Vehicle air inlet system control method and vehicle

Abstract

The invention provides a vehicle air inlet system control method and a vehicle, and relates to the technical field of vehicle control. The working condition information comprises the working state information of the air inlet system, the opening degree information of the grille and the vehicle speed; based on the working condition information, the pollution degree of environmental pollution of the air inlet system is predicted; based on the pollution degree, a target control strategy for the air inlet system is generated, and the air inlet system is controlled to work according to the target control strategy; the target control strategy comprises a grating opening adjustment strategy and an air duct combination strategy. According to the technical scheme, the air inlet quality can be improved, the situation that the air filter element is blocked is avoided, cooperative control over the air duct and the grille based on the pollution degree is achieved, the diversity of control modes is improved, the air conditioner can better adapt to the complex and changeable operation environment, and it is ensured that an engine can work normally.

Description

Technical Field

[0001] This application relates to the field of vehicle control technology, specifically to a vehicle intake system control method and a vehicle. Background Technology

[0002] The vehicle's air intake system is a core system that ensures the normal operation of the engine and plays a role in protecting the engine's key components.

[0003] Currently, most vehicle intake system control strategies are designed based on cooling requirements and weather information, and are executed by adjusting the grille opening and closing state. However, weather information only reflects the external environment of the vehicle and cannot reflect the actual operating status of the vehicle, resulting in inaccurate grille opening and closing state adjustments. Furthermore, vehicle operating environments are complex and diverse, and simple control through the grille cannot adapt to the intake requirements of different operating environments. Summary of the Invention

[0004] In view of this, the embodiments of this application aim to provide a vehicle intake system control method and vehicle, thereby achieving the purpose of improving intake quality and increasing the diversity of control methods.

[0005] In a first aspect, one embodiment of this application provides a vehicle intake system control method, comprising: acquiring vehicle operating condition information; the operating condition information including the operating status information of the intake system, the opening information of the grille, and the vehicle speed; predicting the degree of environmental pollution of the intake system based on the operating condition information; generating a target control strategy for the intake system based on the degree of pollution, and controlling the intake system to operate according to the target control strategy; the target control strategy includes a grille opening adjustment strategy and an air duct combination strategy.

[0006] The vehicle intake system control method provided in this application comprehensively reflects the actual operating status of the vehicle by using the operating status information of the intake system, the opening information of the grille, and the vehicle speed. It predicts the degree of environmental pollution in the intake system using these information, improving the reliability and accuracy of the prediction results. Then, it combines this with a target control strategy for the intake system based on the pollution level. This strategy takes into account the impact of environmental pollution on intake quality when generating the target control strategy, helping to improve intake quality, avoid air filter clogging, and thus improve intake efficiency. Furthermore, the target control strategy includes a grille opening adjustment strategy and an air duct combination strategy. The grille opening adjustment strategy controls the intake volume of the intake system, while the air duct combination strategy controls the airflow direction. This target control strategy achieves coordinated control of the intake volume and airflow direction, increasing the diversity of control methods and ensuring that even in complex and changing operating environments, the impact of environmental pollution on the intake system can be effectively reduced, guaranteeing the normal operation of the engine.

[0007] In conjunction with the first aspect, in some implementations of the first aspect, the degree of environmental pollution of the air intake system is predicted based on operating condition information, including: inputting the operating condition information corresponding to multiple moments within the target time window into a pre-trained dust risk level prediction model to predict the environmental pollution generated by the air intake system in the future time period and obtain the degree of pollution; wherein, the dust risk level prediction model is trained by a temporal convolutional network.

[0008] The dust storm risk level prediction model in this application adopts a temporal convolutional network structure, which can make full use of the temporal characteristics of operating conditions information to capture the dynamic law of dust storm risk changing over time. Furthermore, by taking the operating conditions information at multiple moments within the target time window as input, it can comprehensively reflect the environmental state of the vehicle at different times and the operating status of the air intake system, making the prediction results of pollution level more accurate and reliable.

[0009] In conjunction with the first aspect, in some implementations of the first aspect, before inputting the operating condition information corresponding to multiple moments within the target time window into the pre-trained dust risk level prediction model, the method further includes: obtaining a sample dataset; wherein the sample dataset includes multiple sample data, including the vehicle's operating condition information in historical time periods and the pollution level labels corresponding to the operating condition information in historical time periods; training a temporal convolutional network based on the sample dataset to obtain a basic prediction model; adjusting the weight parameters in the basic prediction model based on the weight adjustment method corresponding to the vehicle type, and determining the basic prediction model after adjusting the weight parameters as the vehicle's dust risk level prediction model; wherein the weight parameters are used to reflect the influence of operating status information, opening information, and vehicle speed on the pollution level.

[0010] In the embodiments of this application, when training the dust risk level prediction model, different adjustment methods for the weight parameters of the basic prediction model are set to take into account the differences of different vehicle models. By adjusting the weight parameters of the basic prediction model, the dust risk level prediction model is obtained, thereby making the dust risk level prediction model more in line with the actual vehicle model requirements and improving the accuracy and reliability of pollution level prediction.

[0011] In conjunction with the first aspect, in some implementations of the first aspect, a target control strategy for the intake system is generated based on the pollution level, including: determining an initial control strategy for controlling the intake system based on the pollution level and operating condition information; performing an optimization operation on the control strategy based on the operating condition information, the initial control strategy, and the vehicle's historical control execution results, and using the obtained optimal control strategy as the target control strategy; the optimization objective of the control strategy is to minimize the operating cost of the intake system, and the factors affecting the operating cost include the amount of air pollution entering, the filter lifespan, and the heat dissipation efficiency.

[0012] This application's embodiments introduce a genetic algorithm, combining population evolution and natural selection mechanisms, to achieve global optimization and efficient search of the control strategy. It can dynamically adjust the control strategy of the intake system under complex and ever-changing operating conditions, thereby ensuring that the optimal control strategy that meets the optimization objective is found, making it more robust and reliable when facing complex operating conditions.

[0013] In conjunction with the first aspect, in some implementations of the first aspect, after determining the initial control strategy for controlling the intake system based on the pollution level and operating condition information, the method further includes: determining the strategy confidence level corresponding to the initial control strategy; if the strategy confidence level does not exceed the confidence level threshold, calling the safety strategy table; and updating the control strategy in the safety strategy table corresponding to the operating condition information and pollution level to the initial control strategy.

[0014] This application embodiment determines the reliability of the initial control strategy through policy confidence. When the reliability of the initial control strategy is poor, the initial control strategy is updated by calling the security policy table, thereby improving the reliability of the initial control strategy. This effectively avoids errors in the optimization direction caused by the deviation of the initial strategy and ensures the accuracy and effectiveness of subsequent optimization operations.

[0015] In conjunction with the first aspect, some implementations of the first aspect also include: detecting the execution time of the optimization operation of the control strategy; if the execution time exceeds a time threshold, stopping the optimization operation and using the initial control strategy as the target control strategy.

[0016] This application embodiment detects the execution time. When the execution time exceeds a threshold, the initial control strategy is directly adopted as the target control strategy. This solves the problem of low optimization efficiency and excessive consumption of computing resources caused by long optimization time under complex working conditions. It not only ensures the effectiveness of the control strategy, but also avoids excessive consumption of computing resources due to excessive optimization, thereby improving the overall control efficiency.

[0017] In conjunction with the first aspect, in certain implementations of the first aspect, an optimization operation of the control strategy is performed based on operating condition information, the initial control strategy, and the vehicle's historical control execution results, and the obtained optimal control strategy is used as the target control strategy. This includes: when the vehicle type is a special environment vehicle, an optimization operation of the control strategy is performed based on operating condition information, the initial control strategy, and the vehicle's historical control execution results, and the obtained optimal control strategy is used as the target control strategy; and / or, when the vehicle type is not a special environment vehicle, the state change rate corresponding to the working state information is determined, and when the state change rate exceeds the change rate threshold, an optimization operation of the control strategy is performed based on operating condition information, the initial control strategy, and the vehicle's historical control execution results, and the obtained optimal control strategy is used as the target control strategy; wherein, special environment vehicles include off-road vehicles and / or mining-specific vehicles.

[0018] The embodiments of this application can flexibly adjust the control strategy generation method according to different vehicle models and operating states, effectively reducing the consumption of computing resources while ensuring the reliability of the target control strategy. For vehicles with special environments, comprehensive optimization is performed directly to adapt to complex operating conditions; for vehicles without special environments, optimization is dynamically determined according to the rate of change of operating states, thereby improving control efficiency and reducing the consumption of computing resources while ensuring the reliability of the control strategy.

[0019] In conjunction with the first aspect, in some implementations of the first aspect, before predicting the degree of environmental pollution of the intake system based on operating condition information, the method further includes: processing the operating condition information to determine the derived feature information of the operating condition information; adding the derived feature information to the operating condition information so as to determine the target control strategy based on the operating condition information after adding the derived feature information.

[0020] This application embodiment processes operating condition information to extract derived feature information and adds this derived feature information to the operating condition information, effectively improving the one-sidedness and limitations that may exist in the original operating condition information, providing richer and more reliable decision support for the generation of control strategies, and thus helping the intake system to maintain efficient and stable operation under various complex operating conditions.

[0021] In conjunction with the first aspect, in some implementations of the first aspect, the working status information includes the wind speed, temperature, humidity, air pressure, particulate matter concentration, and filter element pressure difference of the air intake system duct; the working status information is processed to determine the derived characteristic information of the working status information, including at least one of the following: statistically analyzing the particulate matter concentration within the target time window to determine the slope change, extreme value duration, and abrupt change frequency of the particulate matter concentration, and using the slope change, extreme value duration, and abrupt change frequency as derived characteristic information; fusing the wind speed and vehicle speed within the target time window to obtain the forward dynamic pressure value borne by the air duct within the target time window, and using the forward dynamic pressure value as derived characteristic information; determining the pressure difference change based on the filter element pressure difference within the target time window, and determining the load growth rate of the filter element in the air intake system based on the pressure difference change, and using the load growth rate as derived characteristic information.

[0022] This application provides specific methods for determining derived feature information. By analyzing the slope change of particulate matter concentration, the duration of extreme values, and the frequency of abrupt changes, it effectively reflects the explosive change trend of sandstorms; by using forward dynamic pressure values, it assesses the direct impact of external airflow on the intake system; and by using the load growth rate, it measures the particulate blockage process. Therefore, the derived feature information provided in this application can more comprehensively capture the dynamic characteristics of operating conditions.

[0023] Secondly, this application provides a vehicle intake system control device, comprising: an information acquisition module for acquiring vehicle operating condition information; the operating condition information includes the operating status information of the intake system, the opening information of the grille, and the vehicle speed; a pollution prediction module for predicting the degree of environmental pollution of the intake system based on the operating condition information; and a strategy generation module for generating a target control strategy for the intake system based on the degree of pollution, and controlling the intake system to operate according to the control strategy; the target control strategy includes a grille opening adjustment strategy and an air duct combination strategy.

[0024] Thirdly, one embodiment of this application provides a computer-readable storage medium storing a computer program for performing the method in the first aspect or any possible implementation of the first aspect.

[0025] Fourthly, one embodiment of this application provides a vehicle, the vehicle comprising: a processor; a memory for storing processor-executable instructions; the processor being configured to execute the method in the first aspect or any possible implementation thereof.

[0026] Fifthly, one embodiment of this application provides a computer program product including instructions that, when executed on a vehicle, cause the vehicle to implement the method in the first aspect or any possible implementation of the first aspect.

[0027] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of this application, nor is it intended to limit the scope of this application. Other features of this application will become readily apparent from the following description. Attached Figure Description

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

[0029] Figure 1 The diagram shown is a flowchart of a vehicle intake system control method provided in an embodiment of this application.

[0030] Figure 2 The diagram shown is a flowchart of a vehicle intake system control method provided in another embodiment of this application.

[0031] Figure 3 The diagram shown is a structural schematic of a vehicle intake system control device provided in an embodiment of this application.

[0032] Figure 4 The diagram shown is a structural schematic of a vehicle provided in an embodiment of this application. Detailed Implementation

[0033] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of this application are within the scope of protection of this application.

[0034] The terminology used in the embodiments of this application is for the purpose of describing particular embodiments only and is not intended to be limiting of this application. The singular forms “a,” “the,” and “the” used in the embodiments of this application and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise.

[0035] It should be understood that the term "and / or" used in this document is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Furthermore, the character " / " in this document generally indicates that the preceding and following related objects have an "or" relationship. Additionally, the term "based on" used in this document is not limited to relying solely on one object. For example, determining B based on A can mean: determining B based solely on A, or determining B partially based on A.

[0036] Before introducing the technical solution, an exemplary application scenario can be provided. This technical solution can be applied to scenarios involving controlling the air intake system of a vehicle. For example, the air intake system includes a grille, air ducts, and a filter element, which introduces external air into the engine compartment to provide the engine with the necessary air intake volume.

[0037] It should be noted that when vehicles are driven in areas with heavy dust storms, a large amount of particulate matter can directly enter the engine compartment through the open grille, causing rapid clogging of the air filter, reducing intake efficiency, and even triggering abnormal engine thermal management. Currently, most vehicle intake systems are designed with simple opening and closing adjustments based solely on cooling needs and weather information. Weather information only reflects the external environment of the vehicle and cannot reflect the actual operating status, leading to inaccurate grille opening and closing adjustments. Furthermore, vehicle operating environments are complex and diverse, and simple control via the grille cannot adapt to the intake requirements of different operating environments.

[0038] To address the aforementioned issues, this embodiment can predict the environmental pollution status of the intake system, thereby controlling the opening and closing state of the grille and the air duct combination based on the pollution level. This helps to prevent air filter blockage and improve intake efficiency. Furthermore, the coordinated control of the grille and air duct based on the pollution level increases the diversity of control methods, enabling better adaptation to complex and changing operating environments and ensuring the engine can operate normally.

[0039] The following is combined with Figures 1 to 2 The vehicle intake system control method provided in the embodiments of this application will be described in detail.

[0040] Figure 1 The diagram shown is a schematic flowchart of a vehicle intake system control method provided in an embodiment of this application; as follows: Figure 1 As shown, the method includes the following steps.

[0041] Step S110: Obtain the vehicle's operating condition information.

[0042] Optionally, the operating information includes the operating status of the intake system, the grille opening information, and the vehicle speed. The operating status information includes the airflow speed, temperature, humidity, air pressure, particulate matter concentration, and filter pressure differential of the intake system within the air duct. The opening information reflects the current degree of grille opening, while the vehicle speed reflects the vehicle's travel speed.

[0043] In some embodiments, vehicle operating condition information can be collected using various sensors deployed within the vehicle. For example, particulate matter concentration is collected using a laser scattering sensor installed at the rear of the grille, near the air filter inlet, which reflects the real-time concentration of fine and inhalable particulate matter in the inhaled airflow. Wind speed is measured using a hot-film anemometer fixed in the air intake path at the center of the grille, requiring a dust cover and self-cleaning structure to address dust adhesion issues. Additionally, a multi-functional sensor can be used to collect temperature and humidity data. This multi-functional sensor is positioned upstream of the air duct. Air pressure can be collected using a differential pressure sensor, located in the sections before and after the air filter, to analyze the trend of differential pressure changes in the air duct and aid in airflow condition assessment. Filter differential pressure sensors are deployed at both ends of the air filter, outputting analog signals to estimate the degree of filter clogging. Grille opening information is read using an angle sensor or a stepper motor feedback position sensor. Vehicle speed can be obtained from the vehicle's VCU (Vehicle Control Unit) via the vehicle's CAN (Controller Area Network) bus.

[0044] In practical implementation, different operating condition information can be collected simultaneously and stored according to a time sequence window. The collected operating condition information is then uniformly accessed to the vehicle edge controller. For example, the collection frequency can be 10Hz.

[0045] In this embodiment, the collected operating condition information may suffer from noise interference, short-term power outages, and signal drift. To improve information accuracy and effectively analyze the environmental pollution situation of the air intake system, the operating condition information can be preprocessed. For example, the operating condition information can be a multi-source signal stream from front-end sensors, including three-dimensional components of particulate matter concentration, wind speed, vehicle speed, temperature, humidity, air pressure, filter pressure difference, and grille opening information, totaling eight types of signals. First, outlier detection and imputation are performed on each type of signal. To address short-term missing data, a time-axis linear interpolation method is used to reconstruct continuous signals, ensuring data integrity within the time window. For sensor drift, dynamic upper and lower thresholds are set by referencing multi-vehicle average data to remove data segments exceeding physically reasonable ranges. Signal jitter is particularly pronounced in wind speed and pressure difference signals; a first-order low-pass filter or moving average window can be used to smooth the data, effectively suppressing high-frequency disturbances. After completing the above operations, each type of signal is normalized. Since different signals have different characteristic units and dimensions, all signals are standardized to convert them into dimensionless data with a mean of zero and a variance of one. The normalized data is then used as new operating condition information. For example, for each type of signal, standardization can be performed based on the mean and standard deviation of that type of signal within a time window.

[0046] Step S120: Based on operating condition information, predict the degree of environmental pollution in the intake system.

[0047] In actual vehicle operation, the main cause of air filter clogging is sandstorm weather, with a large amount of particulate matter entering the engine compartment through the grille. Therefore, this embodiment can primarily predict the degree of environmental pollution caused by sandstorms. The degree of pollution can be expressed as a percentage or according to different levels. For example, the degree of pollution can be divided into three levels: low risk, medium risk, and high risk.

[0048] In some embodiments, a physical model of the intake system can be constructed based on fluid dynamics, particulate transport, and chemical reaction theories, and the pollution level of the intake system in future time periods can be predicted through numerical simulation.

[0049] Step S130: Based on the degree of pollution, generate a target control strategy for the intake system, and control the intake system to operate according to the target control strategy.

[0050] The target control strategy includes a grille opening adjustment strategy and an air duct combination strategy. The grille opening adjustment strategy indicates the degree of grille opening to regulate the airflow into the engine compartment. For example, it indicates the grille's open / closed state as fully open, partially open, or fully closed. The air duct combination strategy indicates the open or closed state of different air ducts to optimize airflow paths and reduce the direct impact of particulate matter on the filter element. For example, the air ducts include a main air duct, a secondary air duct, and an internal recirculation bypass air path. The air duct combination strategy is used to indicate whether the main air duct, secondary air duct, and internal recirculation bypass air path are open, respectively.

[0051] It should be noted that after the pollution level prediction is completed, the target control strategy needs to be implemented quickly to achieve refined management of the air inflow path to the engine compartment. The core task of the target control strategy is to translate the current environmental state and predicted risk level into explicit execution commands, controlling the opening and closing of the grilles and the airflow guidance direction of the ducts.

[0052] Specifically, a pre-trained policy generation model can be used to determine the target control policy required for the pollution level. For example, the policy generation model is trained using a GBDT (Gradient Boosting Decision Tree) model. The GBDT model structure consists of multiple decision trees, employing a layer-by-layer iterative training method to construct a non-linear feature space, suitable for handling scenarios with low input dimensionality but complex feature relationships. During training, the split points and leaf weights of each tree are continuously optimized by minimizing the loss function, allowing the model to gradually approximate the actual control policy output. This structure does not require the training resource consumption of deep neural networks, while retaining good generalization ability and robustness. Choosing the GBDT model as the policy generation model offers advantages such as high accuracy, interpretability, and fast response, and can adapt to the combined characteristics of complex environmental variables and execution conditions.

[0053] Optionally, the input to the strategy generation model includes the wind speed, temperature, humidity, air pressure, particulate matter concentration, filter pressure difference, grille opening information, and pollution level of the duct sampled in the most recent frame, to ensure real-time response. The output of the strategy generation model is a grille opening adjustment strategy and a duct combination strategy. The two strategies together constitute a complete control decision to guide the actions of the actuators.

[0054] After generating the target control strategy, it can be sent to the vehicle execution layer to ensure accurate execution of the grille and air duct actions. Specifically, control commands corresponding to the target control strategy can be generated, including grille control parameters and air duct control parameters. The grille control parameters represent the percentage value of the grille opening, adjustable in five-digit steps (21 controllable values ​​between 0 and 100), corresponding to the pulse width modulation signal amplitude or step angle of the motor control system. The air duct control parameters reflect the opening and closing of each air duct. For example, a Boolean vector can be used to reflect the opening and closing status of each air duct, coded as 0 for closed and 1 for open, arranged according to the predefined air duct structure sequence of the vehicle.

[0055] Furthermore, the generated control commands can be sent to the grille actuators and duct actuators. Upon receiving the control commands, each actuator parses them according to the protocol and performs the physical actions. The grille actuators, composed of electric actuators or stepper motors, have position feedback capabilities and transmit the current opening status in real time after executing the control commands. Duct actuators typically employ electric flappers or solenoid valve structures, and their status feedback signals are also uploaded to the central controller via a LIN (Local Interconnect Network) bus or CAN interface, forming a closed-loop control circuit.

[0056] The vehicle intake system control method provided in this application comprehensively reflects the actual operating status of the vehicle by using the operating status information of the intake system, the opening information of the grille, and the vehicle speed. It predicts the degree of environmental pollution in the intake system using these information, improving the reliability and accuracy of the prediction results. Then, it combines this with a target control strategy for the intake system based on the pollution level. This strategy takes into account the impact of environmental pollution on intake quality when generating the target control strategy, helping to improve intake quality, avoid air filter clogging, and thus improve intake efficiency. Furthermore, the target control strategy includes a grille opening adjustment strategy and an air duct combination strategy. The grille opening adjustment strategy controls the intake volume of the intake system, while the air duct combination strategy controls the airflow direction. This target control strategy achieves coordinated control of the intake volume and airflow direction, increasing the diversity of control methods and ensuring that even in complex and changing operating environments, the impact of environmental pollution on the intake system can be effectively reduced, thus ensuring the normal operation of the engine.

[0057] In order to quickly and accurately determine the degree of pollution, in some other embodiments, the specific implementation of predicting the degree of environmental pollution of the air intake system based on operating condition information is as follows: the operating condition information corresponding to multiple moments within the target time window is input into a pre-trained dust risk level prediction model to predict the environmental pollution generated by the air intake system in the future time period and obtain the degree of pollution.

[0058] Optionally, the dust storm risk level prediction model can be trained using a Temporal Convolutional Network (TCN). The TCN structure contains multiple one-dimensional convolutional channels and employs causal convolution to ensure the model focuses only on information from the current and historical time periods, without incorporating future data. An increasing expansion rate strategy is used between layers to progressively expand the receptive field, covering long-term dependencies in the input sequence. Each convolutional layer is followed by an activation function and a normalization module to accelerate convergence and prevent overfitting. A residual connection mechanism introduces a skip structure to improve the stability of deep networks and avoid gradient vanishing.

[0059] In practice, the operating condition information corresponding to multiple moments within the target time window is first acquired. The target time window is a preset time range with the current moment as the cutoff point. The target time window includes multiple moments, and the operating condition information corresponding to each moment can be acquired separately, serving as input information. For example, operating condition information continuously collected within 30 seconds can be input into the dust storm risk level prediction model, covering 300 frames of data. Each frame includes dimensions such as particulate matter concentration, vehicle speed, wind speed, temperature and humidity, air pressure, filter pressure difference, and grille status. This input structure preserves the changing trajectory of dust storm characteristics and the interaction between the vehicle's behavior and the external environment, providing crucial data for risk level evolution.

[0060] The dust storm risk level prediction model outputs the degree of environmental pollution. For example, the model outputs a classification label of low risk, medium risk, or high risk to reflect the level of pollution. The label definition depends on the actual vehicle test results and filter load assessment standards. Low risk indicates that the external particle concentration is within the permissible inhalation range; medium risk indicates that there may be a short-term decrease in filter performance but it will not affect the function of the intake system; high risk indicates that a large amount of particulate matter will enter in a short period of time, requiring the grille to be closed and the air duct adjusted to prevent intrusion.

[0061] It should be noted that, to meet practical deployment requirements, the architecture of the dust storm risk level prediction model has been trimmed and quantized on an automotive-grade chip, removing redundant channels and reducing memory usage. The computation time for each prediction does not exceed one hundred milliseconds, meeting the grid control system's response requirement of within three hundred milliseconds. Verified tests show that the output stability of the dust storm risk level prediction model reaches over 90%, with the misclassification rate remaining within acceptable limits. Under extreme dust storm conditions, the model can still accurately determine high-risk levels, effectively supporting the triggering of subsequent control logic.

[0062] The dust storm risk level prediction model in this application adopts a temporal convolutional network structure, which can make full use of the temporal characteristics of operating conditions information to capture the dynamic law of dust storm risk changing over time. Furthermore, by taking the operating conditions information at multiple moments within the target time window as input, it can comprehensively reflect the environmental state of the vehicle at different times and the operating status of the air intake system, making the prediction results of pollution level more accurate and reliable.

[0063] Currently, a uniform dust storm risk level prediction model is used to predict pollution levels for different vehicles. However, different vehicle models have different sensitivities to pollution. Using a uniform dust storm risk level prediction model to predict pollution levels will result in predictions that are not applicable to the current vehicle, thus affecting the accuracy of subsequent control strategies.

[0064] To accurately predict the pollution levels of different vehicle models and precisely control the air intake system, this embodiment provides a training method for a dust risk level prediction model. Optionally, before inputting the operating condition information corresponding to multiple moments within the target time window into the pre-trained dust risk level prediction model, the method further includes: acquiring a sample dataset; training a temporal convolutional network based on the sample dataset to obtain a basic prediction model; adjusting the weight parameters in the basic prediction model based on the weight adjustment method corresponding to the vehicle model, and determining the basic prediction model with adjusted weight parameters as the vehicle's dust risk level prediction model.

[0065] The sample dataset includes multiple sample data sets, which contain vehicle operating condition information for historical time periods and corresponding pollution level labels. For example, the sample dataset covers various operating conditions and weather conditions. Specifically, it includes sample data related to urban, plateau, sandstorm, and arid plain conditions; it also includes sample data corresponding to weather conditions such as early morning / evening, after rain, and dust storms.

[0066] In this embodiment, cross-validation can be used during the training of the temporal convolutional network to avoid overfitting and data bias, thereby improving the model's generalization ability. The model obtained after training is used as the base prediction model.

[0067] Considering the varying sensitivities of different vehicle models to dust storms, the model's weighting parameters need to be adjusted to match the predicted pollution levels with actual sensitivities. Specifically, the vehicle models for which pollution levels need to be predicted should be identified first. For example, these models could include high-end new energy vehicles, economy vehicles, and SUVs.

[0068] In this embodiment, weighting parameters are used to reflect the influence of operating status information, opening information, and vehicle speed on the pollution level. For example, operating status information includes the airflow speed, temperature, humidity, air pressure, particulate matter concentration, and filter pressure difference of the air intake system duct. Weighting parameters include the weight values ​​corresponding to airflow speed, temperature, humidity, air pressure, particulate matter concentration, filter pressure difference, opening information, and vehicle speed. The weighting adjustment method indicates how the weighting parameters in the basic prediction model are adjusted.

[0069] Different weighting adjustments are applied to different vehicle models. For high-end new energy vehicles, due to their more sophisticated air intake systems, they are more sensitive to dust storms. Therefore, the weighting adjustment for high-end new energy vehicles involves increasing the weight of particulate matter concentration in the basic prediction model to emphasize air quality maintenance. For example, if the weight of particulate matter concentration in the basic prediction model is the target value, the weighting adjustment for high-end new energy vehicles would be to increase the target value by 10%. For economy vehicles, whose design prioritizes a balance between cost and practicality, they are relatively less sensitive to dust storms. Therefore, the weighting adjustment for economy vehicles involves decreasing the weight of particulate matter concentration in the basic prediction model. For example, the target value would be decreased by 5%.

[0070] Furthermore, the pollution thresholds used to determine the degree of pollution differ in the dust storm risk level prediction models corresponding to different vehicle models. For example, high-end new energy vehicles are more sensitive to dust storms, while economy vehicles are relatively less sensitive. Therefore, the pollution threshold for high-end new energy vehicles is set relatively lower than that for economy vehicles.

[0071] The model training method proposed in this application takes into account the differences between different vehicle models when training the dust risk level prediction model. It sets different adjustment methods for the weight parameters of the basic prediction model. By adjusting the weight parameters of the basic prediction model, the dust risk level prediction model is obtained. That is, different vehicle models correspond to different dust risk level prediction models, thereby making the dust risk level prediction model more in line with the actual needs of vehicle models. This solves the problem that the existing technology uses only a uniform model for pollution prediction for different vehicles, resulting in inaccurate prediction results and improves the accuracy of pollution level prediction.

[0072] To further enhance the adaptability of the intake system in complex environments, a combined optimization mechanism can be introduced. This mechanism is multi-objective and searches for the optimal combination of grille and duct configuration within a limited time, taking into account particulate matter interception effect, filter load control, airflow stability, and heat dissipation requirements.

[0073] Specifically, the implementation method for generating the target control strategy for the intake system based on the pollution level is as follows: Based on the pollution level and operating condition information, an initial control strategy for controlling the intake system is determined; based on the operating condition information, the initial control strategy, and the vehicle's historical control execution results, an optimization operation is performed on the control strategy, and the obtained optimal control strategy is taken as the target control strategy; the optimization objective of the control strategy is to minimize the operating cost of the intake system, and the factors affecting the operating cost include the amount of air pollution entering, the filter lifespan, and the heat dissipation efficiency.

[0074] In some implementations, pollution levels and operating conditions can be input into a pre-trained strategy generation model to determine the control strategy required for that pollution level, and the control strategy output by the strategy generation model can be used as the initial control strategy.

[0075] It should be noted that the strategy generation model possesses advantages such as high accuracy, interpretability, and fast response, and can adapt to the combination of complex environmental variables and execution conditions. The deployment structure of the strategy generation model is lightweight, with inference time well below one hundred milliseconds, exhibiting high interpretability. Different strategy generation models can be set for the differences between different vehicle models. For example, for vehicles that prioritize protecting filter lifespan, the corresponding strategy generation model shifts the control strategy for samples with higher risk levels towards a more conservative approach, prioritizing the closure of the grille. For electric drive platform vehicles that focus on cooling efficiency, it allows for partially open grilles under moderate risk conditions, balancing air quality and heat dissipation capacity. Furthermore, depending on the different air intake system layouts of the vehicle models, the output of the strategy generation model can be configured as a three-channel or four-channel combination to adapt to the actual air duct switching structure. The strategy generation model is deployed on the vehicle's edge controller, achieving upstream and downstream decoupling through interface communication with the dust risk level prediction model.

[0076] Furthermore, the initial control strategy, operating condition information, and historical control execution results can form a control state vector, which serves as a constraint reference for individual encoding. Based on the genetic algorithm and the control state vector, an optimization operation is performed on the control strategy, and the resulting final control strategy is taken as the target control strategy.

[0077] The operating information includes the most recently sampled data such as airflow speed, temperature, humidity, air pressure, particulate matter concentration, filter pressure differential, duct resistance coefficient, cooling requirement level, and grille opening information. Historical control execution results are data recorded during past periods of intake system operation, including grille opening, duct configuration, and corresponding intake system operating costs. This data reflects the effectiveness of different control strategies in practical applications.

[0078] It should be noted that the optimization objective of the genetic algorithm is to search for a set of combinations of grille opening and air duct configurations that minimize the operating cost of the intake system. Operating cost is determined by the amount of air pollution entering the system, filter lifespan, and heat dissipation efficiency. Each individual in the genetic algorithm represents a specific set of grille and air duct parameter configurations. The grille opening is encoded using a continuous percentage value, and the air duct selection uses Boolean vectors to represent the activation status of each channel. The entire search space is bounded within the limits allowed by the actual control system to prevent the generation of invalid or unexecutable solutions.

[0079] The initial control strategy is used as the first generation of individuals in the genetic algorithm. A small perturbation is introduced to generate the population, and the remaining individuals are constructed through random sampling across the entire space to enhance population diversity. The fitness function comprehensively considers the weighted scores of air pollution intake, filter lifespan, and heat dissipation efficiency in each generation's evaluation. For example, the weights can be set according to different vehicle models or strategy preferences. For instance, if filter lifespan is emphasized, the particulate matter intake index is set as the primary scoring item; if the vehicle uses an electric drive cooling-sensitive platform, the weight of duct heat exchange efficiency is increased. The evolutionary process employs standard selection, crossover, and mutation operations. Individuals with high fitness are retained through a roulette wheel approach for crossover operations to generate new solutions, while a certain probability of gene mutation is introduced to escape local optima. The entire iterative process is completed within a limited number of generations, for example, set to within fifty generations. The total computational load meets the on-board real-time requirements. The optimal individual is updated to the candidate solution pool after each round of evolution, and the control strategy with the highest score is finally selected as the target control strategy.

[0080] In practical implementation, an optimization decision-making module can be deployed locally on the controller to achieve collaborative work with the GBDT model. This provides the GBDT model with policy direction, while the optimization decision-making module uses a genetic algorithm to fine-tune and refine the scheme, forming an optimization closed loop. Each executed target control strategy, along with its actual execution effect, is written to the historical database for subsequent updates to empirical rules and for the next round of population initialization.

[0081] This application's embodiments introduce a genetic algorithm, combining population evolution and natural selection mechanisms, to achieve global optimization and efficient search of the control strategy. It can dynamically adjust the control strategy of the intake system under complex and ever-changing operating conditions, thereby ensuring that the optimal control strategy that meets the optimization objective is found, making it more robust and reliable when facing complex operating conditions.

[0082] To avoid affecting the accuracy of the optimization results due to an inaccurate initial control strategy, the accuracy of the initial control strategy can be verified after it is determined. Optionally, after determining the initial control strategy for controlling the intake system based on the pollution level and operating condition information, the method further includes: determining the strategy confidence level corresponding to the initial control strategy; if the strategy confidence level does not exceed the confidence level threshold, calling the safety strategy table; and updating the control strategies in the safety strategy table corresponding to the operating condition information and pollution level to the initial control strategy.

[0083] Policy confidence is used to measure the reliability of the initial control policy. It can be determined in various ways. For example, when generating the initial control policy using a policy generation model, the policy confidence of that initial control policy can be output simultaneously.

[0084] In practical implementation, if the strategy confidence level does not exceed the preset confidence threshold, it indicates that the initial control strategy has poor reliability. Optimization based on this initial strategy will affect the accuracy of the optimization results. To ensure the effectiveness and accuracy of the final target control strategy, a safety strategy table can be invoked. This table stores the mapping relationship between pollution levels and control strategies. The safety strategy table can be constructed based on historical control execution results of the intake system, ensuring that reliable control strategies are provided under various operating conditions and combinations of pollution levels to guarantee engine safety.

[0085] Furthermore, the control policy in the safety policy table that matches the current operating conditions and pollution level is identified and updated to the initial control policy. This ensures that even if the initial control policy is not highly reliable, alternative solutions can be provided through the safety policy table, guaranteeing the accuracy and effectiveness of subsequent optimization operations.

[0086] In addition, when the confidence level of the strategy exceeds the preset confidence level threshold, it indicates that the reliability of the initial control strategy is high. In this case, optimization operations can be performed directly based on the initial control strategy to obtain the target control strategy.

[0087] This application embodiment determines the reliability of the initial control strategy through policy confidence. When the reliability of the initial control strategy is poor, the initial control strategy is updated by calling the security policy table, thereby improving the reliability of the initial control strategy. This effectively avoids errors in the optimization direction caused by the deviation of the initial strategy and ensures the accuracy and effectiveness of subsequent optimization operations.

[0088] To improve the efficiency of strategy generation, the vehicle intake system control method also includes: detecting the execution time of the optimization operation of the control strategy; if the execution time exceeds the time threshold, stopping the optimization operation and using the initial control strategy as the target control strategy.

[0089] Specifically, during the optimization of the control strategy, a time threshold is set, which can be reasonably set according to actual control requirements and system performance. During the optimization process, the execution time can be monitored in real time using tools such as timers. When the execution time exceeds the preset time threshold, it indicates that the optimization has not yielded results after a long period, and continuing the optimization may not find a better control strategy. To improve optimization efficiency and avoid consuming excessive computational resources, the optimization operation can be stopped. Since the initial control strategy already meets the basic control requirements, it can be directly used as the target control strategy to ensure that the intake system is effectively controlled within a reasonable timeframe, thus ensuring the normal operation of the engine.

[0090] This application embodiment detects the execution time. When the execution time exceeds a threshold, the initial control strategy is directly adopted as the target control strategy. This solves the problem of low optimization efficiency and excessive consumption of computing resources caused by long optimization time under complex working conditions. It not only ensures the effectiveness of the control strategy, but also avoids excessive consumption of computing resources due to excessive optimization, thereby improving the overall control efficiency.

[0091] It should be noted that, in order to ensure the reliability of the target control strategy while minimizing the consumption of computing resources, different target control strategy generation methods can be set based on different vehicle models and operating states.

[0092] Optionally, based on operating condition information, the initial control strategy, and the vehicle's historical control execution results, an optimization operation is performed on the control strategy, and the obtained optimal control strategy is used as the target control strategy. This includes: when the vehicle type is a special environment vehicle, performing an optimization operation on the control strategy based on operating condition information, the initial control strategy, and the vehicle's historical control execution results, and using the obtained optimal control strategy as the target control strategy; and / or, when the vehicle type is not a special environment vehicle, determining the state change rate corresponding to the operating state information, and when the state change rate exceeds the change rate threshold, performing an optimization operation on the control strategy based on operating condition information, the initial control strategy, and the vehicle's historical control execution results, and using the obtained optimal control strategy as the target control strategy.

[0093] For example, special environment vehicles include off-road vehicles and / or mining vehicles, which typically need to operate in complex and harsh environments, requiring more sophisticated control strategies for the intake system. In addition to special environment vehicles, vehicles operating in normal environments may also be included, such as passenger cars and urban SUVs, whose operating environments are relatively stable and do not require overly sophisticated control strategies for the intake system.

[0094] Different control strategies can be determined for different vehicle models. Specifically, for vehicles designed for special environments, which have higher requirements for control strategies, when the vehicle model requiring intake system control is identified as a special environment vehicle, an optimization operation is performed directly based on operating condition information, the initial control strategy, and the vehicle's historical control execution results. The resulting optimal control strategy is then used as the target control strategy to meet the high requirements of special environment vehicles for intake system control, ensuring that the engine can still operate normally in complex and harsh environments.

[0095] For vehicles not designed for special environments, to maximize the efficiency of control strategy generation and minimize computational resource consumption, the rate of change of the vehicle's operating state information can be further assessed. The rate of change refers to how quickly the vehicle's operating state information changes. For example, the rate of change can be the change in wind speed, temperature, and particulate matter concentration per unit time. When the rate of change exceeds a preset threshold, it indicates that the vehicle is in a rapidly changing operating environment. In this case, based on the operating condition information, the initial control strategy, and the vehicle's historical control execution results, an optimization operation is performed on the control strategy. The resulting optimal control strategy is used as the target control strategy to adapt to the rapidly changing environment and ensure the safe operation of the engine. When the rate of change does not exceed the threshold, it indicates that the vehicle's operating state is relatively stable. In this case, the control strategy generation process can be simplified. For example, the initial control strategy can be directly adopted, or fine-tuned based on historical control execution results, to reduce computational resource consumption and improve control efficiency.

[0096] The embodiments of this application can flexibly adjust the control strategy generation method according to different vehicle models and operating states, effectively reducing the consumption of computing resources while ensuring the reliability of the target control strategy. For vehicles with special environments, comprehensive optimization is performed directly to adapt to complex operating conditions; for vehicles without special environments, optimization is dynamically determined according to the rate of change of operating states, thereby improving control efficiency and reducing the consumption of computing resources while ensuring the reliability of the control strategy.

[0097] In some embodiments, in order to improve the accuracy and reliability of the target control strategy, before predicting the degree of environmental pollution of the intake system based on the operating condition information, the method further includes: processing the operating condition information to determine the derived feature information of the operating condition information; adding the derived feature information to the operating condition information so as to determine the target control strategy based on the operating condition information after adding the derived feature information.

[0098] Among them, derived feature information refers to additional information that can more comprehensively reflect the vehicle's operating status and environmental characteristics, obtained by analyzing and transforming the original operating condition information.

[0099] In practical implementation, the acquired operating condition information can be processed to extract derived feature information. For example, information processing includes time-series analysis of the operating condition information to capture its dynamic trends. For instance, by calculating the moving average, rate of change, or fluctuation range of information such as wind speed, temperature, and particulate matter concentration over a time series, derived feature information reflecting the dynamic characteristics of the operating conditions can be generated. Alternatively, spatial correlation analysis can be performed on the operating condition information, such as combining sensor data from different locations in the air duct to generate spatial features reflecting the uniformity of airflow distribution or the degree of local pollution accumulation. These derived feature information can supplement the correlations and dynamic characteristics not directly reflected in the original operating condition information, providing richer decision-making basis for subsequent control strategy generation.

[0100] This application embodiment processes operating condition information to extract derived feature information and adds this derived feature information to the operating condition information, effectively improving the potential bias and limitations of the original operating condition information. This provides richer and more reliable decision support for the generation of control strategies, thereby helping the intake system maintain efficient and stable operation under various complex operating conditions.

[0101] Optionally, the operating status information includes the airflow velocity, temperature, humidity, air pressure, particulate matter concentration, and filter element pressure difference of the intake system within the target time window. The operating status information is processed to determine derived characteristic information, including at least one of the following: statistical analysis of particulate matter concentration within the target time window to determine the slope change, extreme value duration, and abrupt change frequency of particulate matter concentration, using the slope change, extreme value duration, and abrupt change frequency as derived characteristic information; fusion processing of airflow velocity and vehicle speed within the target time window to obtain the forward dynamic pressure value experienced by the airflow duct within the target time window, using the forward dynamic pressure value as derived characteristic information; determining the pressure difference change based on the filter element pressure difference within the target time window, and based on the pressure difference change, determining the load growth rate of the filter element in the intake system, using the load growth rate as derived characteristic information.

[0102] It should be noted that the slope change of particulate matter concentration reflects the rate of change in particulate matter concentration within the target time window. The duration of extreme values ​​reflects the duration of extremely high or low particulate matter concentrations within the target time window, which is crucial for assessing the intake system's ability to cope with extreme pollution conditions. The frequency of abrupt changes reflects the frequency of particulate matter concentration changes, helping to determine whether the intake system is operating in a frequently fluctuating pollution environment. By introducing the slope change, extreme value duration, and abrupt change frequency, the explosive trends of dust storms can be characterized.

[0103] Forward dynamic pressure is used to assess the direct impact of external airflow on the intake system. In practice, the wind speed and vehicle speed within the target time window can be timestamped, and the wind speed and vehicle speed after timestamping can be vector-synthesized to obtain the forward dynamic pressure value that the air duct experiences within the target time window.

[0104] The pressure difference change is the difference in filter element pressure within the target time window. For example, the filter element pressure difference values ​​at the beginning and end of the target time window can be selected. Subtracting the filter element pressure difference value at the beginning from the end time yields the pressure difference change. Further, the pressure difference change can be multiplied by the air pressure correction factor and the humidity correction factor, and the product is taken as the load growth rate. The load growth rate measures the particle clogging process. A high load growth rate indicates that the filter element may reach a high load state in a short time, affecting the normal operation of the intake system; conversely, a low load growth rate indicates that the filter element clogging process is relatively slow, and the intake system can maintain a stable operating state for a longer period. Accurate assessment of the load growth rate allows for the prediction of filter element replacement or maintenance time in advance, thereby ensuring the continuous and efficient operation of the intake system.

[0105] This application provides specific methods for determining derived feature information. By analyzing the slope change of particulate matter concentration, the duration of extreme values, and the frequency of abrupt changes, it effectively reflects the explosive change trend of sandstorms; by using forward dynamic pressure values, it assesses the direct impact of external airflow on the intake system; and by using the load growth rate, it measures the particulate blockage process. Therefore, the derived feature information provided in this application can more comprehensively capture the dynamic characteristics of operating conditions.

[0106] The following is combined with Figure 2 An example is given to illustrate the control method of a vehicle's air intake system. In this embodiment, the control process of the vehicle's air intake system is executed by an on-board edge controller.

[0107] like Figure 2As shown, firstly, sensors can be deployed at multiple locations on the vehicle to acquire its operating condition information. However, the collected operating condition information suffers from noise interference, short-term interruptions, and signal drift. Therefore, preprocessing such as cleaning, correction, and structuring is necessary. The operating condition information is a multi-source signal stream, including eight types of signals: three-dimensional component of particulate matter concentration, wind speed, vehicle speed, temperature, humidity, air pressure, filter pressure difference, and the current opening state of the grille. During feature construction, to address the dual needs of dust prediction and duct adjustment, secondary derived feature information is further extracted from the operating condition information. All derived feature information is added to the operating condition information, ensuring dimensional consistency across all information. Furthermore, to improve information structure compatibility and processing efficiency, all operating condition information is uniformly organized into a three-dimensional tensor structure. This structure must simultaneously meet the data requirements of the TCN model for temporal convolution and be compatible with the feature reading formats of strategy generation models and genetic algorithms.

[0108] It should be noted that the dimensions of operating condition information differ for different vehicle models. For example, high-end vehicles include visual information such as dust coverage and forward transparency to improve the completeness of the operating condition information. For economy vehicles, only the three-dimensional components of particulate matter concentration, wind speed, vehicle speed, temperature, humidity, air pressure, filter pressure difference, and the current opening status of the grille are retained to reduce computational resource consumption.

[0109] Secondly, a dust storm risk level prediction model is constructed using TCN (Traffic Risk Network), and the pollution level is predicted using this model and operating condition information. A gradient boosting decision tree is selected as the strategy generation model, and the initial control strategy for the intake system is determined using the strategy generation model, pollution level, and operating condition information. Furthermore, the initial control strategy is optimized using a genetic algorithm to obtain the target control strategy.

[0110] Finally, to ensure the long-term effective operation of the air intake system and continuously improve its response accuracy, the dust risk level prediction model, strategy generation model, and genetic algorithm can be adjusted and maintained online. Through remote learning and iterative updates of the model weights, a complete edge-cloud closed-loop learning system is formed.

[0111] For example, for a dust storm risk level prediction model, high-quality data snippets from the most recent hour can be automatically extracted at fixed intervals to construct a mini training set. By fine-tuning the parameters of the last few convolutional layers of the dust storm risk level prediction model, it adapts to changes in local features without altering the core structure, improving the model's short-term prediction performance. For a policy generation model, online parameter adjustment is supported. By monitoring the deviation between the output of the policy generation model and the actual execution feedback, the deviation is recorded in a local log, forming a behavior deviation database. Based on the deviation database, a small-batch incremental model is periodically retrained on the vehicle-mounted edge controller, achieving policy self-correction while maintaining the same response speed. For genetic algorithms, genetic parameters such as population size, crossover rate, and mutation rate can be dynamically adjusted based on the fitness convergence speed, independent variable search efficiency, and stability of execution results under the current environment.

[0112] It should be noted that the dust risk level prediction model, strategy generation model, and genetic algorithm provided in this application embodiment can be applied to different vehicles using transfer learning and module decoupling strategies.

[0113] In order to achieve stable operation of the air intake system in complex environments, this application embodiment organically integrates a dust risk level prediction model, a strategy generation model, and a genetic algorithm to construct a complete control link with self-sensing, self-decision-making, and self-evolution capabilities. The multi-model linkage not only improves control accuracy but also enhances the system's rapid response capability to sudden environmental changes and improves the reliability of the air intake system.

[0114] The above text combined Figures 1 to 2 The present application describes in detail embodiments of the vehicle intake system control method, which are described below in conjunction with... Figure 3 This application provides a detailed description of embodiments of the vehicle intake system control device. It should be understood that the descriptions of the vehicle intake system control method embodiments correspond to the descriptions of the vehicle intake system control device embodiments; therefore, any parts not described in detail can be found in the preceding method embodiments.

[0115] Figure 3 The diagram shown is a structural schematic of a vehicle intake system control device according to an embodiment of this application. Figure 3 As shown, the vehicle intake system control device 30 provided in this application embodiment includes: The information acquisition module 310 is used to acquire the vehicle's operating condition information, including the operating status information of the air intake system, the opening information of the grille, and the vehicle speed. The pollution prediction module 320 is used to predict the degree of environmental pollution in the air intake system based on operating condition information. The strategy generation module 330 is used to generate a target control strategy for the intake system based on the degree of pollution, and to control the operation of the intake system according to the control strategy. The target control strategies include grille opening adjustment strategies and air duct combination strategies.

[0116] In one embodiment of this application, the pollution prediction module 320 is further configured to input the operating condition information corresponding to multiple moments within the target time window into a pre-trained dust risk level prediction model to predict the environmental pollution generated by the air intake system in the future time period and obtain the pollution level; wherein, the dust risk level prediction model is trained by a temporal convolutional network.

[0117] In one embodiment of this application, the pollution prediction module 320 is further configured to acquire a sample dataset before inputting the operating condition information corresponding to multiple times within the target time window into the pre-trained dust risk level prediction model; wherein, the sample dataset includes multiple sample data, the sample data including the vehicle's operating condition information in historical time periods and the pollution degree labels corresponding to the operating condition information in historical time periods; the temporal convolutional network is trained based on the sample dataset to obtain a basic prediction model; the weight parameters in the basic prediction model are adjusted based on the weight adjustment method corresponding to the vehicle model, and the basic prediction model after adjusting the weight parameters is determined as the dust risk level prediction model for the vehicle; wherein, the weight parameters are used to reflect the influence of the operating status information, the opening information, and the vehicle speed on the pollution degree.

[0118] In one embodiment of this application, the strategy generation module 330 is further configured to determine an initial control strategy for controlling the intake system based on the pollution level and operating condition information; and to perform an optimization operation on the control strategy based on the operating condition information, the initial control strategy and the vehicle's historical control execution results, and to take the obtained optimal control strategy as the target control strategy; the optimization objective of the control strategy is to minimize the operating cost of the intake system, and the factors affecting the operating cost include the amount of air pollution entering, the filter life and the heat dissipation efficiency.

[0119] In one embodiment of this application, the device further includes: a strategy table invocation module, configured to determine the strategy confidence level corresponding to the initial control strategy after determining the initial control strategy for controlling the intake system based on the pollution level and operating condition information; if the strategy confidence level does not exceed the confidence level threshold, invoke a safety strategy table; and update the control strategy in the safety strategy table corresponding to the operating condition information and pollution level to the initial control strategy.

[0120] In one embodiment of this application, the device further includes: a time detection module, used to detect the execution time of the optimization operation of the control strategy; if the execution time exceeds a time threshold, the optimization operation is stopped, and the initial control strategy is used as the target control strategy.

[0121] In one embodiment of this application, the strategy generation module 330 is further configured to, when the vehicle type is a special environment vehicle, perform an optimization operation on the control strategy based on the operating condition information, the initial control strategy, and the vehicle's historical control execution results, and use the obtained optimal control strategy as the target control strategy; and / or, when the vehicle type is not a special environment vehicle, determine the state change rate corresponding to the working state information, and when the state change rate exceeds the change rate threshold, perform an optimization operation on the control strategy based on the operating condition information, the initial control strategy, and the vehicle's historical control execution results, and use the obtained optimal control strategy as the target control strategy; wherein, special environment vehicles include off-road vehicles and / or mining area special vehicles.

[0122] In one embodiment of this application, the device further includes: an information processing module, configured to process the operating condition information before predicting the degree of environmental pollution of the intake system based on the operating condition information, determine the derived feature information of the operating condition information, and add the derived feature information to the operating condition information so as to determine the target control strategy based on the operating condition information after adding the derived feature information.

[0123] In one embodiment of this application, the working status information includes the wind speed, temperature, humidity, air pressure, particulate matter concentration, and filter element pressure difference of the air intake system duct; the information processing module is further configured to perform at least one of the following: perform statistical analysis on the particulate matter concentration within the target time window, determine the slope change, extreme value duration, and abrupt change frequency of the particulate matter concentration, and use the slope change, extreme value duration, and abrupt change frequency as derived feature information; perform fusion processing on the wind speed and vehicle speed within the target time window to obtain the forward dynamic pressure value borne by the air duct within the target time window, and use the forward dynamic pressure value as derived feature information; determine the pressure difference change based on the filter element pressure difference within the target time window, determine the load growth rate of the filter element in the air intake system based on the pressure difference change, and use the load growth rate as derived feature information.

[0124] It is worth noting that in the embodiments of the above-mentioned vehicle intake system control device, the various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of each functional unit are only for easy differentiation and are not used to limit the scope of protection of the present invention.

[0125] Below, for reference Figure 4 To describe the vehicle according to embodiments of this application. Figure 4 The diagram shown is a structural schematic of a vehicle provided in an exemplary embodiment of this application.

[0126] like Figure 4 As shown, vehicle 40 includes one or more processors 401 and memory 402.

[0127] The processor 401 may be a central processing unit (CPU) or other form of processing unit with data processing and / or instruction execution capabilities, and may control other components in the vehicle 40 to perform desired functions.

[0128] The memory 402 may include one or more computer program products, which may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and / or cache memory. Non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc. One or more computer program instructions may be stored on the computer-readable storage medium, and the processor 401 may execute the program instructions to implement the vehicle intake system control methods of the various embodiments of this application described above, and / or other desired functions.

[0129] In one example, vehicle 40 may also include input device 403 and output device 404, which are interconnected via a bus system and / or other forms of connection mechanism (not shown).

[0130] The input device 403 may include, for example, a keyboard, a mouse, etc.

[0131] The output device 404 can output various information to the outside. The output device 404 may include, for example, a display, a speaker, a printer, and a communication network and its connected remote output devices, etc.

[0132] Of course, for the sake of simplicity, Figure 4 Only some of the components of the vehicle 40 relevant to this application are shown in this illustration; components such as buses, input / output interfaces, etc., are omitted. In addition, the vehicle 40 may include any other suitable components depending on the specific application.

[0133] In addition to the methods and apparatus described above, embodiments of this application may also be computer program products, which include computer program instructions that, when executed by a processor, cause the processor to perform the steps in the vehicle intake system control methods according to various embodiments of this application as described above.

[0134] Computer program products can be written in any combination of one or more programming languages ​​to perform the operations of the embodiments of this application. The programming languages ​​include object-oriented programming languages ​​such as Java and C++, as well as conventional procedural programming languages ​​such as C or similar languages. The program code can be executed entirely on the user's computing device, partially on the user's computing device, as a standalone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server.

[0135] Furthermore, embodiments of this application may also be computer-readable storage media storing computer program instructions thereon, which, when executed by a processor, cause the processor to perform the steps in the vehicle intake system control method according to the various embodiments of this application described above.

[0136] Computer-readable storage media may take the form of any combination of one or more readable media. A readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may, for example, include, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatuses, or devices, or any combination thereof. More specific examples of readable storage media (a non-exhaustive list) include: electrical connections having one or more wires, portable disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0137] The basic principles of this application have been described above with reference to specific embodiments. However, it should be noted that the advantages, benefits, and effects mentioned in this application are merely examples and not limitations, and should not be considered as essential features of each embodiment of this application. Furthermore, the specific details of the above application are for illustrative and facilitative purposes only, and are not limitations. These details do not limit the application to the necessity of employing the aforementioned specific details for implementation.

[0138] The block diagrams of devices, apparatuses, devices, and systems involved in this application are merely illustrative examples and are not intended to require or imply that they must be connected, arranged, or configured in the manner shown in the block diagrams. As those skilled in the art will recognize, these devices, apparatuses, devices, and systems can be connected, arranged, and configured in any manner. Words such as “comprising,” “including,” “having,” etc., are open-ended terms meaning “including but not limited to,” and are used interchangeably with them. The terms “or” and “and” as used herein refer to the terms “and / or,” and are used interchangeably with them unless the context clearly indicates otherwise. The term “such as” as used herein refers to the phrase “such as but not limited to,” and is used interchangeably with it.

[0139] It should also be noted that in the apparatus, equipment, and methods of this application, the components or steps can be disassembled and / or recombined. These disassemblies and / or recombinations should be considered as equivalent solutions of this application.

[0140] The above description of the claimed aspects is provided to enable any person skilled in the art to make or use this application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the general principles defined herein can be applied to other aspects without departing from the scope of this application. Therefore, this application is not intended to be limited to the aspects shown herein, but rather to be applied within the widest scope consistent with the principles and novel features of this application.

[0141] The above description has been given for illustrative and descriptive purposes. Furthermore, this description is not intended to limit the embodiments of this application to the forms described herein. Although numerous exemplary aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, alterations, additions, and sub-combinations thereof.

Claims

1. A method for controlling a vehicle intake system, characterized in that, include: Obtain vehicle operating condition information; the operating condition information includes the operating status information of the air intake system, the opening information of the grille, and the vehicle speed; Based on the operating condition information, the degree of environmental pollution of the air intake system is predicted; Based on the pollution level, a target control strategy for the air intake system is generated, and the air intake system is controlled to operate according to the target control strategy. The target control strategy includes a grille opening adjustment strategy and an air duct combination strategy.

2. The method according to claim 1, characterized in that, The prediction of the environmental pollution level of the intake system based on the operating condition information includes: The operating condition information corresponding to multiple moments within the target time window is input into a pre-trained dust storm risk level prediction model to predict the environmental pollution generated by the air intake system in the future time period and obtain the pollution level. The dust storm risk level prediction model is obtained by training a temporal convolutional network.

3. The method according to claim 2, characterized in that, Before inputting the operational information corresponding to multiple moments within the target time window into the pre-trained dust storm risk level prediction model, the following steps are also included: Obtain a sample dataset; wherein the sample dataset includes multiple sample data, the sample data including the vehicle's operating condition information over a historical time period and the pollution level label corresponding to the operating condition information over the historical time period; The temporal convolutional network is trained based on the sample dataset to obtain a basic prediction model; Based on the weight adjustment method corresponding to the vehicle model, the weight parameters in the basic prediction model are adjusted, and the basic prediction model after adjusting the weight parameters is determined as the dust risk level prediction model for the vehicle. The weighting parameter is used to reflect the influence of the working status information, the opening information, and the vehicle speed on the degree of pollution.

4. The method according to claim 1, characterized in that, The generation of a target control strategy for the intake system based on the pollution level includes: Based on the pollution level and the operating condition information, an initial control strategy for controlling the intake system is determined. Based on the operating condition information, the initial control strategy, and the vehicle's historical control execution results, an optimization operation is performed on the control strategy, and the obtained optimal control strategy is taken as the target control strategy. The optimization objective of the control strategy is to minimize the operating cost of the intake system. Factors affecting the operating cost include the amount of air pollution entering the system, the lifespan of the filter element, and the heat dissipation efficiency.

5. The method according to claim 4, characterized in that, After determining the initial control strategy for controlling the intake system based on the pollution level and the operating condition information, the method further includes: Determine the policy confidence level corresponding to the initial control policy; If the policy confidence level does not exceed the confidence threshold, the security policy table is invoked; Update the control policy in the safety policy table that corresponds to the operating condition information and the pollution level to the initial control policy.

6. The method according to claim 4, characterized in that, Also includes: Detect the execution time of the optimization operation of the control strategy; If the execution time exceeds a time threshold, the optimization operation is stopped, and the initial control strategy is used as the target control strategy.

7. The method according to claim 4, characterized in that, The step of optimizing the control strategy based on the operating condition information, the initial control strategy, and the vehicle's historical control execution results, and using the obtained optimal control strategy as the target control strategy, includes: When the vehicle is a special environment vehicle, based on the operating condition information, the initial control strategy, and the vehicle's historical control execution results, an optimization operation is performed on the control strategy, and the resulting optimal control strategy is used as the target control strategy; and / or, If the vehicle model is not the special environment model, determine the state change rate corresponding to the working state information. If the state change rate exceeds the change rate threshold, perform an optimization operation on the control strategy based on the working condition information, the initial control strategy and the vehicle's historical control execution results, and take the obtained optimal control strategy as the target control strategy. The special environment vehicles mentioned include off-road vehicles and / or mining area special vehicles.

8. The method according to claim 1, characterized in that, Before predicting the pollution level of the intake system based on the operating condition information, the method further includes: The operating condition information is processed to determine the derived feature information of the operating condition information; The derived feature information is added to the operating condition information so that the target control strategy can be determined based on the operating condition information after adding the derived feature information.

9. The method according to claim 8, characterized in that, The operating status information includes the air velocity, temperature, humidity, air pressure, particulate matter concentration, and filter element pressure difference in the air intake system duct; the information processing of the operating status information to determine the derived feature information of the operating status information includes at least one of the following: Statistical analysis is performed on the particulate matter concentration within the target time window to determine the slope change, extreme value duration, and abrupt change frequency of the particulate matter concentration. The slope change, extreme value duration, and abrupt change frequency are used as the derived feature information. The wind speed and vehicle speed within the target time window are fused to obtain the forward dynamic pressure value that the air duct experiences within the target time window, and the forward dynamic pressure value is used as the derived feature information. The pressure difference change is determined based on the pressure difference of the filter element within the target time window. Based on the pressure difference change, the load growth rate of the filter element in the intake system is determined, and the load growth rate is used as the derived feature information.

10. A vehicle, characterized in that, include: processor; Memory used to store the processor's executable instructions; The processor is used to execute the vehicle intake system control method according to any one of claims 1 to 9.

Images