Vehicle-mounted facility control method and device based on driving behavior pre-judgment and electronic equipment

Abstract

The invention discloses a vehicle-mounted facility control method and device based on driving behavior pre-judgment and electronic equipment, and relates to the field of vehicle control, and the method comprises the steps: collecting driving behaviors and historical data related to the driving behaviors, carrying out the preprocessing of the driving behaviors and the historical data related to the driving behaviors, generating an input data set, and generating a training data set according to the input data set; generating a behavior prediction model according to the training data set; collecting data of a current driving behavior; inputting a behavior prediction model according to the data of the current driving behavior, and obtaining a control intention for the vehicle-mounted facility; according to the control intention of the vehicle-mounted facility, a control instruction of the vehicle-mounted equipment is generated, and whether the control instruction of the vehicle-mounted equipment meets the expectation or not is judged; if yes, the vehicle-mounted equipment executes a control action; if not, the vehicle-mounted equipment does not execute the control action; wherein the judgment result data is fed back to the behavior prediction model by judging whether the control instruction of the vehicle-mounted equipment meets the expectation or not, and the behavior prediction model is iteratively updated.

Description

Technical Field

[0001] The present application relates to the field of vehicle control, and in particular to a method for controlling an on-board facility based on driving behavior prediction, an on-board facility control device based on driving behavior prediction, an electronic device, and a storage medium. Background Art

[0002] With the development of intelligent vehicles, users are increasingly demanding more convenient and personalized vehicle operation. In traditional vehicles, operating multiple features (such as seat adjustment, air conditioning temperature, and audio systems) often requires multiple manual operations. This not only reduces driving convenience but can also distract the driver and increase safety risks. Therefore, it is crucial to develop a system that can intelligently predict and automatically activate relevant in-vehicle features based on the driver's historical operating behavior. Summary of the Invention

[0003] In view of this, the purpose of the present invention is to provide a vehicle-mounted facility control method, device, electronic device and storage medium based on driving behavior prediction. The present invention will use a vehicle facility intelligent prediction and activation system based on the analysis of the owner's historical operation behaviors of multiple vehicle facilities. It aims to automatically predict the owner's subsequent operation intentions through data analysis and machine learning technology, and activate relevant in-vehicle facilities in advance, thereby improving the user experience without the need for additional equipment.

[0004] The present invention provides the following solutions:

[0005] According to one aspect of the present invention, a method for controlling vehicle-mounted facilities based on driving behavior prediction is provided, comprising:

[0006] Collect historical data on driving behavior and driving behavior related data;

[0007] Preprocessing driving behavior and historical data related to driving behavior to generate input data sets;

[0008] Generate a training data set based on the input data set;

[0009] Generate a behavior prediction model based on the training data set;

[0010] Collect data on current driving behavior;

[0011] Based on the current driving behavior data, the behavior prediction model is input to obtain the control intention of the vehicle facilities;

[0012] Generate control instructions for on-board equipment based on the control intention of on-board facilities;

[0013] Determine whether the control instructions of the on-board equipment meet expectations;

[0014] If,it meets the requirements, then the vehicle-mounted device performs the control action;

[0015] If,it does not meet the requirements, the vehicle-mounted device does not perform the control action;

[0016] Among them, based on whether the control instructions of the on-board equipment meet expectations, the judgment result data is fed back to the behavior prediction model, and the behavior prediction model is iteratively updated.

[0017] Further, including:

[0018] Driving behavior and historical data related to driving behavior, including: collecting the driver's driving operations, driving trajectory, vehicle dynamic parameters, operation events of on-board facilities during driving, time and vehicle environment information during operation.

[0019] Furthermore, it also includes:

[0020] The preprocessing of driving behavior and historical data related to driving behavior includes: integrating the collected data and verifying the integrated data.

[0021] Further, including:

[0022] Generating a behavior prediction model, including: the behavior prediction model is constructed based on a neural network;

[0023] The neural network includes: a long short-term memory network or a gated recurrent unit;

[0024] The neural network includes: the number of neurons, the number of layers and the activation function of the input layer, the hidden layer and the output layer.

[0025] Furthermore, it also includes:

[0026] The input layer receives an input data set;

[0027] The hidden layer consists of multiple units that capture the temporal dependencies in the input dataset;

[0028] The output layer outputs the predicted driving behavior, which includes steering angle, acceleration, and air conditioning switch operation.

[0029] Furthermore, it also includes:

[0030] Assign initial values ​​to the connection weights of the neural network.

[0031] Furthermore, it also includes:

[0032] Calculation loss functions include: mean square error and cross entropy loss;

[0033] Get the backpropagation algorithm to backpropagate the loss function gradient, calculate the gradient and update the model parameters.

[0034] According to two aspects of the present invention, there is provided a vehicle-mounted facility control device based on driving behavior prediction, the vehicle-mounted facility control device based on driving behavior prediction comprising:

[0035] Acquisition module, data preprocessing module, model training module, model analysis module, instruction generation module, control judgment module and feedback update module;

[0036] A collection module is used to collect driving behavior and historical data related to driving behavior and collect data on current driving behavior;

[0037] A data preprocessing module is used to preprocess the historical data of driving behavior and driving behavior related data to generate an input data set;

[0038] The model training module is used to generate a training data set based on the input data set, and generate a behavior prediction model based on the training data set;

[0039] The model analysis module is used to input the behavior prediction model based on the current driving behavior data to obtain the control intention of the vehicle facilities;

[0040] An instruction generation module is used to generate control instructions for the vehicle-mounted equipment according to the control intention of the vehicle-mounted facilities;

[0041] A control judgment module is used to judge whether the control instructions of the vehicle-mounted equipment meet expectations;

[0042] The feedback update module is used to judge whether the control instructions of the on-board equipment meet expectations, feed back the judgment result data to the behavior prediction model, and iteratively update the behavior prediction model.

[0043] According to three aspects of the present invention, there is provided an electronic device, comprising: a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other via the communication bus;

[0044] The memory stores a computer program, which, when executed by the processor, causes the processor to execute the steps of the vehicle facility control method based on driving behavior prediction.

[0045] According to four aspects of the present invention, a computer-readable storage medium is provided, which stores a computer program executable by an electronic device. When the computer program runs on the electronic device, the electronic device executes the steps of a method for controlling vehicle facilities based on driving behavior prediction.

[0046] This application utilizes an intelligent predictive activation system for vehicle facilities based on the owner's historical operating behavior. Through big data analysis and machine learning, it accurately predicts the owner's subsequent operating intentions and automatically activates relevant in-vehicle facilities, significantly improving user experience and driving safety. Furthermore, the system requires no additional equipment, reducing implementation costs and possessing broad application prospects and market value.

[0047] This application calculates the loss function, feeds back the judgment result data to the behavior prediction model, iteratively updates the behavior prediction model, measures the deviation between the prediction results of the behavior prediction model and the actual driver's operation, and further improves the accuracy of judging the driver's operation intention.

[0048] This application builds a behavior prediction model to achieve dynamic prediction by combining time dependency and driver habits, thereby improving the accuracy of judging the driver's operating intention. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the specific embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0050] Figure 1 This is a flowchart of a method for controlling vehicle facilities based on driving behavior prediction provided by one or more embodiments of the present invention.

[0051] Figure 2 This is a structural diagram of a vehicle-mounted facility control device based on driving behavior prediction provided by one or more embodiments of the present invention.

[0052] Figure 3 A schematic diagram of a vehicle-mounted facility control device based on driving behavior prediction according to a specific embodiment of the present invention.

[0053] Figure 4 A block diagram of an electronic device structure of a vehicle-mounted facility control device based on driving behavior prediction provided by one or more embodiments of the present invention. DETAILED DESCRIPTION

[0054] The technical solution of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0055] Figure 1This is a flowchart of a method for controlling vehicle facilities based on driving behavior prediction provided by one or more embodiments of the present invention.

[0056] like Figure 1 The vehicle-mounted facility control method shown here based on driving behavior prediction includes:

[0057] Collect historical data on driving behavior and driving behavior related data;

[0058] Preprocessing driving behavior and historical data related to driving behavior to generate input data sets;

[0059] Specifically, it collects driving behavior and historical data related to driving behavior, including: using various sensors and data recorders built into the vehicle to collect the owner's driving information in real time, including vehicle speed, acceleration, braking force, steering angle, etc.

[0060] The system also records the driver's actions on vehicle features, such as seat adjustment, air conditioning temperature settings, and audio system selection, as well as contextual information such as the time these actions occurred and the environment (such as weather and road conditions). This acquired data is stored in the cloud or in the vehicle's local storage system. Cloud computing is used to clean and preprocess the collected data. Multimodal data collection can introduce errors, such as differences in data formats and standards between different data sources. Data cleaning eliminates these inconsistencies by standardizing data formats and standards. Historical data often contains missing values, which may be caused by data collection equipment failures, data transmission interruptions, or human error. Data cleaning can address missing values ​​using appropriate methods, such as interpolation, mean filling, and mode filling. It can also remove records with a large number of missing values ​​and fill them in, or perform inferences based on logical relationships between the data. Duplicate data not only consumes storage resources but also interferes with data analysis results. Data cleaning can identify and remove duplicate records.

[0061] Through data cleaning, historical data in different data formats is converted and standardized. For data such as date and time, it is converted to a unified format by writing scripts or using data processing tools. For numerical data, normalization or standardization may be required, such as mapping the data to the interval [0, 1] or converting it to a standard normal distribution with a mean of 0 and a standard deviation of , to facilitate comparison and analysis between different data sets.

[0062] If abnormal speed data due to sensor failure is used directly for training without cleaning, the neural network may mistakenly associate abnormal speed with specific driving behaviors, leading to prediction bias. Enhancing data consistency ensures that driving behavior data from different sources, such as vehicle sensor data and navigation system data, is input into the model in a unified format, facilitating the model's learning of data features. Improving data integrity by properly filling in missing driving operation records and vehicle status parameters ensures the completeness of model training data and avoids inadequate model learning due to missing data. Removing duplicate data reduces redundant information interference, allowing the model to focus on valid data features, improving training efficiency and prediction accuracy.

[0063] By preprocessing the collected data, accurate operational data is generated to improve the accuracy of model predictions and provide an accurate and consistent data foundation for model construction.

[0064] Further, a training data set is generated based on the input data set;

[0065] Generate a behavior prediction model based on the training data set;

[0066] Specifically, for driving behavior prediction tasks, driving behavior data exhibits distinct time series characteristics, with temporal dependencies between preceding and subsequent actions. For example, a driver's steering operation is often correlated with factors such as previous driving direction and speed. RNNs process sequential data through recurrent connections in their hidden layers, memorizing information from previous moments and applying it to current calculations. LSTMs and GRUs, by introducing gating mechanisms, address the vanishing or exploding gradient issues that plague RNNs when processing long sequences. These mechanisms allow them to better capture long-range temporal dependencies and more accurately learn how driving behavior changes over extended time spans.

[0067] For complex neural network models and large datasets, using appropriate initialization methods can significantly reduce the time and computing resources required for training. For example, in deep convolutional neural networks for image recognition, proper initialization can enable the network to achieve ideal accuracy with fewer training rounds, accelerating model development and deployment.

[0068] Furthermore, data on current driving behavior is collected;

[0069] Based on the current driving behavior data, the behavior prediction model is input to obtain the control intention of the vehicle facilities;

[0070] Generate control instructions for on-board equipment based on the control intention of on-board facilities;

[0071] Specifically, the judgment is based on the vehicle's current state, environmental conditions, and the user's regular usage habits. In terms of air conditioning control, if the vehicle's interior temperature is already in the comfortable range (e.g., 22-26°C), and the system generates frequent temperature adjustment instructions, the system will combine historical data on the user's operating habits at comfortable temperatures and determine that these instructions may be misoperations or abnormal instructions that do not conform to normal usage logic.

[0072] Consistency checks primarily check whether the command matches the user's recent actions and preferences. For example, if a user has consistently adjusted their seat to a specific position consistent with their long-standing sitting habits, and a sudden command for a significant change in seat position is generated, the system will compare historical data with the current command to determine whether there has been a false trigger or anomaly.

[0073] In multimedia system control, if a user has long preferred to play a certain type of music, and suddenly an instruction to switch to a completely different style of music appears, the system will verify the consistency of the instruction based on the user's historical playback records and preference settings. If the difference is too large from the user's habits, the instruction may be judged as not meeting expectations.

[0074] By inputting data from current driving behavior into a behavior prediction model, the system can determine control intent for in-vehicle features. In one embodiment, touch and gesture control captures user actions through the device's interactive interface and maps control intent based on pre-set operational logic. For example, by long-pressing the volume icon on the center console and swiping upward, the system recognizes the intent to "turn up the volume."

[0075] Another embodiment collects user data on past in-vehicle device operations, including air conditioning temperature adjustment habits, seat position preferences, and multimedia playback preferences. For example, if a user consistently sets the air conditioning temperature to 22°C at 8:00 AM on weekdays, the system can predict the user's intention to adjust the air conditioning to that temperature under similar times and environmental conditions. By creating user profiles and digitizing user behavior patterns, association rule algorithms are used to uncover potential connections between operational behaviors. Once a specific behavior is detected, the system can infer potential subsequent operational intentions.

[0076] Regardless of whether the control instruction is ultimately executed, the system will feed back the relevant data and results of the judgment process to the behavior prediction model to achieve iterative updates of the model and improve its prediction accuracy and adaptability.

[0077] Further, it is determined whether the control instructions of the vehicle-mounted equipment are in accordance with expectations;

[0078] If,it meets the requirements, then the vehicle-mounted device performs the control action;

[0079] If,it does not meet the requirements, the vehicle-mounted device does not perform the control action;

[0080] Among them, based on whether the control instructions of the on-board equipment meet expectations, the judgment result data is fed back to the behavior prediction model, and the behavior prediction model is iteratively updated.

[0081] Specifically, if there are repeated instances of unintended commands due to misoperations, the model will learn the patterns and characteristics of these misoperations, improving its ability to identify similar misoperations in subsequent predictions and reducing the probability of generating unreasonable commands. By continuously incorporating feedback data into model training, the behavior prediction model can continuously adapt to new user behavior patterns, vehicle usage scenarios, and environmental changes, providing strong support for generating more accurate and reasonable control commands.

[0082] The judgment, execution, and model iteration of vehicle-mounted device control commands form a complete closed-loop process. Through this process, the intelligent vehicle system can achieve safe and precise control of equipment, continuously improving its intelligence level through optimization, and providing users with a more reliable and convenient driving experience.

[0083] Further, including:

[0084] Driving behavior and historical data related to driving behavior, including: collecting the driver's driving operations, driving trajectory, vehicle dynamic parameters, operation events of on-board facilities during driving, time, and vehicle environment information during operation;

[0085] Furthermore, it also includes:

[0086] The preprocessing of driving behavior and historical data related to driving behavior includes: integrating the collected data and verifying the integrated data.

[0087] Further, including:

[0088] Generating a behavior prediction model, including: the behavior prediction model is constructed based on a neural network;

[0089] The neural network includes: a long short-term memory network or a gated recurrent unit;

[0090] The neural network includes: the number of neurons, the number of layers and the activation function of the input layer, the hidden layer and the output layer.

[0091] Furthermore, it also includes:

[0092] The input layer receives an input data set;

[0093] The hidden layer consists of multiple units that capture the temporal dependencies in the input dataset;

[0094] The output layer outputs the predicted driving behavior, which includes steering angle, acceleration, and air conditioning switch operation.

[0095] Furthermore, it also includes:

[0096] Assign initial values ​​to the connection weights of the neural network.

[0097] Specifically, using an LSTM network as an example, the number of neurons in the input layer is determined based on the dimensionality of the processed driving behavior data. If the input data contains 10 features, such as vehicle speed, acceleration, steering angle, and accelerator pedal position, then the number of input layer neurons is 10. The hidden layer is configured with multiple LSTM units—for example, two layers, each containing 64 LSTM units—to gradually extract time-dependent features and complex patterns in the data. The number of neurons in the output layer depends on the prediction target. For example, if the goal is to predict three driving behaviors: steering angle, acceleration, and air conditioning on / off (encoded as 0 or 1), the number of output layer neurons is 3. Regarding the activation function, the ReLU function can be used from the input layer to the hidden layer to accelerate network convergence and mitigate the vanishing gradient problem. For continuous value predictions such as steering angle and acceleration, a linear activation function can be used in the output layer. For binary classification problems, such as the air conditioning on / off state, a sigmoid activation function can be used to map the output value to the range 0-1, facilitating the determination of the on / off state.

[0098] Before training a neural network, the weighted values ​​of the network connections must be initialized. Using specific methods such as He initialization and Glorot initialization can ensure a more balanced gradient distribution during the initial training phase, preventing training difficulties caused by excessively large or small gradients. For example, He initialization adjusts the initial weights based on the number of neuron inputs, ensuring consistent output variance across each layer. This helps the network converge faster and improves training efficiency.

[0099] Well-designed weight initialization can significantly accelerate network convergence. While random initialization is simple, due to the randomness of the initial weights, the network may require numerous iterations to find the right parameter combination, resulting in a lengthy training process. However, specific initialization methods provide the network with a starting point closer to the optimal solution. Glorot initialization, based on the principles of information theory, minimizes information loss during forward and backward propagation by ensuring the signal variance remains constant. This allows the network to adjust parameters more quickly toward the optimal solution, shortening training time and improving training efficiency.

[0100] Furthermore, it also includes:

[0101] Calculation loss functions include: mean square error and cross entropy loss;

[0102] Get the backpropagation algorithm to backpropagate the loss function gradient, calculate the gradient and update the model parameters.

[0103] Specifically, the LSTM network is trained using the cleaned training data. During the forward propagation phase, the network receives input driver behavior and vehicle data, such as speed and steering wheel angle at a specific moment. Through layer-by-layer calculations, it ultimately outputs a prediction for the next driving action (such as the predicted steering angle, acceleration change, and air conditioning operation). The prediction is compared with the actual driving behavior data to calculate a loss function. For continuous value predictions (steering angle, acceleration), the mean squared error (MSE) loss function can be used to measure the average squared error between the predicted and actual values. For binary classification problems, such as the air conditioning on / off function, the cross-entropy loss function is used to evaluate the difference between the predicted probability and the actual category.

[0104] The backpropagation algorithm propagates the gradient of the loss function from the output layer back through each layer of the network to calculate the gradient of each parameter. Using gradient descent, the network's weights and bias parameters are updated to gradually reduce the loss function. The process of forward propagation, loss calculation, backpropagation, and parameter updates is repeated until a preset number of iterations (e.g., 1000) is reached, or the loss on the validation set no longer decreases significantly, and performance no longer improves. At this point, the network training is considered converged.

[0105] Evaluate the performance of the trained neural network model on the test set, calculating metrics such as accuracy and recall. For classification tasks in driving behavior prediction (such as air conditioning on / off prediction), accuracy reflects the proportion of correctly predicted samples out of the total number of samples, while recall measures the model's ability to correctly predict positive examples. If model performance does not meet expected performance requirements, adjust the neural network structure (such as increasing the number of hidden layers or adjusting the number of neurons), modify training parameters (such as learning rate and number of iterations), or further optimize the data cleaning process and retrain until the model performance meets the requirements.

[0106] Through the above-mentioned neural network training and prediction model construction process based on historical data cleaning, it is possible to effectively utilize the cleaned high-quality data to build a high-performance driving behavior prediction model, providing strong support for intelligent transportation, autonomous driving and other fields.

[0107] Figure 2 This is a structural diagram of a vehicle-mounted facility control device based on driving behavior prediction provided by one or more embodiments of the present invention.

[0108] like Figure 2 The vehicle-mounted facility control device shown in the figure based on driving behavior prediction includes:

[0109] Acquisition module, data preprocessing module, model training module, model analysis module, instruction generation module, control judgment module and feedback update module;

[0110] A collection module is used to collect driving behavior and historical data related to driving behavior and collect data on current driving behavior;

[0111] A data preprocessing module is used to preprocess the historical data of driving behavior and driving behavior related data to generate an input data set;

[0112] The model training module is used to generate a training data set based on the input data set, and generate a behavior prediction model based on the training data set;

[0113] The model analysis module is used to input the behavior prediction model based on the current driving behavior data to obtain the control intention of the vehicle facilities;

[0114] An instruction generation module is used to generate control instructions for the vehicle-mounted equipment according to the control intention of the vehicle-mounted facilities;

[0115] A control judgment module is used to judge whether the control instructions of the vehicle-mounted equipment meet expectations;

[0116] The feedback update module is used to judge whether the control instructions of the on-board equipment meet expectations, feed back the judgment result data to the behavior prediction model, and iteratively update the behavior prediction model.

[0117] It is worth noting that although the present system only discloses an acquisition module, a data preprocessing module, a model training module, a model analysis module, an instruction generation module, a control judgment module and a feedback update module, it does not mean that the present device is limited to the above-mentioned basic functional modules. Rather, what the present invention wants to express is that, based on the above-mentioned basic functional modules, those skilled in the art can arbitrarily add one or more functional modules in combination with the existing technology to form an infinite number of embodiments or technical solutions. In other words, the present system is open rather than closed. Just because the present embodiment only discloses individual basic functional modules, it cannot be considered that the scope of protection of the claims of the present invention is limited to the above-mentioned basic functional modules.

[0118] Figure 3 A schematic diagram of a vehicle-mounted facility control method based on driving behavior prediction according to a specific embodiment of the present invention.

[0119] In a specific embodiment, a method for controlling vehicle-mounted facilities based on driving behavior prediction is disclosed. The method operates as follows:

[0120] Driving behavior analysis is typically based on multiple data sources, including driver operation, driving trajectory, and vehicle dynamic parameters. For example, users may have a preferred seating angle for their seats; for braking, they may have different braking forces and preferred control positions; and when cornering, they may have preferred angles and settings. This data can be collected through hardware devices such as onboard sensors, positioning sensors, and inertial sensors. Driving behavior is also highly integrated with the environment, and the collected driving behavior also includes external environmental information, including current road conditions, weather, and traffic flow. This information can be uploaded to the cloud along with the vehicle's real-time status through driving behavior analysis or logs.

[0121] Data uploaded to the cloud requires data cleansing and correlation. User driving behavior data collected from various data sources (such as on-board sensors, positioning devices, and mobile apps) is integrated to remove invalid and missing data. Missing values ​​can be filled using methods such as interpolation, mean filling, and mode filling, or records with a large number of missing values ​​can be removed. For example, if the average speed during the morning rush hour from 8:00 AM to 9:00 AM in a certain week is 40 km / h per day, and one day during the morning rush hour is missing, the average speed of 40 km / h can be used to fill the gap. After addressing abnormal data, to facilitate subsequent data analysis and modeling, the data needs to be standardized and normalized so that data of different dimensions can be compared and analyzed on the same scale. After data cleaning, the cleaned data needs to be verified and quality controlled. This can be done through random sampling or comparative analysis to ensure that the cleaning results meet expectations.

[0122] To train a neural network and build a predictive model, first select an appropriate neural network type based on the nature of the problem and the characteristics of the dataset. For driving behavior prediction, recurrent neural networks (RNNs) or their variants, such as long short-term memory (LSTM) or gated recurrent units (GRU), are suitable because they can handle temporal dependencies in sequential data. The neural network architecture should be designed, including the number of neurons, layers, and activation functions in the input, hidden, and output layers. For example, an LSTM network could be designed where the input layer receives processed driving behavior data, the hidden layers contain multiple LSTM units to capture temporal dependencies in the data, and the output layer outputs predicted driving behaviors (such as steering angle, acceleration, and air conditioning on / off operation). Before neural network training begins, each connection (i.e., weight) in the network needs to be assigned an initial value. These initial values ​​are typically randomly generated, but some specific initialization methods (such as He initialization and Glorot initialization) can improve network training efficiency.

[0123] The neural network is then trained using the training set data. During the training process, the network receives input such as driver behavior and vehicle data, and calculates the output through forward propagation to predict the user's next action. This output is then compared with the vehicle's actual goal, that is, the user's actual action at the end, and the loss function (such as mean square error, cross entropy loss, etc.) is calculated to measure the network performance. Next, the gradient of the loss function is backpropagated back to the network using the backpropagation algorithm to calculate the gradient and update the model parameters. The training process is repeated until the preset number of iterations is reached or other stopping conditions are met (such as performance on the validation set no longer improving). The performance of the model is evaluated on the test set to verify the generalization ability of the model, and the values ​​of different evaluation indicators (such as accuracy, recall, etc.) are calculated and compared. If the corresponding indicator requirements are not met, this step can be repeated continuously.

[0124] Deploy the trained model to real-world applications, collect feedback data from actual usage, monitor model performance, and make adjustments and optimizations as needed.

[0125] Figure 4 A block diagram of an electronic device structure of a vehicle-mounted facility control method based on driving behavior prediction provided by one or more embodiments of the present invention.

[0126] like Figure 4 As shown, the present application provides an electronic device, comprising: a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other via the communication bus;

[0127] The memory stores a computer program, which, when executed by the processor, causes the processor to execute the steps of the vehicle-mounted facility control method based on driving behavior prediction.

[0128] The present application also provides a computer-readable storage medium storing a computer program executable by an electronic device. When the computer program runs on the electronic device, the electronic device executes the steps of the vehicle-mounted facility control method based on driving behavior prediction.

[0129] For simplicity of description, the method embodiments are described as a series of actions. However, those skilled in the art should be aware that the embodiments of the present invention are not limited by the order of the actions described, because certain steps can be performed in other orders or simultaneously according to the embodiments of the present invention. Secondly, those skilled in the art should also be aware that the embodiments described in this specification are all preferred embodiments, and the actions involved are not necessarily required by the embodiments of the present invention.

[0130] Through the description of the above embodiments, it can be seen that those skilled in the art can clearly understand that the present application can be implemented by means of software plus the necessary general hardware platform. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a storage medium such as ROM / RAM, a magnetic disk, an optical disk, etc., and includes a number of instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in various embodiments of the present application or certain parts of the embodiments.

[0131] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A vehicle-mounted facility control method based on driving behavior prediction, characterized in that: The method comprises: Collect historical data on driving behavior and driving behavior related data; Preprocessing driving behavior and historical data related to driving behavior to generate input data sets; Generate a training data set based on the input data set; Generate a behavior prediction model based on the training data set; Collect data on current driving behavior; Based on the current driving behavior data, the behavior prediction model is input to obtain the control intention of the vehicle facilities; Generate control instructions for on-board equipment based on the control intention of on-board facilities; Determine whether the control instructions of the on-board equipment meet expectations; If,it meets the requirements, then the vehicle-mounted device performs the control action; If,it does not meet the requirements, the vehicle-mounted device will not perform the control action; Among them, based on whether the control instructions of the on-board equipment meet expectations, the judgment result data is fed back to the behavior prediction model, and the behavior prediction model is iteratively updated.

2. The vehicle-mounted facility control method based on driving behavior prediction according to claim 1, characterized in that: Including; historical data of driving behavior and driving behavior, including: collecting the driver's driving operations, driving trajectory, vehicle dynamic parameters, operation events of on-board facilities during driving, time and vehicle environment information during operation.

3. The vehicle-mounted facility control method based on driving behavior prediction according to claim 1, characterized in that: include: Preprocessing of historical data related to driving behavior and driving behavior; Among them, the collected driving behavior and historical data related to driving behavior are integrated; Verify the integrated data.

4. The vehicle-mounted facility control method based on driving behavior prediction according to claim 2, characterized in that: include: Generating a behavior prediction model, including: the behavior prediction model is constructed based on a neural network; The neural network includes: a long short-term memory network or a gated recurrent unit; The neural network includes: the number of neurons, the number of layers and the activation function of the input layer, the hidden layer and the output layer.

5. The vehicle-mounted facility control method based on driving behavior prediction according to claim 4 is characterized in that: include: The input layer receives an input data set; The hidden layer consists of multiple units that capture the temporal dependencies in the input dataset; The output layer outputs the predicted driving behavior, which includes steering angle, acceleration, and air conditioning switch operation.

6. The vehicle-mounted facility control method based on driving behavior prediction according to claim 5 is characterized in that ; Assign initial values ​​to the connection weights of the neural network.

7. The vehicle-mounted facility control method based on driving behavior prediction according to claim 1 is characterized in that ； Calculation loss functions include: mean square error and cross entropy loss; Get the backpropagation algorithm to backpropagate the loss function gradient, calculate the gradient and update the model parameters.

8. A vehicle-mounted facility control device based on driving behavior prediction, characterized in that: The device includes: an acquisition module, a data preprocessing module, a model training module, a model analysis module, an instruction generation module, a control judgment module and a feedback update module; A collection module is used to collect driving behavior and historical data related to driving behavior and collect data on current driving behavior; A data preprocessing module is used to preprocess the historical data of driving behavior and driving behavior related data to generate an input data set; A model training module is used to generate a training data set based on an input data set, and to generate a behavior prediction model based on the training data set; The model analysis module is used to input the behavior prediction model based on the current driving behavior data to obtain the control intention of the vehicle facilities; An instruction generation module is used to generate control instructions for the vehicle-mounted equipment according to the control intention of the vehicle-mounted facilities; A control judgment module is used to judge whether the control instructions of the vehicle-mounted equipment meet expectations; The feedback update module is used to judge whether the control instructions of the on-board equipment meet expectations, feed back the judgment result data to the behavior prediction model, and iteratively update the behavior prediction model.

9. An electronic device, characterized in that: include: A processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other via the communication bus; The memory stores a computer program, and when the processor executes the computer program, the processor executes the steps of the vehicle-mounted facility control method based on driving behavior prediction according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that It stores a computer program that can be executed by an electronic device. When the computer program runs on the electronic device, the electronic device executes the steps of the vehicle-mounted facility control method based on driving behavior prediction according to any one of claims 1 to 7.

Images