User driving brief report generation method based on AI large model and related equipment

Abstract

The invention discloses a user driving brief report generation method based on an AI large model and related equipment, and relates to the technical field of data processing, and the method comprises the steps: obtaining CAN data of each control module in a vehicle in response to a user driving brief report generation instruction, obtaining multi-dimensional data, and generating a user driving brief report based on the multi-dimensional data; the method comprises the steps of determining corresponding multi-dimensional features of a vehicle in the same time period, processing the multi-dimensional features based on a preset AI large model, generating a target natural language description about user driving behaviors corresponding to the vehicle, and generating a personalized user driving brief report based on the target natural language description and a preset brief report template. According to the application, the CAN data of each control module in the vehicle driving process is collected, the multi-dimensional data is obtained, more accurate and comprehensive multi-dimensional features are further obtained, the preset AI large model is called to process the multi-dimensional features, and a personalized and high-quality user driving brief report is finally obtained in combination with the brief report template.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, and particularly relates to a user driving report generation method based on an AI large model and related equipment. BACKGROUND

[0002] With the continuous improvement of the intelligent level of automobiles, vehicles can collect and generate massive data, covering rich information such as vehicle operating states and driver operation habits. Traditional vehicle data applications mainly focus on vehicle performance monitoring and fault diagnosis, and the analysis of driver driving behavior is mostly limited to simple index statistics, such as average speed and the number of sudden accelerations, which is difficult to provide comprehensive, in-depth and personalized driving analysis.

[0003] At the same time, artificial intelligence technology, especially large language models, is developing rapidly and has shown great capabilities in natural language processing and information generation. However, how to organically combine vehicle data with AI large models to generate high-quality and personalized user driving reports has not been fully researched and applied. In related technologies, the data interaction efficiency between the vehicle data collection system and the AI large model is low, and the data processing accuracy is insufficient, resulting in a single and non-targeted driving report content, which cannot meet the user's demand for personalized driving analysis. Therefore, developing a system and method that can effectively collect vehicle data and generate comprehensive and personalized driving reports with the help of AI large models has important practical significance for improving driving experience, optimizing driving behavior and ensuring driving safety.

[0004] The above content is only used to assist in understanding the technical solutions of the present application and does not represent the acknowledgement of the above content as related technology. SUMMARY

[0005] The main purpose of the present application is to provide a user driving report generation method based on an AI large model and related equipment, aiming to solve the technical problem of how to generate high-quality and personalized user driving reports.

[0006] To achieve the above purpose, the present application provides a user driving report generation method based on an AI large model, which comprises: In response to a user driving report generation instruction, CAN data of each control module in the vehicle is obtained to obtain multi-dimensional data; Based on the multi-dimensional data, the multi-dimensional features corresponding to the vehicle in the same time period are determined, and the multi-dimensional features include driving behavior features and vehicle state features; generate the target natural language description about the user driving behavior corresponding to the vehicle based on the multi-dimensional features, the preset AI large model being obtained based on vehicle data and driving behavior cases after training, the training being used to help the preset AI large model learn the mapping relationship between different driving behavior features and corresponding descriptions; generate the personalized user driving report based on the target natural language description and a preset report template.

[0007] In an embodiment, the step of determining the multi-dimensional features corresponding to the vehicle in the same time period based on the multi-dimensional data further comprises: perform data preprocessing on the multi-dimensional data based on a preset data preprocessing strategy to obtain target data, the data preprocessing including data cleaning, format unification, and time alignment; determine the multi-dimensional features corresponding to the vehicle in the same time period based on the target data.

[0008] In an embodiment, the step of performing data preprocessing on the multi-dimensional data based on a preset data preprocessing strategy to obtain target data further comprises: perform a data anomaly judgment operation on the multi-dimensional data based on a preset data preprocessing strategy to determine whether there are abnormal values in the multi-dimensional data; if there are, correct the abnormal values in the multi-dimensional data to obtain first corrected data; determine whether there are missing values in the first data; if there are, fill in the missing values in the first data to obtain second data; perform a format conversion operation on the second data based on a preset format conversion rule to obtain target data.

[0009] In an embodiment, the step of determining the multi-dimensional features corresponding to the vehicle in the same time period based on the target data further comprises: determine the driving behavior corresponding to the vehicle in the same time period based on the target data, and determine the occurrence number and frequency of the driving behavior; determine the driving behavior features and vehicle state features corresponding to the vehicle based on the driving behavior, the occurrence number, and the frequency, to determine the multi-dimensional features corresponding to the vehicle in the same time period.

[0010] In an embodiment, before the step of generating the target natural language description about the user driving behavior corresponding to the vehicle based on the multi-dimensional features, the preset AI large model being obtained based on vehicle data and driving behavior cases after training, the training being used to help the preset AI large model learn the mapping relationship between different driving behavior features and corresponding descriptions, the method further comprises: generate a driving behavior feature vector and a vehicle state feature vector based on the driving behavior feature and the vehicle state feature in the multi-dimensional features; fuse the driving behavior feature vector and the vehicle state feature vector to obtain an initial feature matrix; determine the corresponding trip information of the vehicle in the same time period to generate a trip identification, the trip identification including a trip ID, a trip start time and a trip end time; add the trip identification to the initial feature matrix to obtain a target feature matrix; In an embodiment, the step of processing the multi-dimensional features based on the preset AI large model to generate the target natural language description about the user driving behavior corresponding to the vehicle further includes: processing the target feature matrix based on the preset AI large model to generate the target natural language description about the user driving behavior corresponding to the vehicle.

[0011] In an embodiment, the step of processing the multi-dimensional features based on the preset AI large model to generate the target natural language description about the user driving behavior corresponding to the vehicle further includes: processing the multi-dimensional features based on the embedding layer of the preset AI large model to obtain a multi-dimensional semantic vector; assigning weights to the multi-dimensional semantic vector based on the attention layer of the preset AI large model and the current weight distribution rule; generating a driving behavior label based on the fully connected layer of the preset AI large model, the multi-dimensional semantic vector and the weights; processing the driving behavior label based on the natural language generation layer of the preset AI large model to obtain a structured description; generating an initial natural language description based on the structured description and determining whether there is a logical conflict between the initial natural language description and the multi-dimensional features; If there is, adjust the current weight distribution rule, and based on the adjusted current weight distribution rule, return to the step of processing the multi-dimensional features based on the embedding layer of the preset AI large model to obtain a multi-dimensional semantic vector until there is no logical conflict between the initial natural language description and the multi-dimensional features, and set the initial natural language description as the target natural language description; if there is not, directly set the initial natural language description as the target natural language description.

[0012] In addition, to achieve the above-mentioned purpose, the present application also provides a user driving report generation device based on an AI large model, which comprises: An acquisition module is configured to acquire CAN data of various control modules in a vehicle to obtain multidimensional data in response to a user driving report generation instruction; A determination module is configured to determine multidimensional features corresponding to the vehicle in a same time period based on the multidimensional data, the multidimensional features including driving behavior features and vehicle state features; A data processing module is configured to process the multidimensional features based on a preset AI large model to generate a target natural language description about user driving behavior corresponding to the vehicle, the preset AI large model being obtained by training based on a large amount of vehicle data and driving behavior cases, the training being used to help the preset AI large model learn a mapping relationship between different driving behavior features and corresponding descriptions; A generation module is configured to generate an individualized user driving report based on the target natural language description and a pre-set report template.

[0013] In an embodiment, the determination module further includes: A data preprocessing unit is configured to perform data preprocessing on the multidimensional data based on a preset data preprocessing strategy to obtain target data, the data preprocessing including data cleaning, format unification, and time alignment; A first determination unit is configured to determine multidimensional features corresponding to the vehicle in a same time period based on the target data.

[0014] In an embodiment, the determination module further includes: A data anomaly judgment unit is configured to perform a data anomaly judgment operation on the multidimensional data based on a preset data preprocessing strategy to determine whether there is an abnormal value in the multidimensional data; A correction unit is configured to correct the abnormal value in the multidimensional data to obtain first data after correction if there is an abnormal value; A finding unit is configured to find whether there is a missing value in the first data; A filling unit is configured to fill the missing value in the first data to obtain second data if there is a missing value; A format conversion unit is configured to perform a format conversion operation on the second data based on a preset format conversion rule to obtain target data.

[0015] In an embodiment, the determination module further includes: A second determination unit is configured to determine driving behavior of the vehicle in a same time period based on the target data, and determine a number of occurrences and a frequency corresponding to the driving behavior; The third determining unit is used to determine the driving behavior characteristics and vehicle state characteristics corresponding to the vehicle based on the driving behavior, the number of occurrences, and the frequency, so as to determine the multi-dimensional characteristics of the vehicle within the same time period.

[0016] In one embodiment, the user driving briefing generation device based on the AI ​​large model further includes a feature fusion module, which further includes: The first generation unit is used to generate a driving behavior feature vector and a vehicle state feature vector based on the driving behavior features and vehicle state features in the multi-dimensional features. The feature vector fusion unit is used to fuse the driving behavior feature vector and the vehicle state feature vector to obtain an initial feature matrix; The fourth determining unit is used to determine the trip information corresponding to the vehicle within the same time period, so as to generate a trip identifier, wherein the trip identifier includes a trip ID, a trip start time, and a trip end time; An adding unit is used to add the trip identifier to the initial feature matrix to obtain the target feature matrix; In one embodiment, the generation module further includes: The first data processing unit is used to process the target feature matrix based on a preset AI model to generate a target natural language description of the user's driving behavior corresponding to the vehicle.

[0017] In one embodiment, the generation module further includes: The second data processing unit is used to process the multi-dimensional features based on the embedding layer of a preset AI large model to obtain a multi-dimensional semantic vector. The weighting unit is used to assign weights to the multidimensional semantic vector based on the attention layer of the preset AI large model and the current weight allocation rules; The second generation unit is used to generate driving behavior labels based on the fully connected layer of the preset AI large model, the multidimensional semantic vector, and the weights. The third data processing unit is used to process the driving behavior labels based on the natural language generation layer of the preset AI large model to obtain a structured description. The third generation unit is used to generate an initial natural language description based on the structured description, and to determine whether there is a logical conflict between the initial natural language description and the multi-dimensional features; The logical judgment unit is used to adjust the current weight allocation rule if it exists, and based on the adjusted current weight allocation rule, return to the embedding layer based on the preset AI large model to process the multi-dimensional features and obtain multi-dimensional semantic vectors, until the initial natural language description and the multi-dimensional features do not have logical conflicts, and set the initial natural language description as the target natural language description; if it does not exist, directly set the initial natural language description as the target natural language description.

[0018] Furthermore, to achieve the above objectives, this application also proposes a user driving briefing generation device based on an AI large model, the device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the user driving briefing generation method based on the AI ​​large model as described above.

[0019] In addition, to achieve the above objectives, this application also proposes a storage medium, which is a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the steps of the user driving briefing generation method based on the AI ​​large model described above.

[0020] In addition, to achieve the above objectives, this application also provides a computer program product, which includes a computer program that, when executed by a processor, implements the steps of the user driving briefing generation method based on the AI ​​large model described above.

[0021] One or more technical solutions proposed in this application have at least the following technical effects: This application proposes a method and related equipment for generating user driving reports based on AI large-scale models, relating to the field of data processing technology. In related technologies, traditional vehicle data applications mainly focus on vehicle performance monitoring and fault diagnosis, with driver behavior analysis often limited to simple statistical indicators such as average speed and number of rapid accelerations, failing to provide comprehensive, in-depth, and personalized driving analysis. Meanwhile, artificial intelligence technology, especially large language models, is developing rapidly, demonstrating powerful capabilities in natural language processing and information generation. However, how to organically combine vehicle data with AI large-scale models to generate high-quality, personalized user driving reports has not yet been fully researched and applied. In related technologies, the data interaction efficiency between vehicle data acquisition systems and AI large-scale models is low, and data processing accuracy is insufficient, resulting in driving reports that are simplistic, lack specificity, and cannot provide comprehensive, in-depth, and personalized driving analysis. Compared to meeting users' needs for personalized driving analysis, this application first, in response to a user driving briefing generation command, acquires CAN data from various control modules in the vehicle to obtain multi-dimensional data. Then, based on the multi-dimensional data, determines the multi-dimensional features corresponding to the vehicle within the same time period. The multi-dimensional features include driving behavior features and vehicle state features. Further, based on a preset AI model, the multi-dimensional features are processed to generate a target natural language description of the user's driving behavior corresponding to the vehicle. The preset AI model is obtained after training based on vehicle data and driving behavior cases. The training is used to help the preset AI model learn the mapping relationship between different driving behavior features and corresponding descriptions. Finally, based on the target natural language description and a pre-set briefing template, a personalized user driving briefing is generated.

[0022] It is understood that this application collects CAN data from various control modules during vehicle operation, that is, obtains multi-dimensional data. Based on the multi-dimensional data, more accurate and comprehensive driving behavior characteristics and vehicle status characteristics are obtained, that is, more accurate and comprehensive multi-dimensional characteristics. Based on the multi-dimensional characteristics, a preset AI large model is called to process the multi-dimensional characteristics, and then combined with a briefing template, a personalized, high-quality user driving briefing is finally obtained. Attached Figure Description

[0023] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0024] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the accompanying drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, those skilled in the art can obtain other drawings based on these drawings without creative effort.

[0025] Figure 1 This is a flowchart illustrating an embodiment of the user driving briefing generation method based on an AI large model provided in this application. Figure 2 This is a flowchart illustrating Embodiment 2 of the user driving briefing generation method based on an AI large model provided in this application; Figure 3 This is a flowchart illustrating Embodiment 3 of the user driving briefing generation method based on an AI large model provided in this application; Figure 4 This is a schematic diagram of the module structure of the user driving briefing generation device based on an AI large model according to an embodiment of this application; Figure 5 This is a schematic diagram of the device structure of the hardware operating environment involved in the user driving briefing generation method based on the AI ​​large model in the embodiments of this application.

[0026] The purpose, features, and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0027] It should be understood that the specific embodiments described herein are merely illustrative of the technical solutions of this application and are not intended to limit this application.

[0028] To better understand the technical solution of this application, a detailed description will be provided below in conjunction with the accompanying drawings and specific implementation methods.

[0029] The main solution in this application's embodiments is: In this embodiment, for ease of description, the following description will focus on the user driving briefing generation device based on an AI large model.

[0030] Due to the limitations of related technologies, traditional vehicle data applications primarily focus on vehicle performance monitoring and fault diagnosis. Analysis of driver behavior often remains at the level of simple statistical indicators, such as average speed and the number of rapid accelerations, failing to provide comprehensive, in-depth, and personalized driving analysis. Meanwhile, artificial intelligence technology, especially large language models, is developing rapidly, demonstrating powerful capabilities in natural language processing and information generation. However, how to organically combine vehicle data with large AI models to generate high-quality, personalized user driving reports has not yet been fully researched and applied. In related technologies, the data interaction efficiency between vehicle data acquisition systems and large AI models is low, and the data processing accuracy is insufficient, resulting in driving reports that are simplistic and lack specificity, failing to meet users' needs for personalized driving analysis.

[0031] This application provides a solution in which: first, in response to a user driving briefing generation command, CAN data from various control modules in the vehicle is acquired to obtain multi-dimensional data; then, based on the multi-dimensional data, multi-dimensional features corresponding to the vehicle within the same time period are determined, including driving behavior features and vehicle state features; further, based on a preset AI model, the multi-dimensional features are processed to generate a target natural language description of the user's driving behavior corresponding to the vehicle, wherein the preset AI model is obtained after training based on vehicle data and driving behavior cases, and the training is used to help the preset AI model learn the mapping relationship between different driving behavior features and corresponding descriptions; finally, based on the target natural language description and a pre-set briefing template, a personalized user driving briefing is generated.

[0032] It is understood that this application collects CAN data from various control modules during vehicle operation, that is, obtains multi-dimensional data. Based on the multi-dimensional data, more accurate and comprehensive driving behavior characteristics and vehicle status characteristics are obtained, that is, more accurate and comprehensive multi-dimensional characteristics. Based on the multi-dimensional characteristics, a preset AI large model is called to process the multi-dimensional characteristics, and then combined with a briefing template, a personalized, high-quality user driving briefing is finally obtained.

[0033] It should be noted that the executing entity in this embodiment can be a computing service device with data processing, network communication, and program execution functions, such as a tablet computer, personal computer, or mobile phone, or an electronic device capable of performing the above functions, such as a user driving briefing generation device based on an AI large model. The following description uses a user driving briefing generation device based on an AI large model as an example to illustrate this embodiment and the subsequent embodiments.

[0034] Based on this, embodiments of this application provide a method for generating user driving briefings based on an AI large model, referring to... Figure 1 , Figure 1 This is a flowchart illustrating the first embodiment of the user driving briefing generation method based on an AI large model in this application.

[0035] In this embodiment, the user driving briefing generation method based on AI large model includes steps S100~S400: Step S100: In response to the user's driving briefing generation command, obtain CAN data from various control modules in the vehicle to obtain multi-dimensional data; It should be noted that "responding to the user driving briefing generation command" is equivalent to a triggering mechanism. Only after the user driving briefing generation command is detected will the user driving briefing generation device based on the AI ​​big model execute the subsequent steps.

[0036] It should be noted that the user's command to generate a driving report can be initiated by the user or automatically by the device. For example, the user can click the "Generate Driving Report" button through the vehicle's infotainment system (central control screen) or mobile app, and the system will receive a clear command signal; or the system can be automatically triggered based on preset conditions, such as after each trip (when the vehicle is turned off), at a fixed time each week (e.g., every Monday at 8:00 AM), when the mileage reaches a certain threshold (e.g., every 500 kilometers driven), or when the vehicle starts up, connects to the network successfully, and has sufficient battery power (e.g., SOC > 20%).

[0037] In this application, user commands are sent to the vehicle's main control unit (such as a T-Box or gateway module) via the vehicle's infotainment system or mobile app. Upon receiving the command, the main control unit sends data acquisition requests to each control module.

[0038] It should be noted that CAN (Controller Area Network) bus is a high-speed serial communication protocol used for internal vehicle communication and is widely used in automotive electronic systems. It allows for fast and reliable data exchange between various control modules in a vehicle (such as the engine control unit ECU, body control module BCM, powertrain module, etc.).

[0039] It should be noted that, in this embodiment, data such as engine speed, torque, fuel injection quantity, and throttle opening are acquired based on the engine control unit (ECU); data such as door status, light status, and air conditioning power are acquired based on the body control module (BCM); data such as vehicle speed, acceleration, and braking pressure are acquired based on the powertrain module; data such as battery voltage, current, SOC (state of charge), and cell voltage (for new energy vehicles) are acquired based on the battery management system (BMS); data such as wheel speed, yaw rate, and braking pressure are acquired based on the ABS / ESP module; and data such as vehicle latitude and longitude, altitude, speed, and heading angle are acquired based on the navigation and positioning module (GNSS).

[0040] It should be noted that in this application, firstly, the main control unit (such as a T-Box) sends data request commands to each module. Then, each module responds to the request and sends its currently stored operating data to the main control unit via the CAN bus. Further, the main control unit temporarily stores the received data in a local cache or storage device, preparing for subsequent processing.

[0041] It should be noted that the collected data comes from different units / modules / systems, and the types and dimensions of the collected data are also different; that is, the collected data is multi-dimensional. The multi-dimensional characteristics of the data include: time dimension (data is collected in time series, recording the vehicle's operating status at different points in time), spatial dimension (through the GNSS module, recording the vehicle's operating status at different geographical locations), functional dimension (covering multiple functional systems of the vehicle, such as the power system, braking system, body system, etc.), and physical quantity dimension (including various physical quantities such as speed, acceleration, rotational speed, voltage, current, and temperature).

[0042] It should also be noted that in this embodiment, the collected raw data is stored in the form of CAN messages, including information such as signal name, signal value, and timestamp. The raw CAN messages are parsed into specific physical quantities (such as speed in km / h and voltage in V) using DBC (Database Container) files or ARXML files.

[0043] Understandably, the core of step S100 is to trigger data acquisition and obtain multi-dimensional CAN data from various control modules of the vehicle. This provides fundamental data support for subsequent data preprocessing, feature extraction, and AI model inference. Through a reasonable triggering mechanism and data acquisition process, high-quality, multi-dimensional vehicle operation data can be obtained, laying the foundation for generating personalized driving reports.

[0044] Step S200: Based on the multi-dimensional data, determine the multi-dimensional features of the vehicle within the same time period, wherein the multi-dimensional features include driving behavior features and vehicle state features; It should be noted that before feature extraction, the collected multi-dimensional data needs to be divided into multiple time periods. These time periods can be fixed or dynamic, depending on the application scenario and analysis requirements. For example, a time period can be divided into 30 seconds or minutes; or a time period can be divided into the period from the vehicle's cold start to its shutdown; or a time period can be divided into a continuous range of speeds greater than 5 km / h; or a time period can be divided into a time period with a specific event (such as rapid acceleration or sudden braking) as the start and end point.

[0045] It should be noted that feature extraction involves extracting meaningful metrics from raw data that reflect driving behavior and vehicle status. Feature extraction is divided into driving behavior features and vehicle status features.

[0046] It should be noted that driving behavior characteristics mainly reflect the driver's operating habits and driving style, including but not limited to: the number of times of rapid acceleration, the number of times of rapid deceleration, the number of times of speeding, the number of times of frequent lane changes, the average accelerator depth, and the average braking depth.

[0047] It should be noted that vehicle status characteristics mainly reflect the vehicle's operating status and performance, including but not limited to: average fuel consumption / electricity consumption, maximum single-cell voltage difference of the battery, average engine load rate, vehicle mileage, and braking energy recovery ratio.

[0048] It should be noted that in this application, feature values ​​are extracted through feature calculation. The selected method for feature calculation is: Time series-based computation: For dynamic data (such as velocity, acceleration), calculate its statistics (such as mean, variance, maximum, minimum).

[0049] Event-based computation: For a specific event (such as rapid acceleration or sudden braking), calculate the number of times and frequency of its occurrence.

[0050] Physical model-based calculations: For features such as energy consumption, calculations are performed using physical models (e.g., fuel consumption = fuel consumption / mileage).

[0051] Understandably, feature extraction can transform complex, multi-dimensional data into concise and meaningful metrics that reflect driving behavior and vehicle status. These features will then serve as input for subsequent AI model inference, generating personalized driving reports.

[0052] Step S300: Based on the preset AI big model, the multi-dimensional features are processed to generate a target natural language description of the user's driving behavior corresponding to the vehicle. The preset AI big model is obtained after training based on vehicle data and driving behavior cases. The training is used to help the preset AI big model learn the mapping relationship between different driving behavior features and corresponding descriptions. It should be noted that the preset AI large model is a specially trained deep learning model whose core function is to convert multi-dimensional features into natural language descriptions. This model is trained based on the following: Vehicle data: This includes vehicle operating status data (such as fuel consumption, vehicle speed, etc.).

[0053] Driving behavior case studies: including sample data of different driving behaviors (such as rapid acceleration, sudden braking, etc.).

[0054] Natural language description: Natural language description corresponding to the driving behavior case (e.g., "During this trip, there were many instances of rapid acceleration").

[0055] Understandably, the purpose of training is to enable large AI models to learn the mapping relationship between different driving behavior characteristics and corresponding natural language descriptions.

[0056] It's important to note that the pre-trained AI model is typically a large, pre-trained language model (such as LLaMA, GPT, ChatGLM, etc.). These models are pre-trained on large-scale text data and possess powerful natural language generation capabilities. Based on this pre-trained model, fine-tuning is performed using vehicle data and driving behavior examples. During fine-tuning, the model learns how to convert input feature vectors into specific natural language descriptions.

[0057] It should be noted that the training data includes a large number of feature vectors and corresponding natural language description pairs.

[0058] Specifically, before the step of processing the multi-dimensional features based on a preset AI large model to generate a target natural language description of the user's driving behavior corresponding to the vehicle, steps S210~S240 are further included: Step S210: Based on the driving behavior features and vehicle state features in the multi-dimensional features, generate driving behavior feature vector and vehicle state feature vector; It should be noted that the driving behavior feature vector is a numerical representation extracted from driving behavior features, used to describe the driver's operating habits and driving style. For example: rapid acceleration count (the number of times acceleration exceeds a certain threshold (e.g., 3 m / s²) within a specific time period); rapid deceleration count (the number of times deceleration falls below a certain threshold (e.g., -3 m / s²) within a specific time period); speeding count (the number of times vehicle speed exceeds a certain threshold (e.g., 120 km / h) within a specific time period); frequent lane changes count (the number of times the steering wheel angle change rate exceeds a certain threshold (e.g., 15° / s) within a specific time period); and average throttle depth (the average throttle opening within a specific time period).

[0059] It should be noted that the vehicle state feature vector is a numerical representation extracted from the vehicle state features, used to describe the vehicle's operating state and performance. Examples include: average fuel consumption / electricity consumption (energy consumption per unit distance, such as L / 100km or kWh / 100km within a specific time period); maximum single-cell voltage difference in the battery (the maximum difference in voltage between individual battery cells within a specific time period); average engine load rate (the average engine load rate within a specific time period); vehicle mileage (the distance the vehicle travels within a specific time period); and regenerative braking ratio (the proportion of regenerative braking to total energy consumption within a specific time period).

[0060] Step S220: Fuse the driving behavior feature vector and the vehicle state feature vector to obtain an initial feature matrix; In this embodiment, the driving behavior feature vector and the vehicle state feature vector are merged to form an initial feature matrix. The purpose of this step is to integrate different types of features together for subsequent processing.

[0061] Step S230: Determine the trip information corresponding to the vehicle within the same time period to generate a trip identifier, wherein the trip identifier includes trip ID, trip start time and trip end time; It should be noted that trip information is used to identify the vehicle's travel status within a specific time period, including trip ID, trip start time, and trip end time. Trip ID is a unique identifier for each trip. Trip start time refers to the timestamp indicating the start of the trip. Trip end time refers to the timestamp indicating the end of the trip.

[0062] Step S240: Add the trip identifier to the initial feature matrix to obtain the target feature matrix; In this embodiment, trip identifiers are added to the initial feature matrix to form the target feature matrix. The purpose of this step is to associate trip information with feature data, facilitating subsequent analysis.

[0063] Specifically, the step of processing the multi-dimensional features based on a preset AI large model to generate a target natural language description of the user's driving behavior corresponding to the vehicle further includes step S310: Step S310: Based on a preset AI large model, process the target feature matrix to generate a target natural language description of the user's driving behavior corresponding to the vehicle.

[0064] It should be noted that in step S310, the input to the AI ​​large model is the target feature matrix, which integrates driving behavior features, vehicle state features, and trip information. This matrix contains rich information and can comprehensively reflect the vehicle's operating status and driving behavior within a specific time period.

[0065] In this embodiment, firstly, the numerical features in the target feature matrix are converted into a format that the model can understand (such as a token ID). Then, trip information (such as trip ID, start time, and end time) is used as contextual information to help the model better understand the background of the input data. Further, the model generates a natural language description based on the input feature matrix and the contextual information.

[0066] Understandably, this step allows the system to convert complex numerical features into easily understandable textual information, providing users with intuitive driving behavior analysis and improvement suggestions.

[0067] Step S400: Based on the target natural language description and the pre-set briefing template, generate a personalized user driving briefing.

[0068] It should be noted that a presentation template is a pre-defined framework used to integrate natural language descriptions and other relevant information into a complete report. Presentation templates typically include: a title, body text, charts, etc.

[0069] In this embodiment, firstly, the natural language description generated by the AI ​​model is filled into the corresponding position of the briefing template. Then, other relevant information (such as fuel consumption, mileage, etc.) is added as needed. Furthermore, intuitive charts (such as fuel consumption curves, mileage charts, etc.) are generated based on the data. Finally, a complete personalized driving briefing is generated, which can be a PDF file, a webpage, or other formats.

[0070] This application proposes a method and related equipment for generating user driving reports based on AI large-scale models, relating to the field of data processing technology. In related technologies, traditional vehicle data applications mainly focus on vehicle performance monitoring and fault diagnosis, with driver behavior analysis often limited to simple statistical indicators such as average speed and number of rapid accelerations, failing to provide comprehensive, in-depth, and personalized driving analysis. Meanwhile, artificial intelligence technology, especially large language models, is developing rapidly, demonstrating powerful capabilities in natural language processing and information generation. However, how to organically combine vehicle data with AI large-scale models to generate high-quality, personalized user driving reports has not yet been fully researched and applied. In related technologies, the data interaction efficiency between vehicle data acquisition systems and AI large-scale models is low, and data processing accuracy is insufficient, resulting in driving reports that are simplistic, lack specificity, and cannot provide comprehensive, in-depth, and personalized driving analysis. Compared to meeting users' needs for personalized driving analysis, this application first, in response to a user driving briefing generation command, acquires CAN data from various control modules in the vehicle to obtain multi-dimensional data. Then, based on the multi-dimensional data, determines the multi-dimensional features corresponding to the vehicle within the same time period. The multi-dimensional features include driving behavior features and vehicle state features. Further, based on a preset AI model, the multi-dimensional features are processed to generate a target natural language description of the user's driving behavior corresponding to the vehicle. The preset AI model is obtained after training based on vehicle data and driving behavior cases. The training is used to help the preset AI model learn the mapping relationship between different driving behavior features and corresponding descriptions. Finally, based on the target natural language description and a pre-set briefing template, a personalized user driving briefing is generated.

[0071] It is understood that this application collects CAN data from various control modules during vehicle operation, that is, obtains multi-dimensional data. Based on the multi-dimensional data, more accurate and comprehensive driving behavior characteristics and vehicle status characteristics are obtained, that is, more accurate and comprehensive multi-dimensional characteristics. Based on the multi-dimensional characteristics, a preset AI large model is called to process the multi-dimensional characteristics, and then combined with a briefing template, a personalized, high-quality user driving briefing is finally obtained.

[0072] Based on the first embodiment of this application, in the second embodiment of this application, the content that is the same as or similar to that in Embodiment 1 above can be referred to the above description, and will not be repeated hereafter. Based on this, please refer to... Figure 2 The step of determining the multi-dimensional features of the vehicle within the same time period based on the multi-dimensional data further includes steps A100 to A200: Step A100: Based on a preset data preprocessing strategy, perform data preprocessing on the multi-dimensional data to obtain target data. The data preprocessing includes data cleaning, format unification, and time alignment. It's important to note that data preprocessing is the first step in data processing. Its purpose is to transform raw, multi-dimensional data into clean, uniform, and time-aligned data for subsequent feature extraction and analysis. The preprocessing process includes data cleaning, format standardization, and time alignment.

[0073] It should be noted that the purpose of data cleaning is to remove or correct outliers, noise, and erroneous data to ensure data quality.

[0074] It should be noted that the purpose of format standardization is to convert data from different sources into a consistent format, facilitating subsequent processing. Specific operations include: It should be noted that the purpose of time alignment is to ensure that the timestamps of all data are consistent, which facilitates subsequent analysis.

[0075] Understandably, after data cleaning, format standardization, and time alignment, the resulting target data is clean, uniform, and time-aligned multi-dimensional data that can be directly used for subsequent feature extraction.

[0076] Specifically, the step of preprocessing the multi-dimensional data based on a preset data preprocessing strategy to obtain the target data further includes steps A110 to A150: Step A110: Based on a preset data preprocessing strategy, perform a data anomaly detection operation on the multi-dimensional data to determine whether there are any outliers in the multi-dimensional data. It should be noted that outliers refer to data points that deviate significantly from the normal range, which may be caused by sensor malfunction, data transmission errors, or abnormal operation.

[0077] In this embodiment, the determination method may be: Set thresholds: Set a reasonable threshold range for each feature.

[0078] Statistical methods: Use statistical methods (such as mean ± 3 standard deviations) to identify outliers.

[0079] Machine learning methods: Use anomaly detection algorithms (such as Isolation Forest and DBSCAN) to identify outliers.

[0080] Step A120: If outliers exist, correct them in the multi-dimensional data to obtain the corrected first data. It should be noted that by correcting the identified outliers, the impact of data noise on subsequent analysis can be reduced.

[0081] It should be noted that the correction methods include: replacing outliers with reasonable values, such as the average, median, or the value from the previous time point; or, if outliers cannot be corrected, deleting the record is an option.

[0082] Step A130: Check if there are any missing values ​​in the first data; It should be noted that missing values ​​refer to data points that are not recorded or are missing.

[0083] The method for finding missing values ​​is as follows: Check for null values: Check for null or NaN values ​​in each feature column.

[0084] Check time continuity: Check if the timestamps are continuous. If the timestamps are not continuous, it may indicate that data is lost.

[0085] Step A140: If the missing values ​​exist, fill in the missing values ​​in the first data to obtain the second data; It is understandable that missing values ​​are filled in to ensure data integrity.

[0086] It should be noted that the filling method can be: Linear interpolation: Performing linear interpolation based on the preceding and following data points.

[0087] Spline interpolation: Uses spline interpolation to fill in missing values.

[0088] Mean imputation: Fill in missing values ​​with the mean of the feature.

[0089] Forward / backward filling: Fill missing values ​​with values ​​from the previous or next time step.

[0090] Step A150: Based on preset format conversion rules, perform a format conversion operation on the second data to obtain the target data.

[0091] In this embodiment, the second data is converted into a uniform format for subsequent processing. For example, the data is converted to a uniform unit; or data type conversion is performed to ensure data consistency (converting binary data to decimal); or all timestamps are converted to a uniform time format (such as UTC time).

[0092] Understandably, these steps ensure data quality and consistency, providing a reliable foundation for subsequent feature extraction and analysis.

[0093] Step A200: Based on the target data, determine the multi-dimensional features of the vehicle within the same time period.

[0094] It should be noted that the purpose of feature extraction is to extract multi-dimensional features from the target data that can reflect the vehicle's state and driving behavior. These features will serve as the basis for subsequent analysis and model input.

[0095] Understandably, steps A100 and A200 together ensure the quality and availability of the data, providing a solid foundation for subsequent driving behavior analysis and natural language description generation.

[0096] Specifically, the step of determining the multi-dimensional features of the vehicle within the same time period based on the target data further includes steps A210 to A220: Step A210: Based on the target data, determine the driving behavior of the vehicle within the same time period, and determine the number of times and frequency of occurrence of the driving behavior; It should be noted that in this embodiment, specific driving behaviors are identified from the target data, and the frequency and number of occurrences of these behaviors are calculated. The purpose of this step is to quantify driving behaviors, providing a foundation for subsequent feature extraction.

[0097] It should be noted that the number of occurrences refers to the number of times a certain driving behavior occurs within a specific time period.

[0098] It should be noted that frequency refers to the ratio of the number of times a certain driving behavior occurs to the total driving time or total driving mileage.

[0099] Step A220: Based on the driving behavior, the number of occurrences, and the frequency, determine the driving behavior characteristics and vehicle state characteristics corresponding to the vehicle, so as to determine the multi-dimensional characteristics of the vehicle within the same time period.

[0100] In this embodiment, based on driving behavior and its frequency and occurrence, driving behavior features and vehicle state features are further extracted to form multi-dimensional features. The purpose of this step is to quantify driving behavior and vehicle state into specific feature values ​​for subsequent analysis.

[0101] It should be noted that, in this embodiment, the characteristics can be determined by calculating the average fuel consumption, the highest single-cell voltage difference of the battery, and the average engine load rate.

[0102] Understandably, steps A210 and A220 together ensure that comprehensive and meaningful multi-dimensional features are extracted from the target data, providing a solid foundation for subsequent driving behavior analysis and natural language description generation.

[0103] Based on the first and second embodiments of this application, in the second embodiment of this application, the content that is the same as or similar to that in Embodiment 1 above can be referred to the above description, and will not be repeated hereafter. Based on this, please refer to... Figure 3 The step of processing the multi-dimensional features based on a preset AI large model to generate a target natural language description of the user's driving behavior corresponding to the vehicle further includes steps B100 to B600: Step B100: Based on the embedding layer of the preset AI large model, the multi-dimensional features are processed to obtain a multi-dimensional semantic vector; In this embodiment, firstly, the multi-dimensional features are converted into multi-dimensional semantic vectors that the model can process. The role of the embedding layer is to map the input feature vectors into a high-dimensional space, enabling them to better represent the semantic information of the features.

[0104] Specifically, in this embodiment, multi-dimensional features (such as the number of rapid accelerations, average fuel consumption, etc.) are converted into high-dimensional vectors through an embedding layer. Using a pre-trained embedding matrix, each feature is mapped to a vector space of fixed dimensions.

[0105] Step B200: Based on the attention layer of the preset AI large model and the current weight allocation rules, assign weights to the multidimensional semantic vector; In this embodiment, each feature in the multidimensional semantic vector is assigned a weight to highlight important features and suppress unimportant features. The attention mechanism helps the model focus on the information most useful for generating the description.

[0106] In this embodiment, a preset attention layer and the current weight allocation rule are used to calculate weights for each feature vector. Weights are dynamically assigned based on the importance of the features. For example, the number of rapid accelerations may be more important than the mileage traveled, and therefore receive a higher weight.

[0107] Step B300: Based on the fully connected layer of the preset AI large model, the multidimensional semantic vector, and the weights, generate driving behavior labels; In this embodiment, labels for driving behaviors are generated based on weighted multidimensional semantic vectors. These labels are categorical representations of driving behaviors and are used for subsequent natural language generation. That is, a pre-defined fully connected layer is used to map the weighted multidimensional semantic vectors to driving behavior labels.

[0108] Step B400: Based on the natural language generation layer of the preset AI large model, the driving behavior labels are processed to obtain a structured description; In this embodiment, driving behavior labels are converted into structured natural language descriptions. This step transforms the labels into more detailed text information, providing a foundation for the final natural language description. Specifically, a pre-defined natural language generation layer is used to convert driving behavior labels into structured descriptions.

[0109] Step B500: Based on the structured description, generate an initial natural language description and determine whether there is a logical conflict between the initial natural language description and the multi-dimensional features; In this embodiment, an initial natural language description is generated based on the structured description, and a logical conflict is detected between the description and the original multi-dimensional features. That is, the structured description is converted into a more natural language description.

[0110] It should be noted that logical conflict detection refers to checking whether the generated description is consistent with the original features. For example, if the description mentions "high fuel consumption," but the average fuel consumption in the original features is below the normal range, then there is a logical conflict.

[0111] Step B600: If it exists, adjust the current weight allocation rule, and based on the adjusted current weight allocation rule, return to the embedding layer based on the preset AI large model to process the multi-dimensional features and obtain multi-dimensional semantic vectors, until the initial natural language description and the multi-dimensional features do not have any logical conflict, and set the initial natural language description as the target natural language description; if it does not exist, directly set the initial natural language description as the target natural language description.

[0112] In this embodiment, if a logical conflict exists, the weight allocation rules are adjusted, and the process restarts from the embedding layer until the generated description matches the original features.

[0113] Understandably, through these steps, the system is able to generate accurate and natural descriptions of driving behavior, ensuring that the descriptions are consistent with the logic of the original data.

[0114] It should be noted that the above examples are only for understanding this application and do not constitute a limitation on the user driving briefing generation method based on the AI ​​large model of this application. Any simple modifications based on this technical concept are within the protection scope of this application.

[0115] It should be noted that all user-related data involved in this application (such as user attribute data, user behavior data, and user geographical location, etc., the data types here should be modified according to the adaptability of the solution content) were obtained with the user's permission or consent; that is to say, when this application is applied to specific products or technologies, user permission is required to obtain and process the relevant data, and the processing of the relevant data must comply with the relevant laws, regulations and regulatory standards of the relevant countries and regions.

[0116] For example, when it is necessary to obtain a user's current geographical location, a location acquisition prompt can be displayed on the user's terminal. After receiving confirmation from the user regarding the location acquisition prompt, the terminal can obtain the user's current geographical location.

[0117] This application also provides a user driving briefing generation device based on an AI large model; please refer to [reference needed]. Figure 4 The user driving briefing generation device based on the AI ​​large model includes: The acquisition module 10 is used to acquire CAN data of various control modules in the vehicle in response to the user's driving briefing generation command, and obtain multi-dimensional data. The determining module 20 is used to determine the multi-dimensional features of the vehicle within the same time period based on the multi-dimensional data. The multi-dimensional features include driving behavior features and vehicle state features. Data processing module 30 is used to process the multi-dimensional features based on a preset AI big model to generate a target natural language description of the user's driving behavior corresponding to the vehicle. The preset AI big model is obtained after training on a large amount of vehicle data and driving behavior cases. The training is used to help the preset AI big model learn the mapping relationship between different driving behavior features and corresponding descriptions. The generation module 40 is used to generate personalized user driving briefings based on the target natural language description and a pre-set briefing template.

[0118] In one embodiment, the determining module further includes: The data preprocessing unit is used to perform data preprocessing on the multi-dimensional data based on a preset data preprocessing strategy to obtain target data. The data preprocessing includes data cleaning, format unification, and time alignment. The first determining unit is used to determine the multi-dimensional features of the vehicle within the same time period based on the target data.

[0119] In one embodiment, the determining module further includes: The data anomaly detection unit is used to perform a data anomaly detection operation on the multi-dimensional data based on a preset data preprocessing strategy, so as to determine whether there are any abnormal values ​​in the multi-dimensional data. A correction unit is used to correct outliers in the multi-dimensional data if they exist, to obtain corrected first data. The search unit is used to search for whether there are missing values ​​in the first data; The filling unit is used to fill in the missing values ​​in the first data if they exist, so as to obtain the second data; The format conversion unit is used to perform a format conversion operation on the second data based on a preset format conversion rule to obtain the target data.

[0120] In one embodiment, the determining module further includes: The second determining unit is used to determine the driving behavior of the vehicle within the same time period based on the target data, and to determine the number of times and frequency of occurrence of the driving behavior. The third determining unit is used to determine the driving behavior characteristics and vehicle state characteristics corresponding to the vehicle based on the driving behavior, the number of occurrences, and the frequency, so as to determine the multi-dimensional characteristics of the vehicle within the same time period.

[0121] In one embodiment, the user driving briefing generation device based on the AI ​​large model further includes a feature fusion module, which further includes: The first generation unit is used to generate a driving behavior feature vector and a vehicle state feature vector based on the driving behavior features and vehicle state features in the multi-dimensional features. The feature vector fusion unit is used to fuse the driving behavior feature vector and the vehicle state feature vector to obtain an initial feature matrix; The fourth determining unit is used to determine the trip information corresponding to the vehicle within the same time period, so as to generate a trip identifier, wherein the trip identifier includes a trip ID, a trip start time, and a trip end time; An adding unit is used to add the trip identifier to the initial feature matrix to obtain the target feature matrix; In one embodiment, the generation module further includes: The first data processing unit is used to process the target feature matrix based on a preset AI model to generate a target natural language description of the user's driving behavior corresponding to the vehicle.

[0122] In one embodiment, the generation module further includes: The second data processing unit is used to process the multi-dimensional features based on the embedding layer of a preset AI large model to obtain a multi-dimensional semantic vector. The weighting unit is used to assign weights to the multidimensional semantic vector based on the attention layer of the preset AI large model and the current weight allocation rules; The second generation unit is used to generate driving behavior labels based on the fully connected layer of the preset AI large model, the multidimensional semantic vector, and the weights. The third data processing unit is used to process the driving behavior labels based on the natural language generation layer of the preset AI large model to obtain a structured description. The third generation unit is used to generate an initial natural language description based on the structured description, and to determine whether there is a logical conflict between the initial natural language description and the multi-dimensional features; The logical judgment unit is used to adjust the current weight allocation rule if it exists, and based on the adjusted current weight allocation rule, return to the embedding layer based on the preset AI large model to process the multi-dimensional features and obtain multi-dimensional semantic vectors, until the initial natural language description and the multi-dimensional features do not have logical conflicts, and set the initial natural language description as the target natural language description; if it does not exist, directly set the initial natural language description as the target natural language description.

[0123] The user driving briefing generation device based on an AI large model provided in this application adopts the user driving briefing generation method based on an AI large model in the above embodiments, and can solve the technical problem of user driving briefing generation based on an AI large model. Compared with related technologies, the beneficial effects of the user driving briefing generation device based on an AI large model provided in this application are the same as the beneficial effects of the user driving briefing generation method based on an AI large model provided in the above embodiments, and other technical features in the user driving briefing generation device based on an AI large model are the same as the features disclosed in the methods of the above embodiments, and will not be repeated here.

[0124] This application provides a user driving briefing generation device based on an AI large model. The user driving briefing generation device based on an AI large model includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the user driving briefing generation method based on the AI ​​large model in the first embodiment described above.

[0125] The following is for reference. Figure 5This document illustrates a structural schematic diagram of a user driving briefing generation device based on an AI large model, suitable for implementing embodiments of this application. The user driving briefing generation device based on an AI large model in this application can include, but is not limited to, mobile terminals such as mobile phones, laptops, digital radio receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Description), PMPs (Portable Media Players), in-vehicle terminals (e.g., in-vehicle navigation terminals), and fixed terminals such as digital TVs and desktop computers. Figure 5 The user driving briefing generation device based on the AI ​​large model shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this application.

[0126] like Figure 5 As shown, the AI-based large-scale user driving briefing generation device may include a processing unit 1001 (e.g., a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to a program stored in read-only memory (ROM) 1002 or a program loaded from storage device 1003 into random access memory (RAM) 1004. The RAM 1004 also stores various programs and data required for the operation of the AI-based large-scale user driving briefing generation device. The processing unit 1001, ROM 1002, and RAM 1004 are interconnected via a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to I / O interface 1006: input devices 1007 including, for example, touchscreens, touchpads, keyboards, mice, image sensors, microphones, accelerometers, gyroscopes, etc.; output devices 1008 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 1003 including, for example, magnetic tapes, hard disks, etc.; and communication devices 1009. Communication device 1009 allows the AI-based large-scale user driving briefing generation device to wirelessly or wiredly communicate with other devices to exchange data. Although the figure shows an AI-based large-scale user driving briefing generation device with various systems, it should be understood that it is not required to implement or possess all the systems shown. More or fewer systems can be implemented alternatively.

[0127] Specifically, according to the embodiments disclosed in this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments disclosed in this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device, or installed from storage device 1003, or installed from ROM 1002. When the computer program is executed by processing device 1001, it performs the functions defined in the methods of the embodiments disclosed in this application.

[0128] The user driving briefing generation device based on an AI large model provided in this application, employing the user driving briefing generation method based on an AI large model in the above embodiments, can solve the technical problems. Compared with related technologies, the beneficial effects of the user driving briefing generation device based on an AI large model provided in this application are the same as those of the user driving briefing generation method based on an AI large model provided in the above embodiments, and other technical features in this user driving briefing generation device based on an AI large model are the same as those disclosed in the method of the previous embodiment, and will not be repeated here.

[0129] It should be understood that the various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0130] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0131] This application provides a computer-readable storage medium having computer-readable program instructions (i.e., a computer program) stored thereon, the computer-readable program instructions being used to execute the user driving briefing generation method based on the AI ​​large model in the above embodiments.

[0132] The computer-readable storage medium provided in this application may be, for example, a USB flash drive, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems or devices, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system or device. The program code contained on the computer-readable storage medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination thereof.

[0133] The aforementioned computer-readable storage medium may be included in the AI-based large-scale user driving briefing generation device; or it may exist independently and not be assembled into the AI-based large-scale user driving briefing generation device.

[0134] The aforementioned computer-readable storage medium carries one or more programs that, when executed by the AI-based large-scale model-based user driving briefing generation device, cause the AI-based large-scale model-based user driving briefing generation device to: In response to the user's command to generate a driving briefing, CAN data from various control modules in the vehicle is obtained to acquire multi-dimensional data; Based on the multi-dimensional data, the multi-dimensional features corresponding to the vehicle in the same time period are determined, and the multi-dimensional features include driving behavior features and vehicle state features. Based on a pre-set AI model, the multi-dimensional features are processed to generate a target natural language description of the user's driving behavior corresponding to the vehicle. The pre-set AI model is obtained after training based on vehicle data and driving behavior cases. The training is used to help the pre-set AI model learn the mapping relationship between different driving behavior features and corresponding descriptions. Based on the target natural language description and the pre-defined briefing template, a personalized user driving briefing is generated.

[0135] Computer program code for performing the operations of this application can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a Local Area Network (LAN) or a Wide Area Network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0136] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0137] The modules described in the embodiments of this application can be implemented in software or hardware. The names of the modules do not necessarily limit the functionality of the unit itself.

[0138] The readable storage medium provided in this application is a computer-readable storage medium that stores computer-readable program instructions (i.e., a computer program) for executing the above-described AI-based large-scale user driving briefing generation method, thereby solving the technical problem of generating user driving briefings based on AI large-scale models. Compared with related technologies, the beneficial effects of the computer-readable storage medium provided in this application are the same as those of the AI-based large-scale user driving briefing generation method provided in the above embodiments, and will not be repeated here.

[0139] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the above-described user driving briefing generation method based on an AI large model.

[0140] The computer program product provided in this application can solve the technical problem of generating user driving briefings based on AI large models. Compared with related technologies, the beneficial effects of the computer program product provided in this application are the same as those of the user driving briefing generation method based on AI large models provided in the above embodiments, and will not be repeated here.

[0141] The above description is only a part of the embodiments of this application and does not limit the scope of protection of this application. All equivalent structural transformations made under the technical concept of this application and using the content of this application specification and drawings, or direct / indirect applications in other related technical fields, are included in the scope of protection of this application.

Claims

1. A method for generating user driving briefings based on an AI large model, characterized in that, The method for generating user driving briefings based on AI large models includes: In response to the user's command to generate a driving briefing, CAN data from various control modules in the vehicle is obtained to acquire multi-dimensional data; Based on the multi-dimensional data, the multi-dimensional features corresponding to the vehicle in the same time period are determined, and the multi-dimensional features include driving behavior features and vehicle state features. Based on a pre-set AI model, the multi-dimensional features are processed to generate a target natural language description of the user's driving behavior corresponding to the vehicle. The pre-set AI model is obtained after training based on vehicle data and driving behavior cases. The training is used to help the pre-set AI model learn the mapping relationship between different driving behavior features and corresponding descriptions. Based on the target natural language description and the pre-defined briefing template, a personalized user driving briefing is generated.

2. The user driving briefing generation method based on AI large model as described in claim 1, characterized in that, The step of determining the multi-dimensional features of the vehicle within the same time period based on the multi-dimensional data further includes: Based on a preset data preprocessing strategy, the multi-dimensional data is preprocessed to obtain the target data. The data preprocessing includes data cleaning, format unification, and time alignment. Based on the target data, the multi-dimensional features corresponding to the vehicle within the same time period are determined.

3. The user driving briefing generation method based on an AI large model as described in claim 2, characterized in that, The step of preprocessing the multi-dimensional data based on a preset data preprocessing strategy to obtain the target data further includes: Based on a preset data preprocessing strategy, a data anomaly detection operation is performed on the multi-dimensional data to determine whether there are any outliers in the multi-dimensional data. If any outliers exist, the outliers in the multi-dimensional data are corrected to obtain the corrected first data. Check if there are any missing values ​​in the first data; If missing values ​​exist, fill in the missing values ​​in the first data to obtain the second data; Based on preset format conversion rules, a format conversion operation is performed on the second data to obtain the target data.

4. The user driving briefing generation method based on AI large model as described in claim 2, characterized in that, The step of determining the multi-dimensional features of the vehicle within the same time period based on the target data further includes: Based on the target data, determine the driving behavior of the vehicle within the same time period, and determine the number of times and frequency of occurrence of the driving behavior; Based on the driving behavior, the number of occurrences, and the frequency, the driving behavior characteristics and vehicle state characteristics corresponding to the vehicle are determined, so as to determine the multi-dimensional characteristics of the vehicle within the same time period.

5. The user driving briefing generation method based on AI large model as described in claim 1, characterized in that, Before the step of processing the multi-dimensional features based on a preset AI model to generate a target natural language description of the user's driving behavior corresponding to the vehicle, the method further includes: Based on the driving behavior features and vehicle state features in the multi-dimensional features, a driving behavior feature vector and a vehicle state feature vector are generated. The driving behavior feature vector and the vehicle state feature vector are fused to obtain an initial feature matrix; Determine the trip information corresponding to the vehicle within the same time period to generate a trip identifier, which includes a trip ID, a trip start time, and a trip end time; The trip identifier is added to the initial feature matrix to obtain the target feature matrix; The step of processing the multi-dimensional features based on a preset AI model to generate a target natural language description of the user's driving behavior corresponding to the vehicle further includes: Based on a pre-set AI model, the target feature matrix is ​​processed to generate a target natural language description of the user's driving behavior corresponding to the vehicle.

6. The user driving briefing generation method based on AI large model as described in claim 1, characterized in that, The step of processing the multi-dimensional features based on a preset AI model to generate a target natural language description of the user's driving behavior corresponding to the vehicle further includes: Based on the embedding layer of the preset AI large model, the multi-dimensional features are processed to obtain a multi-dimensional semantic vector; Based on the attention layer of the preset AI large model and the current weight allocation rules, weights are assigned to the multidimensional semantic vector; Based on the fully connected layer of the preset AI model, the multidimensional semantic vector, and the weights, driving behavior labels are generated. The natural language generation layer based on the preset AI large model processes the driving behavior labels to obtain a structured description; Based on the structured description, an initial natural language description is generated, and it is determined whether there is a logical conflict between the initial natural language description and the multi-dimensional features; If such a rule exists, adjust the current weight allocation rule. Based on the adjusted current weight allocation rule, return to the embedding layer based on the preset AI large model to process the multi-dimensional features and obtain multi-dimensional semantic vectors. Continue this process until the initial natural language description and the multi-dimensional features do not have any logical conflict. Then, set the initial natural language description as the target natural language description. If such a rule does not exist, directly set the initial natural language description as the target natural language description.

7. A user driving briefing generation device based on an AI large model, characterized in that, The user driving briefing generation device based on the AI ​​large model includes: The acquisition module is used to acquire CAN data from various control modules in the vehicle in response to the user's driving briefing generation command, and obtain multi-dimensional data. The determining module is used to determine the multi-dimensional features of the vehicle within the same time period based on the multi-dimensional data. The multi-dimensional features include driving behavior features and vehicle state features. The data processing module is used to process the multi-dimensional features based on a preset AI model to generate a target natural language description of the user's driving behavior corresponding to the vehicle. The preset AI model is obtained by training a large amount of vehicle data and driving behavior cases. The training is used to help the preset AI model learn the mapping relationship between different driving behavior features and corresponding descriptions. A generation module is used to generate personalized user driving briefings based on the target natural language description and a pre-set briefing template.

8. A user driving briefing generation device based on an AI large model, characterized in that, The device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the user driving briefing generation method based on any one of claims 1 to 6.

9. A storage medium, characterized in that, The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, it implements the steps of the user driving briefing generation method based on an AI large model as described in any one of claims 1 to 6.

10. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, implements the steps of the user driving briefing generation method based on an AI large model as described in any one of claims 1 to 6.

Images