Scene self-generation method, device and system based on data driving and medium

Abstract

The invention relates to the technical field of vehicles, in particular to a scene self-generation method, device and system based on data driving and a medium, and the method comprises the steps: obtaining first data, and forming a first wide table based on the first data; performing data cleaning and data merging on the first wide table and the second data to form a second wide table; forming a first sequence table based on the first wide table and the second wide table, performing data mining on the first sequence table by adopting an association rule algorithm, and forming a result table based on a data mining result; filtering and screening the scene based on the result table, pre-judging a corresponding scene application according to the screened scene, and recommending the scene application to the user terminal; according to the invention, the personalized scene mode conforming to the habits of the user can be automatically generated, and the intelligence and personalization of the intelligent cabin are improved.

Description

Technical Field

[0001] This invention relates to the field of vehicle technology, and specifically to a data-driven method, apparatus, system, and medium for self-generation of scenes. Background Technology

[0002] With the continuous development of intelligent connected technology, the car cabin will become a digital intelligent terminal and a third living space that meets the multi-sensory experience needs of drivers and passengers, and will show a trend of interconnection.

[0003] Currently, smart cockpits are being used in various scenarios such as driving, social interaction, office work, and entertainment.

[0004] Currently, smart cockpits offer two main types of scenario modes for drivers and passengers. One type involves pre-setting scenario modes on the in-vehicle platform for users to choose from. In this mode, users can only select from pre-set modes and cannot freely adjust them, resulting in a rigid functionality. The other type offers a custom scenario mode function on the in-vehicle platform, allowing users to freely combine and edit them. However, creating custom scenarios is complex, and users cannot effectively utilize the customization function.

[0005] Current smart cockpits cannot automatically generate scene modes based on user habits to reduce repetitive operations each time the user uses the vehicle; or they cannot proactively generate corresponding modes based on certain special scenarios without requiring users to perform complex operations. Summary of the Invention

[0006] In view of this, the purpose of the embodiments of the present invention is to provide a data-driven scene self-generation method, apparatus, system and medium to solve one or more technical problems existing in the prior art and provide at least one beneficial option or create conditions.

[0007] On one hand, embodiments of the present invention provide a data-driven scene self-generation method, the method comprising the following steps: Acquire first data and form a first wide table based on the first data; the first data includes at least one of environmental feature data and vehicle state signal features, and the first wide table includes feature type, feature name and corresponding feature value; The second data is obtained, and the first wide table and the second data are cleaned and merged to form a second wide table; the second data is user behavior tracking data. A first sequence table is formed based on the first wide table and the second wide table. Data mining is performed on the first sequence table using an association rule algorithm, and a result table is formed based on the data mining results. The first sequence table is a sequence table corresponding to user behavior and features. The data mining methods include finding behavior based on features, finding features based on behavior, and finding behavior based on scenarios. The system filters and selects scenarios based on the results table, predicts the corresponding scenario applications based on the selected scenarios, and recommends the scenario applications to the user terminal.

[0008] Optionally, forming a first wide table based on the first data includes: The first data is preprocessed to identify the feature type, feature name and corresponding feature value. The identified feature type, feature name and corresponding feature value are organized into a first wide table according to a preset format. Among them, the environmental feature data includes at least one of weather conditions, in-vehicle temperature, outside vehicle temperature and air humidity, and the vehicle status signal features include at least one of vehicle speed, engine speed, fuel level and battery level.

[0009] Optionally, the step of cleaning and merging the data in the first wide table and the second data to form the second wide table includes: Remove duplicate, erroneous, and incomplete data from the first wide table and the second data. Then, merge the cleaned first wide table and the second data according to preset rules. During the merging process, ensure that the correspondence between user behavior tracking data and environmental feature data or vehicle status signal feature data is accurate, forming a second wide table that contains user behavior tracking data and the corresponding environmental feature or vehicle status signal feature.

[0010] Optionally, the step of using an association rule algorithm to perform data mining on the first sequence list and forming a result table based on the data mining results includes: The association rule algorithm is used to calculate the association index between user behavior and features in the first sequence list; the association index includes at least one of lift, mean of two confidence scores, and imbalance ratio; wherein, lift is the probability of a user behavior occurring given that a certain feature has occurred compared to the probability of the user behavior occurring without any conditions, mean of two confidence scores is the average of the confidence score of a user behavior occurring when a certain feature has occurred and the confidence score of a certain feature occurring when a certain user behavior occurs, and imbalance ratio is an indicator that measures the difference in sample size between a certain feature and a certain user behavior; Based on the aforementioned correlation indicators, results showing a strong correlation between user behavior and features are selected and a results table is generated. The strong correlation is determined by a comprehensive evaluation of the lift, the mean of the two confidence levels, and the imbalance ratio. When the lift is greater than a preset threshold, the mean of the two confidence levels is greater than a preset threshold, and the imbalance ratio is less than a preset threshold, it is determined that there is a strong correlation between user behavior and features.

[0011] Optionally, the step of filtering and selecting scenarios based on the result table, predicting corresponding scenario applications based on the selected scenarios, and recommending the scenario application to the user terminal includes: The result table is filtered according to preset filtering conditions. Based on the current vehicle environment feature data and user behavior data, the scene mode in the filtered result table is matched. The filtering conditions include feature value range limitation and / or user behavior type limitation. When a scenario pattern matching the preset conditions is found, the corresponding scenario application is predicted and recommended to the user terminal; wherein, the preset conditions include feature combination conditions and / or user behavior combination conditions; the scenario application includes at least one of vehicle behavior operation action execution and multimedia usage recommendation.

[0012] Optionally, the method further includes: Given user information, determine the user's intended action under a specific environmental condition based on the result table; Personalized scenarios are generated based on the user's historical behavior data and preferences; When the vehicle's environmental feature data meets the triggering conditions for the personalized scenario, the corresponding operation for that personalized scenario is executed.

[0013] Optionally, the method further includes: The scenario examples are constructed and improved based on the result table; When a user frequently performs a certain action in a certain scenario, add that action to the default action of that scenario; Alternatively, when a user frequently performs multiple operations simultaneously in a certain scenario, these operations can be combined into a new scenario pattern and added to the scenario library.

[0014] On the other hand, embodiments of the present invention provide a data-driven scene self-generation device, comprising: The first module is used to acquire first data and form a first wide table based on the first data; the first data includes at least one of environmental feature data and vehicle state signal features, and the first wide table includes feature type, feature name and corresponding feature value; The second module is used to acquire the second data, clean and merge the data of the first wide table and the second data to form the second wide table; the second data is user behavior tracking data. The third module is used to form a first sequence list based on the first wide table and the second wide table, perform data mining on the first sequence list using an association rule algorithm, and form a result table based on the data mining results; the first sequence list is a sequence list corresponding to user behavior and features, and the data mining methods include feature-based behavior finding, behavior-based feature finding, and scenario-based behavior finding. The fourth module is used to filter and select scenarios based on the result table, predict the corresponding scenario applications based on the selected scenarios, and recommend the scenario applications to the user terminal.

[0015] On the other hand, embodiments of the present invention provide a data-driven scene self-generation system, including: At least one processor; At least one memory for storing at least one program; When the at least one program is executed by the at least one processor, the at least one processor performs the method described above.

[0016] On the other hand, embodiments of the present invention provide a computer-readable storage medium storing a processor-executable program, which, when executed by a processor, is used to perform the above-described method.

[0017] The embodiments of this invention include the following beneficial effects: This invention provides a data-driven method, apparatus, system, and medium for self-generation of scenes. By acquiring and integrating environmental feature data, vehicle status signal features, and user behavior tracking data, this invention utilizes association rule algorithms to deeply mine the potential correlations between user behavior and features, enabling the automatic generation of personalized scene patterns that conform to user habits. This invention not only solves the problems of fixed and rigid scene patterns and complex customization operations in existing smart cockpits, but also achieves automatic scene generation based on user habits, reducing repetitive operations for users each time they use the vehicle and improving the user experience. Furthermore, this invention can proactively generate corresponding patterns based on certain special scenarios, eliminating the need for complex user operations, further enhancing the practicality and convenience of the smart cockpit. Attached Figure Description

[0018] Figure 1 This is a flowchart illustrating the steps of a data-driven scene self-generation method provided in an embodiment of the present invention. Figure 2 This is a structural block diagram of a data-driven scene self-generation device provided in an embodiment of the present invention. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0020] It should be noted that although the device diagram shows a modular division and the flowchart illustrates a logical order, in some cases, the steps shown or described may be performed in a different order than the modular division in the device or the order shown in the flowchart. The terms "first," "second," etc., used in the specification, claims, and the aforementioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence.

[0021] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein is for the purpose of describing embodiments of the invention only and is not intended to limit the invention.

[0022] Furthermore, the described features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. Numerous specific details are provided in the following description to give a full understanding of embodiments of the invention. However, those skilled in the art will recognize that the technical solutions of the invention can be practiced without one or more of the specific details, or other methods, components, apparatuses, steps, etc., can be employed. In other instances, well-known methods, apparatuses, implementations, or operations are not shown or described in detail to avoid obscuring various aspects of the invention.

[0023] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor devices and / or microcontroller devices.

[0024] The flowcharts shown in the accompanying drawings are merely illustrative and do not necessarily include all content and operations / steps, nor do they necessarily have to be performed in the described order. For example, some operations / steps can be broken down, while others can be combined or partially combined; therefore, the actual execution order may change depending on the specific circumstances.

[0025] like Figure 1 As shown, Figure 1 A data-driven scene self-generation method is provided in this embodiment of the invention, the method comprising the following steps: S100, acquire first data, and form a first wide table based on the first data; the first data includes at least one of environmental feature data and vehicle state signal features, and the first wide table includes feature type, feature name and corresponding feature value; S200, Obtain the second data, perform data cleaning and data merging on the first wide table and the second data to form a second wide table; the second data is user behavior tracking data. S300, a first sequence table is formed based on the first wide table and the second wide table, and data mining is performed on the first sequence table using an association rule algorithm, and a result table is formed based on the data mining results; the first sequence table is a sequence table corresponding to user behavior and features, and the data mining methods include feature-based behavior finding, behavior-based feature finding, and scenario-based behavior finding; S400 filters and selects scenarios based on the result table, predicts the corresponding scenario applications based on the selected scenarios, and recommends the scenario applications to the user terminal.

[0026] It should be noted that this invention provides a data-driven scene self-generation method to enhance the intelligence and personalization of smart cockpits. To achieve this, data such as vehicle status signals and environmental characteristics are crucial for improving the intelligence and personalization of smart cockpits. However, the following problems exist in reality: 1. High data acquisition costs; products and operations rely on BI to obtain data for product optimization, a time-consuming and inefficient process. 2. Lack of accurate recommendation timing identification capabilities, making effective personalized recommendations difficult to achieve. 3. Difficulty in identifying the interrelationships and influences of user behaviors, failing to uncover clear behavioral relationships.

[0027] Therefore, the first problem that this invention needs to solve is: to be able to automatically mine and obtain effective data, so as to improve the acquisition rate of effective data, improve the efficiency of acquiring effective data, and accurately identify the time of personalized recommendations; The second problem solved by this invention: accurately identifying the timing of personalized recommendations; The third problem solved by this invention is to improve recommendation accuracy by associating outbound behavior with relationships.

[0028] In steps S100 to S400 of this embodiment of the invention, first data is acquired, and environmental feature data and vehicle state signal features are integrated into a first wide table, providing a structured data foundation for subsequent analysis. User behavior tracking data is introduced as second data, and after data cleaning and merging, a second wide table containing user behavior and corresponding environmental or vehicle state features is formed. This ensures the integrity and accuracy of the data, laying a solid foundation for subsequent data mining. The first sequence table is deeply mined using association rule algorithms, and strong correlations between user behavior and features are accurately identified by calculating association indicators. Based on the mining results table, scenarios that match the current vehicle environment characteristics and user behavior patterns can be intelligently filtered and selected, thereby predicting and recommending corresponding scenario applications to the user terminal. This achieves full-chain automation from data acquisition and processing to scenario recommendation, significantly improving the intelligence level and user experience of the smart cockpit.

[0029] In some embodiments, forming a first wide table based on the first data includes: The first data is preprocessed to identify the feature type, feature name and corresponding feature value. The identified feature type, feature name and corresponding feature value are organized into a first wide table according to a preset format. Among them, the environmental feature data includes at least one of weather conditions, in-vehicle temperature, outside vehicle temperature and air humidity, and the vehicle status signal features include at least one of vehicle speed, engine speed, fuel level and battery level.

[0030] In this embodiment, by performing detailed preprocessing operations on the first data, the specific types, names, and corresponding values ​​of various features can be accurately identified. Organizing this identified information according to a preset format forms a structured first wide table, providing a clear and orderly data foundation for subsequent data analysis and processing.

[0031] In some embodiments, the step of cleaning and merging the data of the first wide table and the second data to form the second wide table includes: Remove duplicate, erroneous, and incomplete data from the first wide table and the second data. Then, merge the cleaned first wide table and the second data according to preset rules. During the merging process, ensure that the correspondence between user behavior tracking data and environmental feature data or vehicle status signal feature data is accurate, forming a second wide table that contains user behavior tracking data and the corresponding environmental feature or vehicle status signal feature.

[0032] In this embodiment, the data cleaning process effectively removes duplicate, erroneous, and incomplete data from the first wide table and the second data set, ensuring data accuracy and reliability. Subsequently, the cleaned data is merged according to preset rules. This process pays particular attention to the correspondence between user behavior tracking data and environmental feature data or vehicle status signal feature data, thus forming a second wide table containing user behavior and its corresponding environmental or vehicle status features. The second wide table has a clear structure and complete data, providing strong support for subsequent data analysis and mining.

[0033] In some embodiments, the step of performing data mining on the first sequence list using an association rule algorithm and forming a result table based on the data mining results includes: The association rule algorithm is used to calculate the association index between user behavior and features in the first sequence list; the association index includes at least one of lift, mean of two confidence scores, and imbalance ratio; wherein, lift is the probability of a user behavior occurring given that a certain feature has occurred compared to the probability of the user behavior occurring without any conditions, mean of two confidence scores is the average of the confidence score of a user behavior occurring when a certain feature has occurred and the confidence score of a certain feature occurring when a certain user behavior occurs, and imbalance ratio is an indicator that measures the difference in sample size between a certain feature and a certain user behavior; Based on the aforementioned correlation indicators, results showing a strong correlation between user behavior and features are selected and a results table is generated. The strong correlation is determined by a comprehensive evaluation of the lift, the mean of the two confidence levels, and the imbalance ratio. When the lift is greater than a preset threshold, the mean of the two confidence levels is greater than a preset threshold, and the imbalance ratio is less than a preset threshold, it is determined that there is a strong correlation between user behavior and features.

[0034] In this embodiment, by utilizing association rule algorithms, various association metrics between user behaviors and features in the first sequence list can be accurately calculated. These metrics include lift, the mean of two confidence levels, and the imbalance ratio, providing an effective tool for quantifying the strength of the association between user behaviors and features. Specifically, lift reflects the increase in the probability of a user behavior occurring under specific feature conditions relative to the unconditional state; the mean of the two confidence levels comprehensively considers the confidence of a user behavior occurring when the feature occurs and the confidence of a feature occurring when the user behavior occurs; and the imbalance ratio measures the difference in sample size between features and user behaviors, thus helping us identify truly strongly correlated combinations. Based on the comprehensive evaluation of these association metrics, strongly correlated user behavior and feature combinations can be screened and compiled into a results table. This process not only improves the accuracy of data mining but also provides strong data support for subsequent scenario recommendations. By setting reasonable preset thresholds, it is ensured that only truly strongly correlated user behavior and feature combinations are included in the results table, thereby further improving the accuracy and effectiveness of scenario recommendations.

[0035] In some embodiments, the filtering and selection of scenarios based on the result table, the prediction of corresponding scenario applications based on the selected scenarios, and the recommendation of the scenario applications to the user terminal include: The result table is filtered according to preset filtering conditions. Based on the current vehicle environment feature data and user behavior data, the scene mode in the filtered result table is matched. The filtering conditions include feature value range limitation and / or user behavior type limitation. When a scenario pattern matching the preset conditions is found, the corresponding scenario application is predicted and recommended to the user terminal; wherein, the preset conditions include feature combination conditions and / or user behavior combination conditions; the scenario application includes at least one of vehicle behavior operation action execution and multimedia usage recommendation.

[0036] Specifically, the result table is filtered according to preset filtering conditions, and scenarios that meet the preset conditions are selected from the filtered result table according to the filtering rules. The filtering conditions include feature value range limitation, user behavior type limitation, and the intersection of the two. The filtering rules include rules based on user behavior frequency, feature occurrence frequency, or a combination of the two. Feature value range limitation is used to filter scenarios where the feature value is within a specific range, user behavior type limitation is used to filter scenarios corresponding to specific types of user behavior, and the intersection of the two is used to filter by considering both feature value and user behavior type at the same time. Based on the correlation between user behavior and features in the results table, scenarios that meet preset conditions are selected; the preset conditions include feature combination conditions, user behavior combination conditions, and the intersection of features and user behavior.

[0037] Based on the selected scenarios, combined with current vehicle environment feature data and user behavior tracking data, the scenario applications that users may need are predicted. The prediction process includes real-time analysis of current environment features and user behavior, and matching them with the selected scenarios. When the matching degree reaches a preset threshold, the corresponding scenario application is determined. In some embodiments, the method further includes: Given user information, determine the user's intended action under a specific environmental condition based on the result table; Personalized scenarios are generated based on the user's historical behavior data and preferences; When the vehicle's environmental feature data meets the triggering conditions for the personalized scenario, the corresponding operation for that personalized scenario is executed.

[0038] It should be further noted that in various specific embodiments of this application, when processing is required based on user information related to the object's identity or characteristics, such as the object's (user's) information, behavioral data, historical data, and location information, the object's permission or consent will be obtained first. Furthermore, the collection, use, and processing of this data will comply with relevant laws, regulations, and standards of the relevant countries and regions. In addition, when embodiments of this application require access to sensitive object information, separate permission or consent from the object will be obtained through pop-ups or redirection to a confirmation page. Only after obtaining the object's separate permission or consent will the necessary object-related data for the proper functioning of the embodiments of this application be acquired.

[0039] In this embodiment, permission to access user information is first obtained. After acquiring the user information, the system can combine the results table to conduct in-depth analysis of the user's potential intended actions under specific environmental conditions. This process not only considers the characteristics of the user's current environment but also integrates the user's historical behavioral data and preference information, thereby generating highly personalized scenario recommendations.

[0040] In some embodiments, the method further includes: The scenario examples are constructed and improved based on the result table; When a user frequently performs a certain action in a certain scenario, add that action to the default action of that scenario; Alternatively, when a user frequently performs multiple operations simultaneously in a certain scenario, these operations can be combined into a new scenario pattern and added to the scenario library.

[0041] In this embodiment, through in-depth analysis of the result table and continuous tracking of user behavior patterns, the system can dynamically construct and improve scenario examples. Specifically, when it detects that a user frequently performs a certain operation in a specific scenario, the system automatically sets that operation as the default action for that scenario, thus automatically triggering it when the user re-enters the scenario, improving the convenience of the user experience. Furthermore, if it is found that a user frequently performs multiple operations simultaneously in a certain scenario, the system intelligently combines these operations into a new scenario pattern and includes it in the scenario library, so as to provide more accurate and comprehensive recommendations for similar scenarios in the future. This dynamic adjustment and optimization mechanism not only enriches the content of the scenario library but also makes scenario recommendations more closely aligned with the user's actual needs and habits.

[0042] The following is a general description of the working process of a data-driven scene self-generation method provided by the present invention.

[0043] like Figure 1 The diagram shown is a flowchart of a scene self-generation method. The following section, in conjunction with the attached diagram, illustrates this method. Figure 1 The method is explained in detail.

[0044] S01: Obtain first data, the first data including environmental features and vehicle status signal features, the vehicle status signal features including features triggered by user behavior operations and features triggered by non-user behavior operations. In this step, after the user issues a start command to the vehicle, the cloud obtains the first data through the vehicle or the in-vehicle system. This first data includes raw data with environmental characteristics and vehicle status signal characteristics. The environmental characteristics include, but are not limited to, weather, air quality, humidity, interior temperature, and exterior temperature. If the vehicle is equipped with a weather sensor, it can monitor the current environmental temperature, humidity, wind speed, etc., in real time, or obtain real-time weather data from external weather services through the in-vehicle system. If the vehicle is equipped with an air quality sensor, it obtains interior air quality data or exterior air quality data through integrated environmental sensors, such as volatile organic compounds, carbon dioxide concentration, and particulate matter. The vehicle's humidity sensor can obtain the moisture content in the air inside or outside the vehicle, or exterior humidity data can be obtained from weather services through the in-vehicle system. Temperature sensors on the vehicle obtain temperature data from different parts of the vehicle, such as the driver's seat, passenger seat, area behind the front bumper, and exterior rearview mirrors.

[0045] In this step, vehicle status signal characteristics include, but are not limited to: initial vehicle status data monitored when the user starts the vehicle, vehicle status data triggered by a user's action during the vehicle's journey, or vehicle status data generated during the user's use of the vehicle. Examples include: window status, air conditioning status, air conditioning internal / external circulation status, driver's seatbelt, passenger's seatbelt, front and rear wiper status, headlight status, battery level, fuel consumption, etc.

[0046] SO2: A first wide table is formed based on the first data; In this step, a first wide table is formed based on the first data. The first wide table includes feature types, feature names, and corresponding feature values. The first wide table includes, but is not limited to, a wide table containing environmental features and vehicle state signal features. An example of a first wide table is shown in Table 1.

[0047] Table 1: First Wide Table

[0048] SO3: Obtain second data, which includes business data points other than device control, multimedia, and vehicle status signal characteristics or device characteristics. In this step, the second data acquired is cabin service tracking data. Based on user behavior data, cabin service tracking is designed to construct more valuable features for the first data, providing high-quality and timely data support for the first wide table. The service tracking includes, but is not limited to, device control, such as windows, air conditioning, seats, wipers, rearview mirrors, fragrance, and lights; it also includes multimedia, such as navigation, music, video, radio, and games. The service tracking also includes some instantaneous scene features. Although these instantaneous scenes are not high-probability events, they may be features appearing in personalized scenarios, such as the monitored state of a person, specifically identifying user fatigue; or the triggering of certain scenario modes, specifically a short nap within a scenario mode.

[0049] It should be noted that in the solution of the present invention, there is no order in which the first data and the second data are acquired. The first data can be acquired first and then the second data, or the second data can be acquired first and then the first data, or the first data and the second data can be acquired simultaneously, or the first data and the second data can be acquired intermittently. The solution does not limit the order, form, or source of the first data and the second data.

[0050] S04: Clean and merge the first data and the second data; In this step, the cloud cleans and merges the first and second data to facilitate data analysis and modeling, thereby improving the acquisition rate of effective data.

[0051] S05: A second wide table is formed based on the cleaning and merging of the first and second data; In this step, based on business needs, new features are constructed after cleaning and merging to form a second wide table, ensuring the consistency and integrity of data logic. An example of the second wide table is shown in Table 1.

[0052] Table 2: Second Wide Table

[0053] S06: A first sequence list is formed based on the first wide table and the second wide table. The first sequence list is a user behavior and feature correspondence sequence list that exists in the form of a list. In this step, a first sequence list is obtained based on the above steps, thereby recording the correspondence between various user behaviors and their characteristics, helping to analyze user behavior patterns, preferences and basic characteristics, and thus providing a foundation for in-depth analysis and decision-making systems.

[0054] S07: Data mining is performed on the first sequence list using an association rule algorithm. The data mining methods include, but are not limited to, feature-based behavior finding, behavior-based feature finding, and scene-based behavior finding, and a result table is formed based on the data mining. In this invention, multiple association metrics can be defined to measure an association rule, such as lift, the mean of two confidence levels, imbalance ratio, the probability of B (second data) occurring given A (first data), and the probability of A occurring given B. Specific implementation schemes of this aspect are as follows: (1) Lift: Lift is the ratio of two probabilities. lift{A→B} describes the probability of B occurring given that A has occurred, divided by the probability of B occurring without any conditions. Formula: lift{A→B} = confidence of {A→B} / P(B) = P(B|A) / P(B) = P(A∩B) / (P(A)*P(B)). Lift reflects the correlation between A and B in the association rule. A lift > 1 and the higher the lift, the higher the positive correlation. A lift < 1 and the lower the lift, the higher the negative correlation. A lift = 1 indicates no correlation (the lift value is affected by zero transactions). In simple terms: The lift of {A→B} reflects the promoting / inhibiting effect of A on the probability of B occurring compared to the probability of B occurring under natural conditions. For example: Lift('In-car temperature < 9 degrees' -----> 'Turn on seat heating') = 5, which means that the user is more likely to turn on the seat heating when the in-car temperature is less than 9 degrees. For example, lift('Turn on seat heating'----->'Turn on seat ventilation')=0.6 means that if a user turns on the seat heating, there is a high probability that they will not turn on the seat ventilation.

[0055] (2) kulc: kulc is the mean of two confidence levels, i.e., KULC(A->B)=0.5*(P(B|A)+P(A|B)). It measures the strength of the association between upstream condition A and downstream result B. The KULC value is between 0 and 1, and the larger the value, the better.

[0056] (3) Imbalance ratio (IR): measures the difference in sample size between upstream condition A and downstream result B.

[0057] IR(A->B)=|sup(A)-sup(B)| / (sup(A)+sup(B)-sup(A∩B)). The IR imbalance ratio has a value between 0 and 1, and the smaller the value, the better.

[0058] (4) pro_adb: The probability of B occurring under condition A. pro_adb('Indoor temperature < 9 degrees' -----> 'Turn on seat heating') = 0.2 means that the probability of 'turning on seat heating' is 0.2 when 'indoor temperature < 9 degrees'.

[0059] (5) pro_bda: The probability of A occurring under condition B. pro_bda('Indoor temperature < 9 degrees' -----> 'Turn on seat heating') = 0.4 The probability of 'Indoor temperature < 9 degrees' when 'turn on seat heating' is 0.4.

[0060] S08: Filtering and selection scenarios based on result tables; S09: Recommend scenario applications to users based on scenario prediction.

[0061] The processing of the first and second data based on the association rule algorithm has the following beneficial effects: 1. It can mine and obtain effective data to improve the acquisition rate and efficiency of effective data, and accurately identify the timing of personalized recommendations. 2. It can accurately identify the timing of personalized recommendations. 3. It can correlate outgoing traffic to improve recommendation accuracy.

[0062] In steps S07, S08, and S09, the association rule algorithm is used to mine the relationships between items in the dataset of the first sequence table to form a result table, thereby discovering the implicit relationships between items in the data in order to better understand user behavior and the impact of the environment on user behavior. Finally, applications are proactively recommended to the user. The recommended application may be to recommend the execution of a certain vehicle behavior operation, or it may be to recommend the use of a certain multimedia.

[0063] Overall, the data-driven scenario self-generation method, applied to vehicle systems, helps automakers and service providers better understand users' driving habits, usage preferences, and consumption behaviors, thereby providing them with more personalized products and services. For example, by analyzing users' driving habits, automakers can provide driving modes and assistance functions that better meet user needs, thus improving driving safety and comfort; by analyzing users' usage preferences, service providers can provide in-vehicle entertainment and information services that better meet user needs, thus improving user satisfaction and loyalty. Furthermore, it can help automakers and service providers better understand users' consumption behaviors.

[0064] Specifically, taking one embodiment as an example to understand the feature-based behavior approach, the method based on the present invention can improve the acquisition rate of effective data.

[0065] Obtain first data, which is environmental characteristic data, including: the weather is sunny, the temperature inside the vehicle is 10 degrees Celsius, ...; Obtain the second data, which is user behavior tracking data, such as: closing windows, turning on air conditioning, seat heating, listening to music, ambient lighting, maps, etc.; The environmental feature data and business tracking data are processed, including a sequence list formed based on the first data and the second data, cleaning and merging based on the first data and the second data; The processed data is mined using an association rule algorithm; Final results table: Car interior temperature: Cold --→ Turn on the air conditioner, Cold --→ Seat heating; Based on this method, when the interior temperature is predicted to be cold, the system can proactively recommend to the user that they turn on the air conditioning or the seat heating.

[0066] During vehicle use, the environmental characteristics data that can be obtained are diverse and complex, and the user's operational behaviors during vehicle use are also diverse. However, by using the method of this invention to clean and merge the obtained data, and then using an association rule algorithm to obtain a filtered and selected dataset, i.e., a result table, we can determine the operations that most users may need under a certain environmental condition, and thus proactively recommend operation options to users.

[0067] Furthermore, given the user information, this method can determine the user's desired action under specific environmental conditions, thereby generating a personalized scenario that automatically assists the user in performing the action without requiring any user intervention. For example, based on this method, if it is known that a user's habitual action when the car interior is cold is to activate the seat heater, then when the cloud senses that the car interior temperature is cold while the user is using the vehicle, the seat heater will be automatically activated.

[0068] Another embodiment is used to understand the behavior feature finding method, which is a scene self-generation method based on the method described in this invention.

[0069] Obtain first data, which is environmental characteristic data, including: the weather is rainy, the temperature inside the vehicle is 10 degrees Celsius, the temperature outside the vehicle is 5 degrees Celsius, etc.; Obtain the second data, which is user behavior tracking data, such as: closing windows, turning on air conditioning, defrosting, etc.; The environmental feature data and business tracking data are processed, including a sequence list formed based on the first data and the second data, cleaning and merging based on the first data and the second data; The processed data is mined using an association rule algorithm; The final results table is as follows: Air vent mode_Defrost--→Weather = Rain / Snow and temperature difference between inside and outside the vehicle ≥5°C; Based on this method, the environmental conditions for activating the defrosting function are rainy or snowy weather and a temperature difference of ≥5°C between the inside and outside of the vehicle.

[0070] During vehicle operation, the defrosting function typically needs to be activated whenever the aforementioned weather characteristics appear. Based on the technical solution of this invention, similar routine operational scenarios can be refined and filtered to identify a specific environmental characteristic as the main trigger for a particular action. This allows for the automatic generation of corresponding scenarios, suggesting desired actions whenever a specific environmental characteristic appears during the vehicle's journey. Especially for novice drivers who are unsure what to do when vehicle windows fog up due to sudden environmental changes, there is no need to worry. The cloud platform will proactively remind the user to activate the defrosting function or activate it automatically for the user.

[0071] Another embodiment for understanding the behavior finding method in a scene is the scene self-generation method based on the method described in this invention.

[0072] Obtain first data, which is environmental characteristic data, including: the weather is rainy, the temperature inside the vehicle is 10 degrees Celsius, the temperature outside the vehicle is 5 degrees Celsius, etc.; Obtain the second data, which is user behavior tracking data, such as: closing windows, turning on air conditioning, heating seats, listening to music, etc.; The environmental feature data and business tracking data are processed, including a sequence list formed based on the first data and the second data, cleaning and merging based on the first data and the second data; The processed data is mined using an association rule algorithm; The final result table is: Turn on seat heating --→ Turn on steering wheel heating.

[0073] During vehicle operation, based on specific scenarios, the correlation between user behaviors can be analyzed to upgrade the scenarios. In the above embodiment, when some scenarios require the seat heating to be turned on, the scenario can be automatically upgraded using scenario generation technology, so that turning on the steering wheel heating may also be required after turning on the seat heating.

[0074] Based on the above method, the following applications can also be made: For example: refine the timing of recommendations for certain card functions; Taking the scenario of "opening the driver's side window and the passenger side window" as an example, where the humidity is between 40% and 60%, the interior temperature is above 28%, and the outside temperature is between 17 and 27 degrees Celsius, it is recommended to open the driver's side window and the passenger side window. For example: related action recommendations; Taking the action of "turn on seat heating --> turn on steering wheel heating" as an example, when a user turns on the seat heating, the system can recommend turning on the steering wheel heating as well.

[0075] Taking the "Ambient Lights - Lights and Music Follow --> Music - My Favorites" feature as an example, when a user turns on the ambient lights and the music follows their turn, the system can recommend turning on other music. Taking the "Bilibili playback ended --> data usage details" option as an example, when a user finishes watching a video, a pop-up window can be displayed to show the user's data usage details. For example: constructing and refining scenario examples; Taking the action of "Open Passenger Rest Mode --> Himalaya" as an example, if a user frequently opens Himalaya while in Passenger Rest Mode, it is advisable to add opening Himalaya to the default action of Passenger Rest Mode to improve the scenario.

[0076] For example: providing reference for UI design; Taking

Adjust seat backrest --&gt; Seat position memory

[0077] In some embodiments, the following methods may also be included: The system acquires third-party data, which is time or time period data. Specifically, the current time or time period can be determined by the vehicle's clock module or in-vehicle system, or based on geographical location and network time protocols. Feature values ​​formed based on the third-party data include morning peak, morning, afternoon, evening peak, and nighttime. This third-party data is incorporated into a first sequence list and participates in data processing based on association rules, which can generate time-related application scenarios. For example, it can automatically recommend more reasonable travel routes based on peak hours.

[0078] The fourth data is spatial data. Specifically, it is road information based on a third-party platform, and the feature values ​​of the journey based on the fourth data include tunnels, customized tracks, etc. This third data is incorporated into the first sequence list and participates in data processing based on association rules, which can generate time-related application scenarios. For example: whether to automatically turn on low beam headlights when entering a tunnel, or to remind the traveler to slow down when approaching a customized track. refer to Figure 2 This invention also provides a data-driven scene self-generation device, comprising: The first module is used to acquire first data and form a first wide table based on the first data; the first data includes at least one of environmental feature data and vehicle state signal features, and the first wide table includes feature type, feature name and corresponding feature value; The second module is used to acquire the second data, clean and merge the data of the first wide table and the second data to form the second wide table; the second data is user behavior tracking data. The third module is used to form a first sequence list based on the first wide table and the second wide table, perform data mining on the first sequence list using an association rule algorithm, and form a result table based on the data mining results; the first sequence list is a sequence list corresponding to user behavior and features, and the data mining methods include feature-based behavior finding, behavior-based feature finding, and scenario-based behavior finding. The fourth module is used to filter and select scenarios based on the result table, predict the corresponding scenario applications based on the selected scenarios, and recommend the scenario applications to the user terminal.

[0079] The content of the above method embodiments is applicable to the device embodiments. The specific functions implemented by the device embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0080] This invention also provides a data-driven scene self-generation system, including a memory, a processor, and a program stored in the memory and executable on the processor. When the program is executed by the processor, it implements the method described in the above embodiments.

[0081] Taking the example of a processor and memory in a vehicle controller being connected via a bus, the memory, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs and non-transitory computer-executable programs. Furthermore, the memory may include high-speed random access memory, and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, the memory may optionally include memory remotely located relative to the control processor, and these remote memories can be connected to the control device via a network.

[0082] The non-transitory software program and instructions required to implement the methods of the above embodiments are stored in memory and executed by the processor to perform the methods of the above embodiments.

[0083] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0084] This invention also provides a vehicle including the control device described in the above embodiments.

[0085] The vehicle can be a private car, such as a sedan, SUV, MPV, or pickup truck. It can also be a commercial vehicle, such as a van, bus, small truck, or large semi-trailer. The vehicle must have an electric motor capable of outputting power or acting as a generator to store mechanical energy. When the vehicle is a new energy vehicle, it can be a hybrid or a pure electric vehicle.

[0086] Since the vehicle applies all the technical solutions of the above-mentioned control device or vehicle controller, it has at least all the beneficial effects brought about by the technical solutions of the above embodiments, which will not be repeated here.

[0087] Furthermore, one embodiment of the present invention provides a computer-readable storage medium storing computer-executable instructions for performing the above-described method.

[0088] It is worth noting that, since the computer-readable storage medium of the present invention is capable of executing the methods of any of the above embodiments, the specific implementation methods and technical effects of the computer-readable storage medium of the present invention can be referred to the specific implementation methods and technical effects of the methods of any of the above embodiments.

[0089] Furthermore, one embodiment of the present invention provides a computer program product, including a computer program or computer instructions, the computer program or computer instructions being stored in a computer-readable storage medium, a processor of a computer device reading the computer program or computer instructions from the computer-readable storage medium, and the processor executing the computer program or computer instructions to cause the computer device to perform the above-described method.

[0090] It is worth noting that, since the computer program product of the present invention can execute the methods of any of the above embodiments, the specific implementation methods and technical effects of the computer program product of the present invention can be referred to the specific implementation methods and technical effects of the methods of any of the above embodiments.

[0091] It will be understood by those skilled in the art that all or some of the steps and systems in the methods disclosed above can be implemented as software, firmware, hardware, and suitable combinations thereof. Some or all of the physical components can be implemented as software executed by a processor, such as a central processing unit, digital signal processor, or microprocessor, or as hardware, or as an integrated circuit, such as an application-specific integrated circuit. Such software can be distributed on a computer-readable medium, which can include computer storage media (or non-transitory media) and communication media (or transient media). As is known to those skilled in the art, the term computer storage media includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules, or other data). Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technologies, CD-ROM, digital versatile disc (DVD) or other optical disc storage, magnetic cartridges, magnetic tape, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and is accessible to a computer. Furthermore, as is known to those skilled in the art, communication media typically include computer-readable instructions, data structures, program modules, or other data in modulated data signals such as carrier waves or other transmission mechanisms, and may include any information delivery medium.

[0092] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

Claims

1. A data-driven scene self-generation method, characterized in that, The method includes the following steps: Acquire first data and form a first wide table based on the first data; the first data includes at least one of environmental feature data and vehicle state signal features, and the first wide table includes feature type, feature name and corresponding feature value; The second data is obtained, and the first wide table and the second data are cleaned and merged to form a second wide table; the second data is user behavior tracking data. A first sequence table is formed based on the first wide table and the second wide table. Data mining is performed on the first sequence table using an association rule algorithm, and a result table is formed based on the data mining results. The first sequence table is a sequence table corresponding to user behavior and features. The data mining methods include finding behavior based on features, finding features based on behavior, and finding behavior based on scenarios. The system filters and selects scenarios based on the results table, predicts the corresponding scenario applications based on the selected scenarios, and recommends the scenario applications to the user terminal.

2. The method of claim 1, wherein, The process of forming a first wide table based on the first data includes: The first data is preprocessed to identify the feature type, feature name and corresponding feature value. The identified feature type, feature name and corresponding feature value are organized into a first wide table according to a preset format. Among them, the environmental feature data includes at least one of weather conditions, in-vehicle temperature, outside vehicle temperature and air humidity, and the vehicle status signal features include at least one of vehicle speed, engine speed, fuel level and battery level.

3. The method of claim 1, wherein, The process of cleaning and merging data from the first wide table and the second data to form the second wide table includes: Remove duplicate, erroneous, and incomplete data from the first wide table and the second data. Then, merge the cleaned first wide table and the second data according to preset rules. During the merging process, ensure that the correspondence between user behavior tracking data and environmental feature data or vehicle status signal feature data is accurate, forming a second wide table that contains user behavior tracking data and the corresponding environmental feature or vehicle status signal feature.

4. The method of claim 1, wherein, The step of using an association rule algorithm to perform data mining on the first sequence list and forming a result table based on the data mining results includes: The association rule algorithm is used to calculate the association index between user behavior and features in the first sequence list; the association index includes at least one of lift, mean of two confidence scores, and imbalance ratio; wherein, lift is the probability of a user behavior occurring given that a certain feature has occurred compared to the probability of the user behavior occurring without any conditions, mean of two confidence scores is the average of the confidence score of a user behavior occurring when a certain feature has occurred and the confidence score of a certain feature occurring when a certain user behavior occurs, and imbalance ratio is an indicator that measures the difference in sample size between a certain feature and a certain user behavior; Based on the aforementioned correlation indicators, results showing a strong correlation between user behavior and features are selected and a results table is generated. The strong correlation is determined by a comprehensive evaluation of the lift, the mean of the two confidence levels, and the imbalance ratio. When the lift is greater than a preset threshold, the mean of the two confidence levels is greater than a preset threshold, and the imbalance ratio is less than a preset threshold, it is determined that there is a strong correlation between user behavior and features.

5. The method of claim 1, wherein, The filtering and screening of the scenes based on the result table, the pre-judgment of the corresponding scene application according to the screened scenes, and the recommendation of the scene application to the user terminal, comprising: Filtering the result table according to the preset filtering condition, matching the scene mode in the filtered result table according to the current vehicle environment characteristic data and the user behavior buried point data; the filtering condition includes the characteristic value range limitation and / or the user behavior type limitation; When the scene mode matching the preset condition is matched, the corresponding scene application is pre-judged, and the scene application is recommended to the user terminal; wherein the preset condition includes the characteristic combination condition and / or the user behavior combination condition; the scene application includes at least one of the action execution of the vehicle behavior operation and the use recommendation of the multimedia.

6. The method of claim 1, wherein, The method further comprises: Under the condition of obtaining the user information, determining the intended operation of the user under a certain environmental condition based on the result table; Generating a personalized scene according to the historical behavior data and the preference of the user; When the vehicle environment characteristic data meets the trigger condition of the personalized scene, performing the operation corresponding to the personalized scene.

7. The method of claim 1, wherein, The method further comprises: Based on the result table, the construction and improvement of the scene case are performed; When the user frequently performs a certain operation in a certain scene, the operation is added to the default action of the scene; Or, when the user frequently performs multiple operations simultaneously in a certain scene, the multiple operations are combined into a new scene mode and added to the scene library.

8. A data-driven scene self-generation apparatus, characterized in that, The device comprises: A first module for obtaining first data and forming a first wide table based on the first data; the first data includes at least one of environmental characteristic data and vehicle state signal characteristics, and the first wide table includes feature types, feature names and corresponding feature values; A second module for obtaining second data, performing data cleaning and data merging on the first wide table and the second data, and forming a second wide table; the second data is user behavior buried point data; A third module for forming a first sequence table based on the first wide table and the second wide table, performing data mining on the first sequence table using an association rule algorithm, and forming a result table based on the data mining result; the first sequence table is a sequence table corresponding to user behavior and features, and the data mining method includes feature finding behavior, behavior finding feature, and scene finding behavior; A fourth module for filtering and screening scenes based on the result table, pre-judging the corresponding scene application according to the screened scenes, and recommending the scene application to the user terminal. 9.A data-driven scene self-generation system, characterized in that, The computer program is executed by the processor to implement the method of any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, the computer program comprising instructions that, when executed by a computer, cause the computer to perform the method of any one of claims 1 to 9. The computer program is executed by the processor to implement the method of any one of claims 1 to 7.

Images