Field

[0001] This application relates to the field of vehicle control technology, and in particular to a method for switching driving parameters, a method and apparatus for model training and mapping relationship generation. Background Technology

[0002] With the rapid development of modern intelligent driving technology, the dynamic adjustment of driving parameters has become the key to improving driving experience and energy efficiency.

[0003] Currently, adjusting driving parameters mainly relies on manual adjustment by the driver. This requires the driver to analyze the current scenario based on their own experience and switch driving parameters, which is time-consuming and laborious. In some scenarios, the switching response is not timely enough, resulting in a poor driving experience. In complex road conditions, it may also cause safety risks due to distracted operation or delayed adjustment. Summary of the Invention

[0004] The driving parameter switching method, model training and mapping relationship generation method and apparatus provided in this application address the problems of cumbersome operation, delayed response and safety risks caused by manually adjusting driving parameters. They achieve accurate scene recognition through real-time operating conditions and automatically switch driving parameters using the recognized scene, thereby improving the timeliness and accuracy of driving parameter adjustment, reducing the driving burden, and optimizing the driving experience and energy efficiency.

[0005] In a first aspect, embodiments of this application provide a method for switching driving parameters, comprising: monitoring real-time operating condition data of a vehicle during operation; identifying the scene type of the real-time operating condition data based on a pre-trained scene classification model; based on the scene type of the real-time operating condition data, searching for the corresponding switching operation from a locally cached mapping relationship between scenes and switching operations, or sending a query request containing the scene type of the real-time operating condition data to the cloud to query the corresponding switching operation from the cloud-stored mapping relationship between scenes and switching operations; and executing the found or received switching operation corresponding to the real-time operating condition data; wherein the mapping relationship is generated based on the scene type and corresponding switching operation of the vehicle's historical operating condition data obtained from the time node when the driver switched driving parameters during the vehicle's historical operation.

[0006] In one possible implementation, both real-time operating condition data and historical operating condition data include environmental perception data and vehicle motion state parameters, which include at least one of driving speed, acceleration, and tilt angle.

[0007] In one possible implementation, the environmental perception data includes road image data and ambient light data.

[0008] In one possible implementation, based on a pre-trained scene classification model, the scene type of real-time operating data is identified, including: identifying the number of vehicles, road type, traffic signs and construction areas in road image data; extracting features from ambient light data and extracting image brightness features from road image data to obtain an environmental feature vector; extracting features from motion state parameters to obtain a motion feature vector; and concatenating the environmental feature vector and the motion feature vector and inputting them into the pre-trained scene classification model to obtain the scene type of the real-time operating data.

[0009] In one possible implementation, the method further includes: obtaining historical operating condition data of the vehicle based on the time nodes when the driver switched driving parameters during the vehicle's historical operation; uploading the historical operating condition data and the corresponding switching operations to the cloud, training a machine learning model based on the historical operating condition data and its corresponding scene labels to obtain a scene classification model, and generating and storing a mapping relationship between scenes and switching operations based on the scene labels and the corresponding switching operations of the historical operating condition data.

[0010] In one possible implementation, the method further includes: if a driver's manual switching operation is detected within a preset time after the switching operation corresponding to the real-time operating data is executed, then the mapping relationship is updated based on the manual switching operation and the scenario type of the real-time operating data.

[0011] In one possible implementation, the driving parameters include at least one of driving mode, recovery intensity, and suspension settings.

[0012] Secondly, embodiments of this application provide a model training and mapping relationship generation method, including: acquiring historical operating condition data of the vehicle during operation; training a machine learning model based on the historical operating condition data and its corresponding scene labels to obtain a scene classification model and sending it to the vehicle; generating and storing a mapping relationship between the scene and the switching operation based on the scene labels corresponding to the historical operating condition data and the switching operation corresponding to the historical operating condition data.

[0013] In one possible implementation, a mapping relationship between scenes and switching operations is generated and stored based on scene labels corresponding to historical operating condition data and switching operations corresponding to historical operating condition data. This includes: clustering the scene labels and switching operations corresponding to historical operating condition data using a clustering algorithm to obtain multiple clusters, each cluster containing multiple switching operations corresponding to the same scene label; for each cluster, determining the target switching operation associated with the scene label of the cluster based on the multiple switching operations in the cluster; and generating a mapping relationship based on the target switching operations associated with the scene labels of each cluster and storing it in a database.

[0014] Thirdly, embodiments of this application provide a driving parameter switching device, comprising: a working condition data monitoring module for monitoring real-time working condition data during vehicle operation; a scene recognition module for identifying the scene type of the real-time working condition data based on a pre-trained scene classification model; a switching operation determination module for searching for the corresponding switching operation based on the scene type of the real-time working condition data from a locally cached mapping relationship between scenes and switching operations, or sending a query request containing the scene type of the real-time working condition data to the cloud to query the corresponding switching operation; and a switching execution module for executing the switching operation corresponding to the found or received real-time working condition data; wherein the mapping relationship is generated based on the scene type and corresponding switching operation of the vehicle's historical working condition data obtained from the time node when the driver switched driving parameters during the vehicle's historical operation.

[0015] In one possible implementation, the scene recognition module is specifically used to: identify the number of vehicles, road type, traffic signs, and construction areas in road image data; extract features from ambient light data and extract image brightness features from road image data to obtain an environmental feature vector; extract features from motion state parameters to obtain a motion feature vector; and concatenate the environmental feature vector and the motion feature vector and input them into a pre-trained scene classification model to obtain the scene type of the real-time working condition data.

[0016] In one possible implementation, the device further includes a historical data uploading module, used to: acquire historical operating condition data of the vehicle based on the time nodes when the driver switched driving parameters during the vehicle's historical operation; upload the historical operating condition data and the corresponding switching operations to the cloud, so as to train a machine learning model based on the historical operating condition data and its corresponding scene labels to obtain a scene classification model, and generate and store a mapping relationship between scenes and switching operations based on the scene labels and the corresponding switching operations of the historical operating condition data.

[0017] In one possible implementation, the device further includes a mapping update module, which is used to: if a driver's manual switching operation is detected within a preset time after the switching operation corresponding to the real-time operating data is executed, update the mapping relationship based on the manual switching operation and the scenario type of the real-time operating data.

[0018] Fourthly, embodiments of this application provide a model training and mapping relationship generation apparatus, comprising: a historical data acquisition module for acquiring historical operating condition data of a vehicle during operation; a model training module for training a machine learning model based on the historical operating condition data and its corresponding scene labels to obtain a scene classification model and send it to the vehicle; and a mapping relationship generation module for generating and storing a mapping relationship between a scene and a switching operation based on the scene labels corresponding to the historical operating condition data and the switching operations corresponding to the historical operating condition data.

[0019] In one possible implementation, the mapping relationship generation module is specifically used for: clustering the scene labels and switching operations corresponding to historical working condition data based on a clustering algorithm to obtain multiple clusters, each cluster containing multiple switching operations corresponding to the same scene label; for each cluster, determining the target switching operation associated with the scene label of the cluster based on the multiple switching operations in the cluster; and generating a mapping relationship based on the target switching operations associated with the scene labels of each cluster and storing it in the database.

[0020] Fifthly, embodiments of this application provide an electronic device, including: a memory, a processor; the memory stores computer-executable instructions; the processor executes the computer-executable instructions stored in the memory, causing the processor to perform the method provided in the first or second aspect above, and / or, various possible implementations of the first or second aspect.

[0021] In a sixth aspect, embodiments of this application provide a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement the methods provided in the first or second aspect above, and / or various possible implementations of the first or second aspect.

[0022] In a seventh aspect, embodiments of this application provide a computer program product, including a computer program that, when executed by a processor, implements the methods provided in the first or second aspect above, and / or various possible implementations of the first or second aspect.

[0023] The driving parameter switching method, model training, and mapping relationship generation method and apparatus provided in this application pre-establish a mapping relationship between the scene and the switching operation by utilizing the historical operating condition data of the vehicle collected when the driver manually switches driving parameters, thereby enabling the driver to learn their operating habits. During vehicle operation, the scene is accurately identified by monitoring real-time operating condition data, and the driving parameters are automatically switched using the identified scene and the aforementioned mapping relationship. This improves the timeliness and accuracy of driving parameter adjustment, reduces the driving burden, and optimizes the driving experience and energy efficiency. Attached Figure Description

[0024] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0025] Figure 1 Flowchart of the driving parameter switching method provided in this application Figure 1 ;

[0026] Figure 2A schematic diagram illustrating the interaction process between the vehicle-mounted system and the cloud, provided in an embodiment of this application;

[0027] Figure 3 Flowchart of the driving parameter switching method provided in this application Figure 2 ;

[0028] Figure 4 A flowchart illustrating the model training and mapping relationship generation method provided in this application. Figure 1 ;

[0029] Figure 5 A schematic diagram of the driving parameter switching device provided in this application;

[0030] Figure 6 A schematic diagram of the structure of the electronic device provided in this application.

[0031] The accompanying drawings illustrate specific embodiments of this application, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the concept in any way, but rather to illustrate the concept of this application to those skilled in the art through reference to particular embodiments. Detailed Implementation

[0032] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. 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.

[0033] This application applies to adaptive adjustment systems for vehicle driving parameters in any driving scenario, including intelligent driving scenarios and ordinary driving scenarios.

[0034] In normal driving scenarios, drivers need to manually adjust driving parameters based on experience, such as recovery intensity, driving mode, and suspension mode, to adapt to real-time changes in road conditions.

[0035] In intelligent driving scenarios, drivers still need to adjust some driving parameters, such as recovery intensity, driving mode, and suspension mode, to compensate for the shortcomings of intelligent driving functions in personalized adaptation and adaptability to complex road conditions, and to achieve a balance between safety, comfort and energy efficiency in human-machine collaboration.

[0036] For example, when driving on a highway and encountering a long, gentle slope with the vehicle in front traveling at a relatively high speed, the driver often manually switches the driving mode from Eco to Sport and the suspension mode from Comfort (soft) to a firmer one. This ensures the vehicle can keep up with the speed of the vehicle in front on the slope and avoids speed fluctuations caused by frequent hard acceleration due to insufficient power. The firmer suspension also improves vehicle stability and prevents sensor misalignment caused by slight road bumps at high speeds, thus avoiding system misjudgments of road conditions.

[0037] For example, when driving on city roads during rush hour, where traffic is heavy and vehicle speeds fluctuate within a low range, drivers will manually adjust the energy recovery intensity from low to high and switch the driving mode from economy to comfort mode to better recover braking energy, improve range, and enhance passenger comfort.

[0038] In both intelligent driving and normal driving scenarios, manually adjusting driving parameters by the driver is cumbersome and results in delayed responses. To address this issue, this application provides a driving parameter switching method that automatically identifies scenarios based on real-time perception data. By utilizing a pre-established mapping relationship between scenarios and switching operations, it automatically switches to the driver's preferred operation in similar scenarios, reducing the driver's workload, improving the accuracy of parameter adaptation to real-time conditions and personalized driving habits, and ensuring the safety, comfort, and energy efficiency of intelligent driving.

[0039] The technical solution of this application and how the technical solution of this application solves the above-mentioned technical problems are described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of this application will now be described with reference to the accompanying drawings.

[0040] Figure 1 Flowchart of the driving parameter switching method provided in this application Figure 1 This method, when applied to vehicles, can be executed by the vehicle's infotainment system, such as by the vehicle's processor, control unit, or controller. Figure 1 As shown, the method includes:

[0041] Step S101: Monitor real-time operating data of the vehicle during operation.

[0042] Real-time operating data refers to data collected in real time during vehicle operation, such as during intelligent driving, that is related to the external environment and vehicle status, and may include data such as road images and driving speed.

[0043] Real-time operating data of the vehicle can be continuously collected at preset intervals using hardware devices such as sensors and controllers mounted on the vehicle.

[0044] For example, ADAS systems can use cameras to capture road images, use millimeter-wave radar to collect data such as distance to the vehicle in front, relative speed, and obstacles, and use onboard weather sensors to collect weather data such as temperature, humidity, and rainfall. They can also use the Body Control Module (BCM) to collect data such as speed and acceleration.

[0045] Real-time operating data collected can be transmitted in real time to the processor or module responsible for switching driving parameters via vehicle Ethernet or CAN bus.

[0046] Step S102: Based on the pre-trained scene classification model, identify the scene type of the real-time working condition data.

[0047] The scene classification model is a machine learning model used to identify the scene of the input real-time operating data and obtain the scene type of the real-time operating data.

[0048] First, feature extraction can be performed on the real-time operating condition data to obtain the operating condition feature vector. Then, the operating condition feature vector can be input into the scene classification model to determine the scene type of the real-time operating condition data.

[0049] For example, scene classification models can be decision tree models, support vector machines (SVM), convolutional neural network (CNN) models, etc.

[0050] Optionally, the operating condition data (including real-time operating condition data and subsequent historical operating condition data) includes environmental perception data and vehicle motion state parameters, which include at least one of driving speed, acceleration and tilt angle.

[0051] Environmental perception data refers to the data of the vehicle's external environment collected by the environmental perception module. This can include visual data, such as images and point cloud data, as well as data collected by sensors, such as ambient light data collected by an ambient light sensor and ambient temperature and humidity data collected by a temperature and humidity sensor.

[0052] Vehicle motion state data is used to characterize the vehicle's motion state, including speed, acceleration, and tilt angle. The tilt angle includes the climbing angle when the vehicle is driving on a slope, and may also include the turning angle of the vehicle body when turning, to identify slopes and curves.

[0053] In scene recognition, features can be extracted from environmental perception data and motion state data. For example, features such as road type, traffic signs, and weather type can be identified in the collected road images. Features of image brightness, ambient light data, and motion state parameters can be extracted. The extracted features are then input into a pre-trained scene classification model, which determines the matching scene type based on the input features.

[0054] By using environmental perception data and vehicle motion state parameters as feature inputs for scene recognition, the comprehensiveness and accuracy of scene recognition are improved.

[0055] Optionally, environmental perception data includes road image data and ambient light data.

[0056] Ambient light data is collected by the vehicle's ambient light sensor (ALS), while road image data is collected by the vehicle's camera.

[0057] An ambient light sensor is integrated into the top of the vehicle's windshield or the rearview mirror bracket. It can monitor data such as the intensity, uniformity, and color temperature of light outside the vehicle in real time. During scene recognition, it can determine the lighting conditions based on the ambient light data and distinguish between day and night.

[0058] Road image data includes images of the road ahead of the vehicle, which serve as the core visual basis for scene recognition. It is used to identify the type of road the vehicle is traveling on, nearby traffic signs, the density of vehicles ahead, weather type, construction signs, etc.

[0059] By integrating ambient light and image-based environmental perception data as the basis for scene recognition, the shortcomings of single visual data in complex lighting scenarios are effectively compensated for, thereby improving the discrimination accuracy and robustness of scene classification models.

[0060] Step S103: Based on the scenario type of the real-time operating data, find the corresponding switching operation from the mapping relationship between scenarios and switching operations cached locally, or send a query request containing the scenario type of the real-time operating data to the cloud to query the corresponding switching operation from the mapping relationship between scenarios and switching operations stored in the cloud.

[0061] The mapping relationship is generated based on the scenario type and corresponding switching operation of the vehicle's historical operating data obtained from the time node when the driver switched driving parameters during the vehicle's historical operation.

[0062] When a driver manually switches driving parameters, the system records the driver's switching operation and the time of the switch. Based on the vehicle's operating condition data collected during the time period corresponding to that time point (to distinguish it from the operating condition data used for matching in the future using this mapping relationship, the operating condition data here is recorded as historical operating condition data), the system determines the corresponding scenario type and updates the mapping relationship between the scenario type and the driver's manual switching operation. For example, it adds the scenario type and the driver's manual switching operation to the mapping relationship, or updates the switching operation that already matches the scenario type in the mapping relationship.

[0063] The complete mapping relationship between scenes and switching operations can be stored directly on the vehicle's local terminal or in the cloud.

[0064] When the mapping relationship is stored in the cloud, the vehicle terminal caches the switching operation corresponding to a scene type after each request, and continuously updates the local cached mapping relationship between scenes and switching operations.

[0065] For example, Figure 2 This is a schematic diagram illustrating the interaction process between the vehicle-mounted system and the cloud, as provided in the embodiments of this application. Figure 2 As shown, in the mapping relationship generation stage, the vehicle's infotainment system references historical operating condition data collected when the driver manually switches driving parameters in intelligent driving mode. This data is then sent to the cloud. Data is collected via cameras, vehicle sensors, microphones, etc., and the collected historical operating condition data is fed back to the cloud. The cloud performs scene and switching operation identification on the historical operating condition data. By clustering switching operations with scenes, a mapping relationship between scenes and switching operations is obtained and stored in the database. After the mapping relationship is generated, the online automatic switching stage begins. Real-time operating condition data of the vehicle during intelligent driving is collected through data collection. Scene type is identified through scene recognition. The cloud then searches for matching switching operations from the mapping relationship stored in the database and sends the found switching operations back to the vehicle's infotainment system. The vehicle's infotainment system executes the switching operations fed back from the cloud, enabling automatic switching of vehicle driving parameters, such as recovery intensity, driving mode, and suspension settings, in similar scenarios.

[0066] After generating the mapping relationship, the scenario type of the driving condition data can be obtained based on the driving condition data collected when the driver manually switches driving parameters. The mapping relationship stored in the database can then be updated based on the scenario type and switching operation of the driving condition data. The mapping relationship can be updated periodically based on the scenario type and switching operation of the driving condition data collected within the current period.

[0067] Updating the mapping relationship can replace the switching operation corresponding to a certain scene type in the mapping relationship with the latest collected switching operation corresponding to that scene type, or add a mapping for the switching operation corresponding to the tone of a newly added scene type.

[0068] Step S104: Perform the switching operation corresponding to the found or received real-time operating data.

[0069] The driving parameter switching operation is used to switch the vehicle's driving parameters. These parameters can include various parameters that can be manually adjusted during intelligent driving, or commonly used parameters, such as driving modes.

[0070] Optionally, driving parameters include at least one of driving mode, recovery intensity, and suspension settings.

[0071] Regeneration intensity, specifically energy recovery intensity, is a driving parameter for new energy vehicles. It refers to the efficiency level at which the vehicle switches from drive motor to engine mode during deceleration, coasting, or braking, converting the vehicle's kinetic energy into electrical energy and storing it in the power battery. Regeneration intensity can be categorized into low, medium, and high levels.

[0072] Suspension settings are core parameters of a vehicle's chassis system. They refer to the changes in a vehicle's shock absorption performance, body posture, and off-road capability achieved by adjusting the suspension's damping coefficient, spring stiffness, or chassis height. Suspension settings are typically divided into three damping modes: soft, standard, and stiff. Some vehicles also specify an off-road mode, which can actively raise the chassis.

[0073] Driving modes can include Eco, Comfort, Sport, Anti-Slip, and Off-Road modes. Different driving modes correspond to different control strategies for the vehicle's powertrain and steering systems, adapting to the diverse needs of intelligent driving scenarios.

[0074] In economy mode, the control strategy prioritizes reducing energy consumption by limiting the maximum output power of the engine or motor to improve energy recovery efficiency. At the same time, it optimizes the acceleration and deceleration slope to reduce rapid acceleration and deceleration, balancing range and cruising smoothness, and is suitable for constant speed driving scenarios.

[0075] In Comfort mode, the control strategy is to improve driving comfort, with smooth power output and reduced steering system damping to reduce vehicle jerking. It also works in conjunction with the soft suspension mode to cushion road bumps, making it suitable for scenarios such as urban congestion and daily commuting.

[0076] In Sport mode, the control strategy enhances power response and handling. The power output threshold of the engine or electric motor is increased, the throttle response speed is accelerated, and the steering system damping is increased, thereby improving the vehicle's stability when cornering. This is suitable for scenarios such as high-speed overtaking and mountain curves.

[0077] In anti-slip mode, the control strategy is to improve driving safety on low-traction surfaces by automatically increasing the following distance, reducing the sensitivity of braking intervention, and limiting torque output of the power system to prevent wheel slippage. Combined with the suspension hard mode, it reduces the risk of vehicle sideslip and is suitable for scenarios such as rain, snow, and icy roads.

[0078] In off-road mode, the control strategy is to improve the passability of complex road conditions. The suspension automatically raises the chassis height to avoid scraping the bottom of the vehicle, the power system switches to a high torque output mode, and some intelligent driving functions are automatically weakened to prioritize the vehicle's ability to get out of trouble. It is suitable for scenarios such as rural unpaved roads and gravel roads.

[0079] In addition to outputting the scene type of real-time operating data, the scene classification model can also output the probability that the real-time operating data belongs to that scene type, i.e., the confidence level.

[0080] When the confidence level of the scene type in the real-time operating condition data output by the scene classification model is greater than a preset threshold, such as 85%, then based on the scene type in the real-time operating condition data, the system first searches for a switching operation that matches the scene type from the locally cached mapping relationship between scenes and switching operations. If a match is found, the found switching operation is executed to achieve automatic switching of driving parameters. If no match is found, a query request containing the scene type is generated and sent to the cloud. The cloud then responds to the query request, searches for a switching operation that matches the scene type from the cloud-stored mapping relationship between scenes and switching operations, and returns the found switching operation to the vehicle's infotainment system, which then executes the received switching operation.

[0081] The driving parameter switching method provided in this application pre-establishes a mapping relationship between the scene and the switching operation by utilizing the historical operating condition data of the vehicle collected when the driver manually switches driving parameters, thereby enabling the driver to learn their operating habits. During vehicle operation, the method achieves accurate scene recognition by monitoring real-time operating condition data, and automatically switches driving parameters using the recognized scene and the aforementioned mapping relationship. This improves the timeliness and accuracy of driving parameter adjustment, reduces the driving burden, and optimizes the driving experience and energy efficiency.

[0082] Figure 3 Flowchart of the driving parameter switching method provided in this application Figure 2 In this embodiment Figure 2 Based on the embodiments, the driving parameter switching method is described in detail, such as... Figure 3 As shown, the method includes:

[0083] Step S301: Based on the time points when the driver switched driving parameters during the vehicle's historical operation, obtain the vehicle's historical operating condition data.

[0084] Historical operating condition data and real-time operating condition data have the same data content, only the collection time is different. They can include environmental perception data and vehicle motion status data. In addition to road image data and ambient light data, environmental perception data can also include audio data, which is data collected by the vehicle's speakers, such as the driver's voice.

[0085] Each time the driver manually switches the driving parameters of the intelligent driving vehicle, the vehicle's operating condition data for the corresponding time period is obtained. Once a certain amount of operating condition data is obtained, a step of generating or updating a mapping relationship is performed.

[0086] Assuming the driver manually switches the vehicle's driving parameters at time t0, the operating condition data for the time period [t0-t1, t0+t1] can be obtained as the historical operating condition data corresponding to this manual switching of driving parameters. Each manual switching operation of driving parameters by the driver and its corresponding historical operating condition data can be stored in the historical adjustment record. Once the data stored in the historical adjustment record reaches a preset quantity, this preset quantity of data is uploaded to the cloud, where the mapping relationship is updated based on the uploaded data. The cloud can also actively interact with the vehicle's infotainment system to obtain the latest historical adjustment records stored on the vehicle and update the mapping relationship stored in the database based on the latest historical adjustment records.

[0087] Step S302: Upload historical operating condition data and the corresponding switching operations to the cloud so that the cloud can train a machine learning model based on the historical operating condition data and its corresponding scene labels to obtain a scene classification model, and generate and store the mapping relationship between scenes and switching operations based on the scene labels and the corresponding switching operations of the historical operating condition data.

[0088] The switching operation refers to the operation corresponding to the vehicle's driving parameters. The switching operation corresponding to historical operating condition data is the switching operation that triggers the acquisition of that historical operating condition data.

[0089] The cloud can train machine learning models based on a large amount of historical operating data uploaded by multiple vehicles and their corresponding scene tags to obtain scene classification models for scene recognition.

[0090] For example, scene labels may include: downhill, traffic jam, fast driving, turning, driving in the rain, driving at night, intersection, construction area, etc.

[0091] During model training, historical work condition data can be preprocessed, including but not limited to outlier removal, missing value imputation, and data standardization. Then, feature engineering is used to extract discriminative features from the historical work condition data, such as driving speed, acceleration fluctuations, distance to the vehicle in front, traffic density, and slope, as well as image brightness features, and features corresponding to the number of vehicles, road type, traffic signs, and construction signs in the image. This yields training samples for each historical work condition data point. Scene labels for each training sample are obtained through manual annotation. The training samples are divided into training and validation sets. Model parameters are continuously optimized using the training set, and the model's performance is validated using the validation set. Once validated, the scene classification model is obtained.

[0092] Taking decision trees as an example, during model training, the root node is first constructed. Based on the training samples and their scene labels, the information gain ratio (IVR) of all candidate features is calculated. The candidate feature with the largest IVR is selected as the root node, for example, whether the driving speed is greater than or equal to 80 km / h. The training set is then divided into two branches through the root node. For each branch, based on the training samples and their scene labels under that branch, the IVR of the remaining candidate features is calculated. The feature with the largest IVR is selected as the node corresponding to the current branch, and that branch is divided into two sub-branches. This process continues until the branching process meets the training termination condition, at which point the recursion stops, and the current node is marked as a leaf node. The training termination condition can be that the scene labels of all training samples under the node are completely consistent, the number of remaining training samples under the node is less than a set threshold, or the IVR of all candidate features is less than a preset threshold.

[0093] To improve the model's generalization ability, cross-validation can be used for model training and parameter optimization, avoiding overfitting or underfitting caused by improper partitioning of the training and validation sets. For example, all labeled training samples collected from the cloud can be randomly divided into 5 non-overlapping subsets. Four subsets can be used alternately as the training set and one subset as the validation set, and the training can be repeated 5 times. The average accuracy of the 5 validations can then be used as the final performance metric of the model, ensuring that the model does not depend on a single data partition.

[0094] After successful verification, the scene classification model is lightweighted and deployed on the vehicle's infotainment system to identify the scene type of real-time operating data.

[0095] In addition to training the scene classification model in the cloud, it is also necessary to generate a mapping relationship between the scene and the switching operation corresponding to each driver based on the scene label of the driver's historical working condition data and the switching operation corresponding to the historical working condition data, and cache it in the database.

[0096] Step S303: Monitor real-time operating data of the vehicle during operation; wherein, real-time operating data includes road image data, ambient light data and vehicle motion state parameters.

[0097] Step S304: Identify the number of vehicles, road type, traffic signs, and construction areas in the road image data; extract the features of the ambient light data; and extract the image brightness features of the road image data to obtain the environmental feature vector.

[0098] Based on image processing algorithms such as object detection and semantic segmentation, the collected road images can be processed to identify the number of vehicles, road type, traffic signs, and construction areas.

[0099] Object detection algorithms can be used to identify all targets in an image, such as vehicles, traffic signs, and construction signs. The output includes the number of traffic signs and vehicles, the identified construction areas, and vehicle density. Semantic segmentation algorithms can be used to divide the image into road regions, and the road type can be determined by combining features such as lane number and road surface texture.

[0100] Ambient light data can be directly quantified to obtain its characteristics, such as light intensity. Image brightness features can include average image brightness value, brightness standard deviation, and the proportion of highlight areas.

[0101] The features obtained from image recognition, the features of ambient light data, and the features of image brightness are arranged in a fixed order to form a standardized vector, resulting in an environmental feature vector, such as [number of vehicles, road type, traffic signs, light intensity, average image brightness value, and percentage of highlight areas].

[0102] Step S305: Extract features from the motion state parameters to obtain motion feature vectors.

[0103] For the driving speed in the motion state parameters, features such as the current driving speed, the driving speed range, and vehicle speed stability can be extracted.

[0104] For acceleration in motion state parameters, features such as current acceleration value, acceleration switching frequency, and average acceleration can be extracted.

[0105] The tilt angle can include the climbing angle and the turning angle, and the extracted features can include the current value, range, and magnitude of change of the climbing angle or the turning angle.

[0106] By arranging the extracted features in a certain order, we can obtain the motion feature vector.

[0107] Step S306: After concatenating the environmental feature vector and the motion feature vector, input them into the pre-trained scene classification model to obtain the scene type of the real-time working condition data.

[0108] The environmental feature vector and motion feature vector are concatenated to obtain scene features. These scene features are then input into a scene classification model to achieve scene recognition and obtain the scene type of real-time operating data.

[0109] Step S307: Based on the scenario type of the real-time operating data, search for the corresponding switching operation in the local cached mapping relationship between scenarios and switching operations. If found, proceed to step S308; otherwise, proceed to step S309.

[0110] Step S308: Execute the switching operation corresponding to the found real-time operating data.

[0111] Step S309: Send a query request for the scenario type containing real-time operating data to the cloud, so as to query the switching operation corresponding to the real-time operating data from the mapping relationship between scenarios and switching operations stored in the cloud.

[0112] Step S310: Perform the switching operation corresponding to the received real-time operating data.

[0113] After determining the scenario type of the real-time operating data, the system searches the local cache of the scenario-to-switching operation mapping relationship on the vehicle's infotainment system to see if there is a matching switching operation. If so, the switching operation is executed. If not, a query request carrying the scenario type of the real-time operating data is generated and sent to the cloud. The cloud then searches its stored mapping relationship for a matching switching operation and sends the found switching operation back to the vehicle's infotainment system. The vehicle's infotainment system executes the received switching operation and caches the mapping relationship between the scenario type of the real-time operating data and the switching operation.

[0114] Furthermore, if there is no switching operation in the cloud storage mapping relationship that matches the scene type of the real-time operating condition data, then there is no need to switch the driving parameters. Instead, wait for the real-time operating condition data collected in the next cycle to obtain the scene type of the real-time operating condition data in the next cycle through scene recognition, and then find the switching operation that matches the scene type, and so on.

[0115] In this embodiment, a scene classification model is trained using a large amount of operating condition data collected during time periods corresponding to manual switching of driving parameters by drivers, improving the accuracy and fit of scene recognition. Simultaneously, using the driver as a dimension, scene tags and corresponding switching operations from the driver's historical operating condition data are used to generate a mapping relationship between scenes and switching operations for that driver. This enables the learning of driver driving habits and provides accurate and personalized decision-making basis for automatic parameter switching in similar scenarios. Features extracted from multi-dimensional data such as images, ambient light, and vehicle cloud parameters are used to achieve online scene recognition of real-time operating condition data using the scene classification model, improving the accuracy and comprehensiveness of scene recognition. Based on the identified scene type, switching operations are first matched from the locally cached mapping relationship. If a match is found, the matched switching operation is executed directly, resulting in fast response and low overhead. If no match is found locally, a matching switching operation is searched through interaction with the cloud, improving the comprehensiveness of scene matching and increasing the coverage of automatic switching operation execution.

[0116] Furthermore, the method also includes: if a driver's manual switching operation is detected within a preset time period corresponding to the switching operation of the real-time operating data, then the mapping relationship is updated based on the manual switching operation and the scenario type of the real-time operating data.

[0117] The preset time can be a short time, such as 1 second, 3 seconds, etc.

[0118] If, after matching the scenario type corresponding to the real-time operating data according to the aforementioned method and executing the switching operation matching that scenario type, a manual switching operation by the driver is detected within a short period of time, indicating that the driver is not satisfied with the currently automatically executed switching operation, then the scenario type and the manual switching operation are uploaded to the cloud. This allows the cloud to use the driver's manual switching operation and the scenario type of the real-time operating data as a new sample to update the mapping relationship between scenarios and switching operations. Alternatively, the cloud can directly update the mapping relationship stored locally based on the scenario type and the manual switching operation, and simultaneously update the mapping relationship stored in the cloud.

[0119] The mapping relationship can be updated by replacing the switching operation corresponding to the scene type in the mapping relationship with the manual switching operation, or updating the switching operation corresponding to the scene type to the weighted result of the original switching operation and the manual switching operation, or updating the switching operation matching the scene type in the mapping relationship based on the preset number of manual switching operations when the number of manual switching operations by the driver reaches a preset number under the same scene type.

[0120] By detecting the driver's manual switching operation shortly after the automatic switching operation is executed, a feedback loop is achieved. The mapping relationship is updated using the driver's feedback, which improves the accuracy of the mapping relationship and the driver's driving habits. It can better capture the driver's real-time preferences, making the automatic switching operation more in line with the driver's actual needs, reducing the frequency of driver manual operation, and improving the driving experience.

[0121] Figure 4 A flowchart illustrating the model training and mapping relationship generation method provided in this application. Figure 1 The method can be executed by a server deployed in the cloud, such as... Figure 4 As shown, the model training and mapping relationship generation method specifically includes the following steps:

[0122] Step S401: Obtain historical operating condition data of the vehicle during operation.

[0123] Among them, the historical operating condition data is the vehicle's operating condition data collected when the driver manually switches driving parameters, and is obtained based on the time points when the driver switches driving parameters during the vehicle's historical operation.

[0124] Step S402: Train a machine learning model based on historical operating condition data and its corresponding scene labels to obtain a scene classification model and send it to the vehicle.

[0125] Using a large amount of historical operating condition data uploaded by vehicles as raw data, feature engineering was used to obtain a large number of training samples, and scene labels were obtained for the training samples through manual annotation. The scene features of the training samples were used as input, and the scene type was used as output to train a machine learning model. During training, cross-validation was used to optimize the model parameters, ensuring the model's generalization ability under different road conditions and vehicles, ultimately resulting in a scene classification model that can output corresponding scene types based on real-time operating condition data.

[0126] After the model training is completed and passed the verification, the verified scene classification model will be distributed to the vehicle's in-vehicle terminal via OTA (Over-the-Air) or local transmission to ensure that the in-vehicle terminal has the ability to recognize local scenes.

[0127] Step S403: Based on the scene labels corresponding to the historical working condition data and the switching operations corresponding to the historical working condition data, generate and store the mapping relationship between the scene and the switching operation.

[0128] The scene labels corresponding to the historical operating condition data are the scene types of the historical operating condition data that were manually labeled.

[0129] Based on the driver's perspective, the scene tags and switching operations corresponding to the historical working data of the same driver are organized into standardized mapping relationships, forming a structured mapping relationship library. These mapping relationships are stored in the cloud or distributed to the corresponding vehicles. The cloud stores the full set of mapping relationships, while the vehicle's infotainment system can cache mapping relationships for high-frequency scenarios for rapid local matching, improving response speed.

[0130] Optionally, based on the scene labels corresponding to historical operating condition data and the switching operations corresponding to historical operating condition data, a mapping relationship between scenes and switching operations is generated and stored, including: clustering the scene labels and switching operations corresponding to historical operating condition data based on a clustering algorithm to obtain multiple clusters, each cluster containing multiple switching operations corresponding to the same scene label; for each cluster, determining the target switching operation associated with the scene label of the cluster based on the multiple switching operations in the cluster; and generating a mapping relationship based on the target switching operations associated with the scene labels of each cluster and storing it in the database.

[0131] In historical operating data, the same or similar scene labels or scene types may correspond to various switching operations of the driver. Clustering algorithms can automatically determine the commonalities of these switching operations, group similar operations into a cluster, and then select the most representative target switching operation from the cluster as the switching operation that matches the scene type, thereby reducing the interference of the driver's occasional switching operations on the mapping relationship.

[0132] By using clustering algorithms, we were able to extract the commonalities of switching operations in similar scenarios, effectively filter out occasional operations, improve the reliability of mapping relationships, and reduce storage pressure.

[0133] To reduce the amount of mapping data stored in the cloud, multi-dimensional data clustering algorithms, such as K-Means and DBSCAN (Density-Based Spatial Clustering of Applications with Noise), can be used as inputs to perform joint clustering, identify user groups with similar switching operations in similar scenarios, and obtain multiple clusters. Each cluster corresponds to a set of mapping relationships between scenarios and switching operations.

[0134] For each cluster, the switching operations of all drivers within the cluster under various scenario types are extracted. The operations with the highest frequency and similarity under each scenario type are selected as target switching operations. A mapping relationship between the target switching operation and the scenario type is established to obtain the general mapping relationship for the cluster, which is then stored. This general mapping relationship can also be distributed to the vehicle-mounted terminals corresponding to each driver within the cluster, allowing scenario type matching to determine automatically executed switching operations.

[0135] Figure 5 A schematic diagram of the driving parameter switching device provided in this application is shown below. Figure 5 As shown, the driving parameter switching device provided in this embodiment includes: a working condition data monitoring module for monitoring real-time working condition data during vehicle operation; a scene recognition module for identifying the scene type of the real-time working condition data based on a pre-trained scene classification model; a switching operation determination module for finding the corresponding switching operation based on the scene type of the real-time working condition data from the locally cached mapping relationship between scenes and switching operations, or sending a query request containing the scene type of the real-time working condition data to the cloud to query the corresponding switching operation; and a switching execution module for executing the switching operation corresponding to the found or received real-time working condition data; wherein, the mapping relationship is generated based on the scene type and corresponding switching operation of the vehicle's historical working condition data obtained from the time node when the driver switched driving parameters during the vehicle's historical operation.

[0136] In one possible implementation, the scene recognition module is specifically used to: identify the number of vehicles, road type, traffic signs, and construction areas in road image data; extract features from ambient light data and extract image brightness features from road image data to obtain an environmental feature vector; extract features from motion state parameters to obtain a motion feature vector; and concatenate the environmental feature vector and the motion feature vector and input them into a pre-trained scene classification model to obtain the scene type of the real-time working condition data.

[0137] In one possible implementation, the device further includes a historical data uploading module, used to: acquire historical operating condition data of the vehicle based on the time nodes when the driver switched driving parameters during the vehicle's historical operation; upload the historical operating condition data and the corresponding switching operations to the cloud, so as to train a machine learning model based on the historical operating condition data and its corresponding scene labels to obtain a scene classification model, and generate and store a mapping relationship between scenes and switching operations based on the scene labels and the corresponding switching operations of the historical operating condition data.

[0138] In one possible implementation, the device further includes a mapping update module, which is used to: if a driver's manual switching operation is detected within a preset time after the switching operation corresponding to the real-time operating data is executed, update the mapping relationship based on the manual switching operation and the scenario type of the real-time operating data.

[0139] The driving parameter switching device provided in this embodiment can execute the driving parameter switching method provided in the above method embodiment. Its implementation principle and technical effect are similar, and will not be described in detail here.

[0140] This application embodiment also provides a model training and mapping relationship generation device, including: a historical data acquisition module, used to acquire historical operating condition data of the vehicle during operation; a model training module, used to train a machine learning model based on the historical operating condition data and its corresponding scene labels, obtain a scene classification model and send it to the vehicle; and a mapping relationship generation module, used to generate and store a mapping relationship between the scene and the switching operation based on the scene labels corresponding to the historical operating condition data and the switching operation corresponding to the historical operating condition data.

[0141] In one possible implementation, the mapping relationship generation module is specifically used for: clustering the scene labels and switching operations corresponding to historical working condition data based on a clustering algorithm to obtain multiple clusters, each cluster containing multiple switching operations corresponding to the same scene label; for each cluster, determining the target switching operation associated with the scene label of the cluster based on the multiple switching operations in the cluster; and generating a mapping relationship based on the target switching operations associated with the scene labels of each cluster and storing it in the database.

[0142] Figure 6 A schematic diagram of the structure of the electronic device provided in this application. Figure 6 As shown, the control device 60 provided in this embodiment includes at least one processor 601 and a memory 602. Optionally, the control device 60 further includes a communication component 603. The processor 601, memory 602, and communication component 603 are connected via a bus 604.

[0143] In a specific implementation, at least one processor 601 executes computer execution instructions stored in memory 602, causing at least one processor 601 to perform the above-described method.

[0144] The specific implementation process of processor 601 can be found in the above method embodiments, and its implementation principle and technical effect are similar. It will not be repeated here.

[0145] In the above embodiments, it should be understood that the processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), etc. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in this invention can be directly implemented by a hardware processor, or implemented by a combination of hardware and software modules within the processor.

[0146] The memory may include random access memory (RAM) and may also include non-volatile memory (NVM), such as at least one disk storage device.

[0147] The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of illustration, the buses shown in the accompanying drawings are not limited to a single bus or a single type of bus.

[0148] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described method.

[0149] This application also provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the above-described method.

[0150] The aforementioned readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The readable storage medium can be any available medium accessible to a general-purpose or special-purpose computer.

[0151] An exemplary readable storage medium is coupled to a processor, enabling the processor to read information from and write information to the readable storage medium. Of course, the readable storage medium can also be a component of the processor. The processor and the readable storage medium can reside in an Application Specific Integrated Circuit (ASIC). Alternatively, the processor and the readable storage medium can exist as discrete components in the device.

[0152] The division of units is merely a logical functional division; in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces, devices, or units, and may be electrical, mechanical, or other forms.

[0153] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0154] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0155] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0156] Those skilled in the art will understand that all or part of the steps of the above-described method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments; and the aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0157] Finally, it should be noted that other embodiments of the invention will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This invention is intended to cover any variations, uses, or adaptations of the invention that follow the general principles of the invention and include common knowledge or customary techniques in the art not disclosed herein, and is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of the invention is limited only by the appended claims.

Claims

1. A method for switching driving parameters, characterized in that, include: Monitor real-time operating data of the vehicle during operation; Based on a pre-trained scene classification model, the scene type of the real-time operating condition data is identified; Based on the scenario type of the real-time operating data, the corresponding switching operation can be found from the mapping relationship between scenarios and switching operations cached locally, or a query request containing the scenario type of the real-time operating data can be sent to the cloud to query the corresponding switching operation from the mapping relationship between scenarios and switching operations stored in the cloud. Execute the switching operation corresponding to the found or received real-time operating data; The mapping relationship is generated based on the scenario type and corresponding switching operation of the vehicle's historical operating condition data obtained from the time node when the driver switched driving parameters during the vehicle's historical operation.

2. The method according to claim 1, characterized in that, Both the real-time operating data and the historical operating data include environmental perception data and the vehicle's motion state parameters, which include at least one of driving speed, acceleration, and tilt angle.

3. The method according to claim 2, characterized in that, The environmental perception data includes road image data and ambient light data.

4. The method according to claim 3, characterized in that, The pre-trained scene classification model identifies the scene type of the real-time operating data, including: The system identifies the number of vehicles, road type, traffic signs, and construction areas in the road image data, extracts features from the ambient light data, and extracts image brightness features from the road image data to obtain an environmental feature vector. The motion state parameters are subjected to feature extraction to obtain a motion feature vector; The environmental feature vector and the motion feature vector are concatenated and then input into a pre-trained scene classification model to obtain the scene type of the real-time working condition data.

5. The method according to claim 1, characterized in that, The method further includes: Based on the time points when the driver switched driving parameters during the vehicle's historical operation, the vehicle's historical operating condition data is obtained. The historical operating condition data and the corresponding switching operations are uploaded to the cloud. A machine learning model is trained based on the historical operating condition data and its corresponding scene labels to obtain the scene classification model. A mapping relationship between scenes and switching operations is generated and stored based on the scene labels and the corresponding switching operations of the historical operating condition data.

6. The method according to any one of claims 1-5, characterized in that, The method further includes: If a driver's manual switching operation is detected within a preset time after the switching operation corresponding to the real-time operating data is executed, the mapping relationship is updated based on the manual switching operation and the scenario type of the real-time operating data.

7. The method according to any one of claims 1-5, characterized in that, The driving parameters include at least one of driving mode, recovery intensity, and suspension settings.

8. A method for model training and mapping relationship generation, characterized in that, include: Acquire historical operating condition data of the vehicle during operation; A machine learning model is trained based on the historical operating condition data and its corresponding scene labels to obtain the scene classification model and then sent to the vehicle. Based on the scene tags corresponding to the historical operating condition data and the switching operations corresponding to the historical operating condition data, a mapping relationship between scenes and switching operations is generated and stored.

9. The method according to claim 8, characterized in that, The step of generating and storing a mapping relationship between scenes and switching operations based on the scene tags corresponding to the historical operating condition data and the switching operations corresponding to the historical operating condition data includes: Based on the clustering algorithm, the scene labels and switching operations corresponding to the historical working condition data are clustered to obtain multiple clusters, and each cluster contains multiple switching operations corresponding to the same scene label. For each cluster, based on multiple switching operations in the cluster, a target switching operation associated with the scene label of the cluster is determined; Based on the target switching operation associated with the scene labels of each cluster, the mapping relationship is generated and stored in the database.

10. A driving parameter switching device, characterized in that, include: The operating condition data monitoring module is used to monitor real-time operating condition data of the vehicle during operation; The scene recognition module is used to identify the scene type of the real-time operating condition data based on a pre-trained scene classification model. The switching operation determination module is used to find the switching operation corresponding to the real-time operating conditions data based on the scenario type of the real-time operating conditions data, from the mapping relationship between scenarios and switching operations cached locally, or to send a query request containing the scenario type of the real-time operating conditions data to the cloud to query the switching operation corresponding to the real-time operating conditions data. The switching execution module is used to perform the switching operation corresponding to the found or received real-time operating condition data; The mapping relationship is generated based on the scenario type and corresponding switching operation of the vehicle's historical operating condition data obtained from the time node when the driver switched driving parameters during the vehicle's historical operation.

11. A model training and mapping relationship generation device, characterized in that, include: The historical data acquisition module is used to acquire historical operating condition data of the vehicle during operation; The model training module is used to train a machine learning model based on the historical working condition data and its corresponding scene labels, obtain the scene classification model, and send it to the vehicle. The mapping relationship generation module is used to generate and store the mapping relationship between the scene and the switching operation based on the scene label corresponding to the historical working condition data and the switching operation corresponding to the historical working condition data.

12. An electronic device, characterized in that, include: Memory, processor; The memory stores computer-executed instructions; The processor executes computer execution instructions stored in the memory, causing the processor to perform the method as described in any one of claims 1-9.

13. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, implement the method described in any one of claims 1-9.

14. A computer program product, characterized in that, Includes a computer program that, when executed by a processor, implements the method according to any one of claims 1-9.

Images