Vehicle control method, vehicle and computer readable storage medium

Abstract

The embodiment of the invention provides a vehicle control method, a vehicle and a computer readable storage medium, and relates to the technical field of vehicle control. The method comprises the following steps: acquiring vehicle speed information and environment information of a target vehicle; based on the vehicle speed information and the environment information, a state feature set is constructed, and the state feature set is used for representing multi-dimensional features associated with the oxygen generation demand of the target vehicle; the state feature set is input into an oxygen generation mode prediction model, a target oxygen generation mode is obtained, and the oxygen generation mode prediction model is a classification model obtained based on supervised learning pre-training; and controlling the target vehicle to enter the target oxygen generation mode. The technical problem that the safety and comfort of a driver and passengers are affected due to the fact that a cabin oxygen generation strategy is rigid in the prior art is solved.

Description

Technical Field

[0001] This application relates to the field of vehicle control technology, and more specifically, to a vehicle control method, a vehicle, and a computer-readable storage medium. Background Technology

[0002] In-vehicle air quality management is an increasingly important aspect of modern vehicle design, especially given the trend towards higher quality of life and better health. Oxygen generation control, as a key technology for improving the cabin environment, has a direct impact on the safety, comfort, and even health of passengers.

[0003] However, existing vehicle cabin oxygen generation technologies generally suffer from overly rigid control strategies and difficulty adapting to changing environments. Most current vehicle cabin oxygen generation systems employ control logic based on fixed thresholds; for example, when the oxygen concentration inside the vehicle is detected to be below a certain set value, the oxygen generator is activated to supplement oxygen. While this control method is simple and intuitive, it lacks flexibility and struggles to effectively respond to changes in vehicle operating conditions and the environment.

[0004] There is currently no good solution to the above problems. Summary of the Invention

[0005] This application provides a vehicle control method, a vehicle, and a computer-readable storage medium to at least solve the technical problem in the related art of rigid cabin oxygen generation strategies, which in turn affect the safety and comfort of drivers and passengers.

[0006] According to one aspect of the embodiments of this application, a vehicle control method is provided, comprising: acquiring vehicle speed information and environmental information of a target vehicle; constructing a state feature set based on the vehicle speed information and environmental information, wherein the state feature set is used to characterize multidimensional features associated with the oxygen production demand of the target vehicle; inputting the state feature set into an oxygen production mode prediction model to obtain a target oxygen production mode, wherein the oxygen production mode prediction model is a classification model obtained based on supervised learning pre-training; and controlling the target vehicle to enter the target oxygen production mode.

[0007] Furthermore, the environmental information includes: oxygen concentration information, vital signs information of the in-vehicle population, temperature information, air quality information, and air humidity information. Based on the vehicle speed information and environmental information, a set of state features is constructed, including: normalizing the vehicle speed information, oxygen concentration information, vital signs information of the in-vehicle population, temperature information, air quality information, and air humidity information to obtain a first processing result; and constructing a set of state features based on the first processing result.

[0008] Furthermore, controlling the target vehicle to enter the target oxygen production mode includes: querying a preset control parameter mapping table to determine the control parameters of the vehicle execution unit corresponding to the target oxygen production mode, wherein the vehicle execution unit is used to characterize the unit in the target vehicle used to realize the oxygen production function; and controlling the vehicle execution unit to operate according to the control parameters.

[0009] Furthermore, the vehicle control method also includes: acquiring historical vehicle speed information and historical environmental information of the target vehicle; performing a rationality check on the historical vehicle speed information and historical environmental information according to preset check rules to obtain a check result, wherein the check result is used to characterize whether the historical vehicle speed information and historical environmental information meet the preset check rules; in response to the check result indicating that the historical vehicle speed information and historical environmental information meet the preset check rules, performing normalization processing on the historical vehicle speed information and historical environmental information to obtain a second processing result; constructing multiple historical state feature sets based on the second processing result; labeling the multiple historical state feature sets according to preset labeling rules to obtain a pre-training dataset; and pre-training the initial classification model based on the pre-training dataset to obtain an oxygen production mode prediction model.

[0010] Furthermore, according to preset annotation rules, multiple historical state feature sets are annotated, including: determining the oxygen generation mode corresponding to each historical state feature set in the multiple historical state feature sets according to preset annotation rules, wherein the oxygen generation mode includes a first oxygen generation mode, a second oxygen generation mode, and a third oxygen generation mode, the oxygen generation capacity of the first oxygen generation mode is greater than that of the second oxygen generation mode, and the oxygen generation capacity of the second oxygen generation mode is greater than that of the third oxygen generation mode; and annotating each historical state feature set according to the oxygen generation mode corresponding to each historical state feature set.

[0011] Furthermore, the initial classification model is a random forest model. Based on the pre-training dataset, the initial classification model is pre-trained to obtain an oxygen production mode prediction model, including: dividing the pre-training dataset into a training dataset and a test dataset according to a preset partitioning rule; and using the training dataset to pre-train the random forest model to obtain an oxygen production mode prediction model.

[0012] Furthermore, the random forest model is pre-trained using the training dataset to obtain an oxygen production mode prediction model, including: initializing the forest structure of the random forest model to obtain an initial forest structure, wherein the initial forest structure includes multiple initial decision trees; training the multiple initial decision trees in parallel based on the training dataset to generate multiple target decision trees; and generating an oxygen production mode prediction model based on the multiple target decision trees.

[0013] Furthermore, the vehicle control method also includes: calculating the classification accuracy of the oxygen production mode prediction model on the test dataset; in response to the classification accuracy being greater than or equal to a preset accuracy threshold, calculating the mode index value of the oxygen production mode prediction model on the test dataset; in response to the mode index value being greater than or equal to a preset index threshold, determining that the oxygen production mode prediction model meets the prediction requirements, wherein the mode index value is used to measure the ability of the oxygen production mode prediction model to identify multiple oxygen production modes; and in response to the classification accuracy being less than a preset accuracy threshold, retraining the oxygen production mode prediction model.

[0014] According to another aspect of the embodiments of this application, a vehicle control system is also provided, including: an acquisition module for acquiring vehicle speed information and environmental information of a target vehicle; a construction module for constructing a state feature set based on the vehicle speed information and environmental information, wherein the state feature set is used to characterize multi-dimensional features associated with the oxygen production demand of the target vehicle; a prediction module for inputting the state feature set into an oxygen production mode prediction model to obtain a target oxygen production mode, wherein the oxygen production mode prediction model is a classification model obtained based on supervised learning pre-training; and a control module for controlling the target vehicle to enter the target oxygen production mode.

[0015] According to another aspect of the embodiments of this application, a vehicle is also provided, including: a memory storing an executable program; and a processor for running the program, wherein the program executes the methods in various embodiments of this application when it runs.

[0016] According to another aspect of the embodiments of this application, a computer-readable storage medium is also provided, the computer-readable storage medium including a stored executable program, wherein, when the executable program is running, it controls the device where the computer-readable storage medium is located to perform the methods of various embodiments of this application.

[0017] In this embodiment, the vehicle speed and environmental information of the target vehicle are acquired; based on the vehicle speed and environmental information, a state feature set is constructed, wherein the state feature set is used to characterize the multidimensional features associated with the oxygen production demand of the target vehicle; the state feature set is input into an oxygen production mode prediction model to obtain the target oxygen production mode, wherein the oxygen production mode prediction model is a classification model obtained based on supervised learning pre-training; and the target vehicle is controlled to enter the target oxygen production mode. This application firstly acquires the vehicle speed and environmental information of the target vehicle, enabling a comprehensive perception of the dynamic environment and vehicle speed inside and outside the vehicle cabin, thereby avoiding decision-making biases caused by incomplete or delayed information in traditional control strategies, ensuring that oxygen production control matches the current environmental conditions. Secondly, based on the acquired vehicle speed and environmental information, a state feature set is constructed, and these multidimensional features are parameterized, effectively characterizing the complex states related to the vehicle's oxygen production demand. The construction of this state feature set provides an accurate data foundation for subsequent model prediction, avoiding the one-sidedness of decision-making caused by controlling a single or few parameters, improving the refinement and comprehensiveness of the oxygen production strategy, and thus improving the safety and comfort experience of drivers and passengers. Secondly, the set of state features is input into an oxygen production mode prediction model pre-trained based on supervised learning. Through the model's nonlinear mapping capability, it can output the target oxygen production mode that best matches the current environmental and vehicle speed information, achieving dynamic matching between oxygen production demand, environmental conditions, and vehicle speed. Finally, based on the predicted target oxygen production mode, the target vehicle is controlled to enter the corresponding oxygen production mode. This closed-loop control strategy ensures that the vehicle can dynamically adjust the control parameters related to oxygen production according to real-time environmental demands and vehicle speed changes, achieving precise satisfaction of oxygen production needs. Therefore, this application achieves the technical effect of dynamically adjusting the cabin oxygen production strategy to ensure the safety and comfort of passengers, thereby solving the technical problem of rigid cabin oxygen production strategies in related technologies, which affects the safety and comfort of passengers. Attached Figure Description

[0018] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0019] Figure 1 This is a flowchart of a vehicle control method according to an embodiment of this application;

[0020] Figure 2 This is a flowchart of another vehicle control method according to an embodiment of this application;

[0021] Figure 3 This is a structural block diagram of a vehicle control system according to an embodiment of this application. Detailed Implementation

[0022] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0023] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0024] According to an embodiment of this application, a method embodiment for vehicle control is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0025] This embodiment provides a vehicle control method. Figure 1 This is a flowchart of a vehicle control method according to an embodiment of this application, such as... Figure 1 As shown, the method includes the following steps:

[0026] Step S10: Obtain the target vehicle's speed information and environmental information;

[0027] Step S11: Based on vehicle speed information and environmental information, construct a set of state features, wherein the set of state features is used to characterize the multidimensional features associated with the oxygen production requirements of the target vehicle.

[0028] Step S12: Input the set of state features into the oxygen production mode prediction model to obtain the target oxygen production mode. The oxygen production mode prediction model is a classification model obtained based on supervised learning pre-training.

[0029] Step S13: Control the target vehicle to enter the target oxygen generation mode.

[0030] The speed information mentioned above refers to the current speed of the target vehicle.

[0031] In one optional embodiment, vehicle speed information is collected by a vehicle speed sensor to determine the vehicle's driving status, such as stationary, low speed, or high speed, thereby affecting the oxygen concentrator's operating strategy.

[0032] The aforementioned environmental information includes oxygen concentration, temperature, air quality, and air humidity.

[0033] In one alternative embodiment, environmental information is acquired by environmental sensing sensors (such as oxygen sensors, temperature sensors, humidity sensors, and air quality monitoring sensors) integrated on the target vehicle to comprehensively assess the environmental condition.

[0034] The aforementioned set of state features refers to a multi-dimensional set of features composed of vehicle speed information and environmental information.

[0035] In one optional embodiment, the collected vehicle speed information and environmental information are preprocessed, and a set of state features is constructed based on the input requirements of the oxygen production mode prediction model and the preprocessed vehicle speed information and environmental information.

[0036] In one optional embodiment, the set of state features includes: vehicle speed, in-vehicle oxygen concentration, out-of-vehicle oxygen concentration, number of occupants, in-vehicle temperature, out-of-vehicle temperature, out-of-vehicle air quality index, in-vehicle humidity, and out-of-vehicle humidity.

[0037] The oxygen generation mode prediction model described above is a classification model obtained through supervised learning pre-training. This model can quickly predict the most suitable oxygen generation mode for the current vehicle operating state, i.e., the target oxygen generation mode, based on the set of state features.

[0038] The aforementioned target oxygen production mode refers to the optimal oxygen production operation mode predicted by the oxygen production mode prediction model based on the current environment and operating status.

[0039] The target oxygen generation modes include, but are not limited to, low-intensity oxygen generation mode, medium-intensity oxygen generation mode, and high-intensity oxygen generation mode.

[0040] In one optional embodiment, an oxygen production mode prediction model is pre-trained on the server side, and then the trained oxygen production mode prediction model is loaded onto the target vehicle to enable the target vehicle to predict the oxygen production mode.

[0041] In another optional embodiment, the model is pre-trained directly on the target vehicle to obtain an oxygen production mode prediction model, and the target vehicle directly uses the trained oxygen production mode prediction model to predict the oxygen production mode.

[0042] Furthermore, the target vehicle is controlled to enter the target oxygen production mode. For example, based on the predicted target oxygen production mode, the oxygen production system parameters of the target vehicle, such as oxygen generator power, damper opening, and blower speed, are adjusted to meet the target oxygen supply demand.

[0043] It should be noted that after the target vehicle enters the target oxygen production mode, it is necessary to continuously monitor changes in the environment and vehicle speed, adjust the control strategy based on feedback, form a closed-loop control mechanism, and ensure continuous optimization of the oxygen production effect.

[0044] In addition, when the target vehicle is detected to be in an extreme situation (such as severe oxygen deficiency or highly polluted environment), it can automatically switch to the preset emergency oxygen generation mode to ensure the safety of the occupants.

[0045] This application first acquires the vehicle speed and environmental information of the target vehicle, enabling a comprehensive perception of the dynamic environment and vehicle speed inside and outside the vehicle cabin. This avoids decision-making biases caused by incomplete or delayed information in traditional control strategies, ensuring that oxygen production control matches the current environmental conditions. Second, based on the acquired vehicle speed and environmental information, a set of state features is constructed. These multi-dimensional features are parameterized, effectively representing the complex states related to the vehicle's oxygen production needs. The construction of this set of state features provides an accurate data foundation for subsequent model predictions, avoiding the biased decision-making caused by controlling a single or few parameters, improving the refinement and comprehensiveness of the oxygen production strategy, and thus enhancing the safety and comfort of passengers. Third, the set of state features is input into an oxygen production mode prediction model pre-trained based on supervised learning. Through the nonlinear mapping capability of this model, the target oxygen production mode that best matches the current environmental and vehicle speed information can be output, achieving dynamic matching of oxygen production needs with environmental conditions and vehicle speed. Finally, based on the predicted target oxygen production mode, the target vehicle is controlled to enter the corresponding oxygen production mode. This closed-loop control strategy ensures that the vehicle can dynamically adjust the control parameters related to oxygen production according to real-time environmental needs and vehicle speed changes, achieving precise satisfaction of oxygen production requirements. Therefore, this application achieves the technical effect of dynamically adjusting the cabin oxygen production strategy to ensure the safety and comfort of passengers, thereby solving the technical problem of rigid cabin oxygen production strategies in related technologies, which in turn affect the safety and comfort of passengers.

[0046] The vehicle control method in the embodiments of this application will be further described below.

[0047] Optionally, the environmental information includes: oxygen concentration information, vital signs information of objects inside the vehicle, temperature information, air quality information, and air humidity information. In step S11, based on the vehicle speed information and environmental information, a set of state features is constructed, including:

[0048] Step S111: Normalize the vehicle speed information, oxygen concentration information, vital signs information of the person inside the vehicle, temperature information, air quality information, and air humidity information to obtain the first processing result.

[0049] Step S112: Based on the first processing result, construct a set of state features.

[0050] The oxygen concentration information mentioned above includes both the oxygen concentration inside and outside the vehicle. The vital signs information mentioned above refers to the occupants inside the vehicle. The vital signs information mentioned above refers to the number of occupants inside the vehicle.

[0051] The temperature information above includes both the interior and exterior temperatures. The air quality information above refers to the exterior air quality index. The humidity information above includes both the interior and exterior humidity.

[0052] The normalization process described above is a data preprocessing technique used to standardize the data distribution and map it to the same interval, usually [0,1]. This can prevent features with larger values ​​from dominating the model training process and ensure that the model has the same sensitivity to all features.

[0053] In an optional embodiment, the normalization operation can be expressed as:

[0054] =(x-min) / (max-min) (1)

[0055] Where min represents the minimum value and max represents the maximum value. This represents the normalization result.

[0056] The first processing result mentioned above refers to the normalized information.

[0057] After normalizing all the information, a set of state features is created, which includes the normalized results of vehicle speed, in-vehicle oxygen concentration, outside-vehicle oxygen concentration, number of occupants, in-vehicle temperature, outside-vehicle temperature, outside-vehicle air quality index, in-vehicle humidity, and outside-vehicle humidity.

[0058] The above steps effectively solved the problems of inconsistent dimensions and large differences in range of the original data, laying the foundation for subsequent model applications. Normalization ensures that each piece of information has an equal weight in the model, preventing bias due to numerical magnitude, thereby improving the fairness and accuracy of decision-making. Constructing a set of state features integrates scattered environmental and vehicle speed information into an ordered whole, facilitating model understanding and processing.

[0059] Optionally, in step S13, controlling the target vehicle to enter the target oxygen generation mode includes:

[0060] Step S131: Query the preset control parameter mapping table to determine the control parameters of the vehicle execution unit corresponding to the target oxygen production mode, wherein the vehicle execution unit is used to characterize the unit in the target vehicle used to realize the oxygen production function.

[0061] Step S132: Control the vehicle execution unit to operate according to the control parameters.

[0062] The aforementioned control parameter mapping table is a predefined database that links different oxygen generation modes with the detailed parameter settings required to control the vehicle's actuators. The design of the control parameter mapping table is based on extensive experimental data and expert rules, ensuring that in specific oxygen generation modes, actuators (such as oxygen generators, ventilation systems, and dampers) operate with the most suitable parameters to achieve the desired safety, comfort, or energy efficiency goals.

[0063] The aforementioned vehicle execution unit refers to the collection of hardware components on a vehicle responsible for performing the oxygen generation function, including but not limited to onboard oxygen generators, blowers, and dampers. These components directly determine the oxygen generation effect, airflow status, and energy consumption level, and are the key carriers for achieving real-time response and dynamic adjustment.

[0064] In one optional embodiment, a control parameter mapping table is queried to extract detailed control parameters corresponding to the target oxygen generation mode. For example, a high-intensity mode requires the oxygen generator power to be set to maximum, the blower speed to high speed, and the damper opening to be fully open to ensure rapid and sufficient oxygen supply and air circulation. A low-power mode, on the other hand, tends to use the oxygen generator at its lowest power, the blower at idle speed, and a moderate damper opening to maximize energy efficiency while ensuring oxygen quality.

[0065] The aforementioned control parameters include, but are not limited to, specific values ​​such as the power rating of the oxygen generator, the speed of the blower, and the opening degree of the damper. These constitute the basic instructions for the dynamic control of the execution unit.

[0066] In one alternative embodiment, once the control parameters are determined, they are sent to the control module of the vehicle's actuator unit. The control module receives instructions from the central processing unit and converts them into executable signals to drive the hardware execution. For example, upon receiving the "high-intensity mode" instruction, the oxygen generator control module adjusts its power to maximum, while the damper control module adjusts the damper opening to 100%, and the blower control module adjusts the blower speed to high speed. This process requires an efficient and reliable internal communication mechanism to ensure that instructions are transmitted instantly and executed accurately by the actuator unit.

[0067] Through the above steps, precise control of the vehicle's actuators can be achieved, ensuring that the target oxygen production mode is executed accurately. The control parameter mapping table acts as a bridge, simplifying the conversion process from oxygen production mode to execution commands and ensuring the accuracy and efficiency of the conversion.

[0068] Optionally, the vehicle control method further includes:

[0069] Step S141: Obtain the target vehicle's historical speed information and historical environmental information;

[0070] Step S142: According to the preset inspection rules, the historical vehicle speed information and historical environmental information are inspected for reasonableness, and the inspection results are obtained. The inspection results are used to characterize whether the historical vehicle speed information and historical environmental information meet the preset inspection rules.

[0071] Step S143: In response to the test result that the historical vehicle speed information and historical environmental information meet the preset test rules, the historical vehicle speed information and historical environmental information are normalized to obtain the second processing result.

[0072] Step S144: Based on the second processing result, construct multiple sets of historical state features;

[0073] Step S145: According to the preset annotation rules, annotate multiple historical state feature sets to obtain a pre-training dataset;

[0074] Step S146: Based on the pre-trained dataset, pre-train the initial classification model to obtain the oxygen production mode prediction model.

[0075] The aforementioned historical vehicle speed information refers to the speed data recorded during the target vehicle's historical driving process. The aforementioned historical environmental information includes historical occupant count, historical oxygen concentration, historical temperature, historical air quality, and historical air humidity. The acquired historical vehicle speed information and historical environmental information are consistent in terms of time dimension.

[0076] The aforementioned pre-defined verification rules refer to a set of standards used to verify the validity and reasonableness of data, ensuring the physical correctness and statistical representativeness of the data.

[0077] In one optional embodiment, after obtaining the historical speed information and historical environmental information of the target vehicle, a rationality check needs to be performed based on preset check rules to check the rationality of the data distribution.

[0078] In one optional embodiment, the preset inspection rules are as follows: the oxygen concentration inside the vehicle is limited to between 400 ppm (fresh air level) and 5000 ppm (safe tolerance limit); the temperature inside the vehicle ranges from -10°C to 45°C, with a comfortable range typically between 22-26°C; the relative humidity inside the vehicle is set between 20% (dry) and 90% (humid); the oxygen concentration outside the vehicle ranges from 400 to 2000 ppm; the outside temperature ranges from -20°C to 50°C; the outside humidity ranges from 10% to 100%; the number of occupants ranges from 0 to the vehicle's approved passenger capacity of 7; the vehicle speed range is set from 0 km / h (stationary) to 120 km / h (highway speed limit); and the outside air quality index follows the standard air quality index, ranging from 0 (excellent) to 500 (severe pollution).

[0079] Optionally, outliers in historical vehicle speed information and historical environmental information can be filtered based on preset verification rules.

[0080] In one optional embodiment, when the historical vehicle speed information and historical environment information meet preset verification rules, the historical vehicle speed information and historical environment information are normalized to obtain normalized historical vehicle speed information and historical environment information. The specific operation of the normalization process can be referred to the above embodiment.

[0081] Based on the second processing result, multiple historical state feature sets are constructed. The vehicle speed and environmental features included in each historical state feature set are consistent over time. The environmental features include: in-vehicle oxygen concentration, outside-vehicle oxygen concentration, number of occupants, in-vehicle temperature, outside-vehicle temperature, outside air quality index, in-vehicle humidity, and outside-vehicle humidity.

[0082] Furthermore, based on preset annotation rules, each historical state feature set is automatically labeled and marked as one of several preset oxygen generation modes, forming a pre-training dataset. These preset oxygen generation modes include, but are not limited to, low-intensity oxygen generation mode, medium-intensity oxygen generation mode, and high-intensity oxygen generation mode.

[0083] The aforementioned initial classification model refers to a machine learning model that has not yet been trained, such as a random forest model or a support vector machine.

[0084] In one optional embodiment, a supervised learning method is used, with the set of historical state features in the pre-training dataset as input and the corresponding oxygen production mode label as output, to train the initial classification model and obtain the oxygen production mode prediction model.

[0085] Through the above steps, a pre-training dataset was constructed, and an oxygen production mode prediction model was trained based on this dataset. The final oxygen production mode prediction model can intelligently predict the operating mode of the oxygen production system based on real-time vehicle speed and environmental information, thereby ensuring the health and safety of passengers while achieving efficient utilization of vehicle resources.

[0086] Optionally, multiple sets of historical state features are labeled according to preset labeling rules, including:

[0087] Step S1451: According to the preset labeling rules, determine the oxygen production mode corresponding to each historical state feature set in the multiple historical state feature sets. The oxygen production mode includes a first oxygen production mode, a second oxygen production mode and a third oxygen production mode. The oxygen production capacity of the first oxygen production mode is greater than that of the second oxygen production mode, and the oxygen production capacity of the second oxygen production mode is greater than that of the third oxygen production mode.

[0088] Step S1452: Label each historical state feature set according to the oxygen production mode corresponding to each historical state feature set.

[0089] The aforementioned preset labeling rules refer to the oxygen production mode decision-making criteria formulated based on the priority principle of safety > comfort > energy saving to ensure the health and comfort of passengers, and are used to determine the oxygen production mode corresponding to the set of historical state characteristics.

[0090] Optionally, when historical state characteristic sets determine that the in-vehicle oxygen deficiency is less than 19%, the external environment is severely polluted, the occupant load is greater than 5 people, or the external environment is oxygen-deficient (oxygen concentration less than 19.5%), the current oxygen generation mode corresponding to the historical state characteristic set is determined to be the high-intensity oxygen generation mode (i.e., the first oxygen generation mode) to ensure safety. In the first oxygen generation mode, the oxygen generator operates at maximum power, the damper opening is adjusted to the maximum, and the blower speed is increased to ensure the safety of occupants and oxygen supply in emergencies. When moderate oxygen deficiency occurs (oxygen concentration less than 20%), low-speed high pollution, moderate occupant load (more than 2 people), or unsuitable temperature and humidity, the current oxygen generation mode corresponding to the historical state characteristic set is determined to be the medium-intensity oxygen generation mode (i.e., the second oxygen generation mode) to balance comfort and energy consumption. The second oxygen generation mode, while ensuring safety, moderately increases the oxygen production and adjusts the ventilation strategy to balance occupant comfort and system energy consumption, providing a suitable in-vehicle air environment. When all environmental parameters are in good condition and the load is low, the current oxygen generation mode corresponding to the historical state feature set is determined as the low-power oxygen generation mode (i.e., the third oxygen generation mode) to optimize energy efficiency. In the third oxygen generation mode, the oxygen generator operates at the lowest power, the damper opening is appropriate, and the blower speed is reduced to achieve energy-saving operation, while ensuring the air quality and passenger comfort inside the vehicle.

[0091] Based on a defined oxygen production mode, the historical state feature set is labeled to provide target values ​​for training samples for subsequent machine learning model training.

[0092] In one optional embodiment, each historical state feature set is labeled according to its corresponding oxygen generation mode. For example, for a historical state feature set, if it is determined to correspond to the first oxygen generation mode according to a preset labeling rule, it is labeled as "1" in the dataset; the second oxygen generation mode is labeled as "2"; and the third oxygen generation mode is labeled as "3". The purpose of this is to transform the abstract environmental states and decision-making logic into numerical forms that the machine learning model can understand, making it easier for the model to "learn" the most suitable oxygen generation mode under different environmental conditions.

[0093] Through the above steps, the determined oxygen production mode is used as a label and associated with the corresponding historical state feature set, thereby constructing high-quality training data. This labeling process based on expert rules effectively transforms the decision-making wisdom of human experts into learning material for the model, enabling the model to master multi-objective trade-off decision-making strategies during the training phase.

[0094] Optionally, the initial classification model is a random forest model. Based on the pre-training dataset, the initial classification model is pre-trained to obtain an oxygen production mode prediction model, including:

[0095] Step S1461: Divide the pre-training dataset into a training dataset and a test dataset according to the preset partitioning rules;

[0096] Step S1462: Using the training dataset, pre-train the random forest model to obtain the oxygen production mode prediction model.

[0097] Splitting the dataset is an important preprocessing step in machine learning, designed to create a training dataset and a test dataset. This process ensures that the model can learn during the training phase and its generalization performance can be evaluated on the independent test dataset.

[0098] In one alternative embodiment, the pre-training dataset is divided into a training dataset and a test dataset in an 8:2 ratio. The training dataset contains 80% of the data and is used to train the initial classification model; the test dataset contains the remaining 20% ​​of the data and is used for performance evaluation of the model, ensuring that the model can not only learn patterns in the data but also be effectively applied to unseen data environments.

[0099] The random forest model described above consists of multiple decision trees. It employs a guided clustering strategy to reduce the risk of overfitting and enhance the model's generalization ability. Each tree is trained on randomly selected features and sample subsets, and the final classification result is determined through a voting mechanism. This makes the random forest model perform exceptionally well in complex classification tasks.

[0100] In one optional embodiment, a random forest model is pre-trained using a training dataset to obtain an oxygen production pattern prediction model. The performance of the oxygen production pattern prediction model is then evaluated using a test dataset.

[0101] Through the above steps, the random forest model is pre-trained using the training dataset to obtain an oxygen production mode prediction model, which is used to predict oxygen production modes, thereby improving the flexibility of oxygen production control.

[0102] Optionally, the random forest model is pre-trained using the training dataset to obtain an oxygen production pattern prediction model, including:

[0103] Step S14621: Initialize the forest structure of the random forest model to obtain the initial forest structure, which includes multiple initial decision trees;

[0104] Step S14622: Based on the training dataset, train multiple initial decision trees in parallel to generate multiple target decision trees;

[0105] Step S14623: Generate an oxygen production mode prediction model based on multiple target decision trees.

[0106] In one alternative embodiment, a forest structure containing 100 initial decision trees is initialized, with the maximum depth of a single tree set to 20 layers. Each tree is trained using two-thirds of all training samples, and the remaining one-third is used as out-of-bag samples for internal validation.

[0107] Meanwhile, during the node splitting process of each tree, the square root of the total number of features in the feature set (e.g., 3 features) is randomly selected as a subset of candidate features. In addition, based on the principle of minimizing Gini impurity, the optimal split point is selected from the candidate features, and the decision tree is recursively constructed, where each leaf node must contain at least 2 samples.

[0108] By generating 100 target decision trees with different structures through parallel training, they are combined to form a random forest classification model, namely an oxygen production mode prediction model, thereby completing the learning of nonlinear mapping relationships from multidimensional feature space to multiple oxygen production modes.

[0109] When a new set of state features is input into the oxygen production mode prediction model, each objective decision tree independently predicts the oxygen production mode. The final prediction mode is determined by voting on these 100 predictions, and the mode with the most votes is selected as the model's prediction.

[0110] Through the above steps, the trained oxygen production mode prediction model can quickly and accurately identify the optimal oxygen production mode when faced with real-time data.

[0111] Optionally, the vehicle control method further includes:

[0112] Step S151: Calculate the classification accuracy of the oxygen production mode prediction model on the test dataset;

[0113] Step S152: In response to the classification accuracy being greater than or equal to a preset accuracy threshold, calculate the pattern index value of the oxygen production mode prediction model on the test dataset. In response to the pattern index value being greater than or equal to a preset index threshold, determine that the oxygen production mode prediction model meets the prediction requirements. The pattern index value is used to measure the ability of the oxygen production mode prediction model to identify multiple oxygen production modes.

[0114] Step S153: In response to the classification accuracy being less than a preset accuracy threshold, the oxygen production mode prediction model is retrained.

[0115] The classification accuracy mentioned above refers to the proportion of correctly classified samples out of the total number of samples when the oxygen production mode prediction model predicts oxygen production mode on the test dataset. Classification accuracy is a straightforward indicator for evaluating the overall performance of the model, directly reflecting the model's accuracy in predicting input data.

[0116] In one optional embodiment, each sample in the test dataset is classified and predicted using an oxygen production mode prediction model. The predicted oxygen production mode output by the model is compared with the actual sample label. The number of correctly predicted samples in all test samples is calculated and then divided by the total number of samples in the test set to obtain the classification accuracy.

[0117] The above-mentioned indicator values ​​include the indicator values ​​corresponding to each oxygen production mode.

[0118] Optionally, the model index value (referred to as the F1 score) corresponding to any oxygen generation mode is the harmonic mean of precision and recall. Precision refers to the proportion of samples predicted as a certain oxygen generation mode that actually belong to that mode. Recall refers to the proportion of samples that actually belong to a certain oxygen generation mode that were correctly predicted by the model.

[0119] In one optional embodiment, when the classification accuracy of the oxygen generation mode prediction model is greater than or equal to a preset accuracy threshold (e.g., 95%), a more detailed analysis of the model's performance is performed. A classification reporting tool is used to generate precision, recall, and F1 scores for each oxygen generation mode. The F1 score for each oxygen generation mode is checked to ensure it is greater than or equal to a preset threshold (e.g., 0.93) to ensure the model's ability to identify all modes meets a high standard. If the F1 scores for all modes meet the threshold, the model is confirmed to meet the prediction requirements.

[0120] When the classification accuracy of the oxygen production mode prediction model is less than the preset accuracy threshold, it indicates that the model's generalization ability or data fitting is insufficient, and retraining is required to improve performance.

[0121] Through the above steps, an effective evaluation of the oxygen production mode prediction model was achieved.

[0122] Figure 2 This is a flowchart of another vehicle control method according to an embodiment of this application, such as... Figure 2 As shown, the implementation of the vehicle control method includes two stages.

[0123] The first stage involves offline model training, which includes: data collection and rationality checks; data normalization (normalization calculation formula is X = X - min / max - min); automatic labeling based on rule base (labeling principle is safety > comfort > energy saving); dataset partitioning, with the training set accounting for 80% and the test set accounting for 20%; training the random forest model; evaluating whether the model meets the standards; if the model evaluation meets the standards, saving the model as a .pkl file; if the model evaluation does not meet the standards, retraining the random forest model.

[0124] The second phase involves online deployment and real-time control, which specifically includes: real-time data acquisition from vehicle-mounted sensors; data normalization processing; loading a pre-trained random forest model; model inference and prediction of the oxygen production mode; and execution of a composite control instruction set.

[0125] In one optional embodiment, the evaluated random forest model is serialized into a pkl format file using the joblib tool and stored. This model file is integrated into the memory of the in-vehicle embedded system and automatically loaded into memory upon system startup. In the real-time control loop, a state feature set is constructed every 500 milliseconds, normalized, and then input into the loaded model. Within 10 milliseconds, the oxygen generation mode inference and prediction are completed. Finally, based on the output mode command, a preset control parameter mapping table is queried to dynamically adjust the three actuator parameters: oxygen generator power, damper opening, and blower speed.

[0126] It's easy to understand that by pre-training the complex computational process offline and deploying only a lightweight inference model on the vehicle side, the resource constraints of in-vehicle embedded systems are perfectly suited. The inherent robustness of the random forest algorithm, combined with the coordinated control of pre-set composite instruction sets for actuators such as oxygen generators and dampers, ensures that the system can make fast, stable, and coordinated responses when facing various complex and dynamic driving environments, significantly improving the overall reliability, energy efficiency, and passenger comfort of the system.

[0127] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, use and processing of the relevant data must comply with the relevant laws, regulations and standards of the relevant countries and regions, and corresponding operation entry points are provided for users to choose to authorize or refuse.

[0128] According to an embodiment of this application, a system embodiment of a vehicle control system is provided. It should be noted that this system can be used to execute the above-described vehicle control method.

[0129] Figure 3 This is a structural block diagram of a vehicle control system according to an embodiment of this application, such as... Figure 3 As shown, the vehicle control system 300 includes: an acquisition module 301 for acquiring vehicle speed information and environmental information of the target vehicle; a construction module 302 for constructing a state feature set based on the vehicle speed information and environmental information, wherein the state feature set is used to characterize multi-dimensional features associated with the oxygen production demand of the target vehicle; a prediction module 303 for inputting the state feature set into an oxygen production mode prediction model to obtain the target oxygen production mode, wherein the oxygen production mode prediction model is a classification model obtained based on supervised learning pre-training; and a control module 304 for controlling the target vehicle to enter the target oxygen production mode.

[0130] Optionally, the environmental information includes: oxygen concentration information, information on vital signs of the person inside the vehicle, temperature information, air quality information, and air humidity information. The construction module 302 is also used to: normalize the vehicle speed information, oxygen concentration information, information on vital signs of the person inside the vehicle, temperature information, air quality information, and air humidity information to obtain a first processing result; and construct a set of state features based on the first processing result.

[0131] Optionally, the control module 304 is further configured to: query a preset control parameter mapping table to determine the control parameters of the vehicle execution unit corresponding to the target oxygen generation mode, wherein the vehicle execution unit is used to characterize the unit in the target vehicle used to achieve the oxygen generation function; and control the vehicle execution unit to operate according to the control parameters.

[0132] Optionally, the vehicle control system also includes a pre-training module (not shown in the figure). The pre-training module is used to: acquire historical vehicle speed information and historical environmental information of the target vehicle; perform a reasonableness check on the historical vehicle speed information and historical environmental information according to preset check rules, and obtain a check result, wherein the check result is used to characterize whether the historical vehicle speed information and historical environmental information meet the preset check rules; in response to the check result that the historical vehicle speed information and historical environmental information meet the preset check rules, perform normalization processing on the historical vehicle speed information and historical environmental information to obtain a second processing result; construct multiple historical state feature sets based on the second processing result; label the multiple historical state feature sets according to preset labeling rules to obtain a pre-training dataset; and pre-train the initial classification model based on the pre-training dataset to obtain an oxygen production mode prediction model.

[0133] Optionally, the pre-training module is further configured to: determine the oxygen generation mode corresponding to each historical state feature set in a plurality of historical state feature sets according to preset annotation rules, wherein the oxygen generation mode includes a first oxygen generation mode, a second oxygen generation mode and a third oxygen generation mode, the oxygen generation capacity of the first oxygen generation mode is greater than that of the second oxygen generation mode, and the oxygen generation capacity of the second oxygen generation mode is greater than that of the third oxygen generation mode; and annotate each historical state feature set according to the oxygen generation mode corresponding to each historical state feature set.

[0134] Optionally, the pre-training module is also used to: divide the pre-training dataset into a training dataset and a test dataset according to a preset partitioning rule; and use the training dataset to pre-train the random forest model to obtain an oxygen production mode prediction model.

[0135] Optionally, the pre-training module is also used to: initialize the forest structure of the random forest model to obtain an initial forest structure, wherein the initial forest structure includes multiple initial decision trees; perform parallel training on the multiple initial decision trees based on the training dataset to generate multiple target decision trees; and generate an oxygen production mode prediction model based on the multiple target decision trees.

[0136] Optionally, the pre-training module is further configured to: calculate the classification accuracy of the oxygen generation mode prediction model on the test dataset; calculate the mode index value of the oxygen generation mode prediction model on the test dataset in response to the classification accuracy being greater than or equal to a preset accuracy threshold; determine that the oxygen generation mode prediction model meets the prediction requirements in response to the mode index value being greater than or equal to a preset index threshold, wherein the mode index value is used to measure the ability of the oxygen generation mode prediction model to identify multiple oxygen generation modes; and retrain the oxygen generation mode prediction model in response to the classification accuracy being less than a preset accuracy threshold.

[0137] Embodiments of this application also provide a vehicle, including: a memory storing an executable program; and a processor for running the program, wherein the program executes the methods described in various embodiments of this application when it runs.

[0138] Embodiments of this application also provide a computer-readable storage medium including a stored executable program, wherein, when the executable program is running, it controls the device where the computer-readable storage medium is located to perform the methods of various embodiments of this application.

[0139] In the above embodiments of this application, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0140] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units can be a logical functional division, and 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 displayed or discussed mutual couplings, direct couplings, or communication connections may be through some interfaces; indirect couplings or communication connections between units or modules may be electrical or other forms.

[0141] 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 units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0142] Furthermore, the functional units in the various embodiments of this application 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. The integrated unit can be implemented in hardware or as a software functional unit.

[0143] If the integrated unit 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 application, in essence, or the part that contributes to the prior art, or all or 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 application. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, read-only memory (ROM), random access memory (RAM), portable hard drive, magnetic disk, or optical disk.

[0144] The above are merely preferred embodiments of this application. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A vehicle control method, characterized in that, include: Obtain the target vehicle's speed and environmental information; Based on the vehicle speed information and the environmental information, a state feature set is constructed, wherein the state feature set is used to characterize multidimensional features associated with the oxygen production requirements of the target vehicle; The set of state features is input into the oxygen production mode prediction model to obtain the target oxygen production mode, wherein the oxygen production mode prediction model is a classification model obtained based on supervised learning pre-training; Control the target vehicle to enter the target oxygen generation mode.

2. The vehicle control method according to claim 1, characterized in that, The environmental information includes: oxygen concentration information, vital signs information of the in-vehicle population, temperature information, air quality information, and air humidity information. Based on the vehicle speed information and the environmental information, the state feature set is constructed, including: The vehicle speed information, oxygen concentration information, vital signs information of the person inside the vehicle, temperature information, air quality information, and air humidity information are normalized to obtain a first processing result; Based on the first processing result, the state feature set is constructed.

3. The vehicle control method according to claim 1, characterized in that, Controlling the target vehicle to enter the target oxygen generation mode includes: Query the preset control parameter mapping table to determine the control parameters of the vehicle execution unit corresponding to the target oxygen production mode, wherein the vehicle execution unit is used to characterize the unit in the target vehicle used to realize the oxygen production function; The vehicle actuator is controlled to operate according to the control parameters.

4. The vehicle control method according to claim 1, characterized in that, The vehicle control method further includes: Obtain the historical speed information and historical environmental information of the target vehicle; According to the preset inspection rules, the historical vehicle speed information and the historical environmental information are subjected to a reasonableness inspection to obtain an inspection result, wherein the inspection result is used to characterize whether the historical vehicle speed information and the historical environmental information meet the preset inspection rules. In response to the test result that the historical vehicle speed information and the historical environment information meet the preset test rules, the historical vehicle speed information and the historical environment information are normalized to obtain a second processing result; Based on the second processing result, multiple sets of historical state features are constructed; According to the preset annotation rules, the multiple historical state feature sets are labeled to obtain the pre-training dataset; Based on the pre-trained dataset, the initial classification model is pre-trained to obtain the oxygen production mode prediction model.

5. The vehicle control method according to claim 4, characterized in that, According to the preset annotation rules, the multiple historical state feature sets are annotated, including: According to the preset labeling rules, the oxygen production mode corresponding to each historical state feature set in the plurality of historical state feature sets is determined. The oxygen production mode includes a first oxygen production mode, a second oxygen production mode and a third oxygen production mode. The oxygen production capacity of the first oxygen production mode is greater than that of the second oxygen production mode, and the oxygen production capacity of the second oxygen production mode is greater than that of the third oxygen production mode. Each historical state feature set is labeled according to the oxygen production mode corresponding to each historical state feature set.

6. The vehicle control method according to claim 4, characterized in that, The initial classification model is a random forest model. Based on the pre-training dataset, the initial classification model is pre-trained to obtain the oxygen production mode prediction model, which includes: According to the preset partitioning rules, the pre-trained dataset is divided into a training dataset and a test dataset; The random forest model is pre-trained using the training dataset to obtain the oxygen production mode prediction model.

7. The vehicle control method according to claim 6, characterized in that, Using the training dataset, the random forest model is pre-trained to obtain the oxygen production mode prediction model, including: The forest structure of the random forest model is initialized to obtain an initial forest structure, wherein the initial forest structure includes multiple initial decision trees; Based on the training dataset, the multiple initial decision trees are trained in parallel to generate multiple target decision trees; Based on the multiple target decision trees, the oxygen production mode prediction model is generated.

8. The vehicle control method according to claim 6, characterized in that, The vehicle control method further includes: Calculate the classification accuracy of the oxygen production mode prediction model on the test dataset; In response to the classification accuracy being greater than or equal to a preset accuracy threshold, the mode index value of the oxygen production mode prediction model on the test dataset is calculated. In response to the mode index value being greater than or equal to a preset index threshold, it is determined that the oxygen production mode prediction model meets the prediction requirements. The mode index value is used to measure the ability of the oxygen production mode prediction model to identify multiple oxygen production modes. In response to the classification accuracy being less than the preset accuracy threshold, the oxygen production mode prediction model is retrained.

9. A vehicle, characterized in that, include: Memory, which stores executable programs; A processor for running the executable program, wherein the executable program, when running, performs the vehicle control method according to any one of claims 1 to 8.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored executable program, wherein, when the executable program is executed, it controls the device on which the computer-readable storage medium is located to perform the vehicle control method according to any one of claims 1 to 8.

Images