An intelligent scenario-based vehicle application recommendation system and method

Abstract

The application relates to an application recommendation method for a vehicle, comprising: receiving data from a data source of the vehicle in real time; constructing a driving scene based on the received data; monitoring scene data in the driving scene; updating the scene data when the scene data changes, and determining whether to trigger an application recommendation function; and recommending at least one application to a user of the vehicle in response to determining to trigger the application recommendation function.

Description

Technical Field

[0001] This invention relates to the field of vehicles, and more specifically to an intelligent, scenario-based vehicle application recommendation system and method. Background Technology

[0002] With the development of technology, intelligent driving cockpits have attracted widespread attention. Currently, existing automotive application recommendation functions are basically based on the user's previous usage history. This recommendation mechanism is not intelligent enough and lacks personalization, thus failing to provide drivers with a more intelligent and comfortable application recommendation experience. Summary of the Invention

[0003] This invention relates to an application recommendation method for vehicles, comprising: receiving data in real time from a data source of the vehicle; constructing a driving scenario based on the received data; monitoring scenario data in the driving scenario; updating the scenario data when the scenario data changes, and determining whether to trigger an application recommendation function; and recommending at least one application to the user of the vehicle in response to determining that the application recommendation function has been triggered.

[0004] As described above, the scenario data includes one or more of the following: driver status, vehicle driving status, passenger information, date, and external environment.

[0005] As described above, the driver status includes one or more of the following: driver identity, driver mood, driver fatigue level, and driver distraction level; the vehicle driving status includes one or more of the following: vehicle gear, driving duration, vehicle speed, start time of driving, and end time of driving; the passenger information includes: whether there is someone in the front passenger seat, whether there are people in the back seat, whether there are children in the back seat, and the number of people in the back seat; the date and external environment include one or more of the following: traffic congestion status, weekday, holiday, road type, GPS location, navigation origin, and navigation destination.

[0006] The method described above, in determining whether to trigger the application recommendation function when the scene data changes, further includes: recording the time of the last recommendation; and determining to trigger the application recommendation function when the time since the last recommendation exceeds a predetermined time threshold.

[0007] The method described above further includes collecting application usage information, including: when the state of an application changes, recording a set of scene data and the state of all applications at that time, and storing the scene data and the state of all applications in a database.

[0008] As described above, the state of the application includes whether the application is open or closed.

[0009] In the method described above, the database is a local database that retains a fixed amount of up-to-date data.

[0010] The method described above, in response to determining that the application recommendation function is triggered to recommend at least one application to the user of the vehicle, includes: recommending at least one application to the user of the vehicle based on the similarity between current scene data and historical scene data in the database, wherein the historical scene data includes one or more sets of historical scene data, wherein each set of historical scene data includes: a set of scene data and the usage status of all corresponding applications.

[0011] The method described above, which recommends at least one application to a user of the vehicle based on the similarity between current scene data and historical scene data, includes: defining a similarity function; selecting a first part of the scene data as first-level feature data; selecting a second part of the scene data as second-level feature data; using the similarity function, based on the first-level feature data, calculating the similarity between the current scene data and the historical scene data to filter out the top m groups of historical scene data with the highest similarity; and using the similarity function, based on the second-level feature data, calculating the similarity between the current real-time scene data and the filtered top m groups of historical scene data to further filter out the top n groups of historical scene data with the highest similarity.

[0012] The method described above further includes: accumulating the similarity of each application among multiple applications in the top n sets of historical scene data; sorting the applications based on the recommendation value obtained by accumulating the similarity of each application, wherein the higher the recommendation value, the higher the ranking in the recommendation list; and providing the recommendation list to the user.

[0013] The present invention also relates to an application recommendation system for vehicles, the system including a processor configured to perform any of the methods described above.

[0014] The present invention also relates to a computer-readable medium including instructions stored thereon, which, when executed, cause a vehicle processor to perform any of the methods described above. Attached Figure Description

[0015] To further illustrate the advantages and features of the various embodiments of the present invention, a more detailed description of the embodiments will be presented with reference to the accompanying drawings. It should be understood that these drawings depict only typical embodiments of the invention and are therefore not intended to limit the scope of protection claimed by the invention.

[0016] Figure 1This is a system block diagram of scenario-based vehicle application recommendation according to an embodiment of the present invention;

[0017] Figure 2 This is a specific system architecture diagram of a scenario-based vehicle application recommendation according to an embodiment of the present invention; and

[0018] Figure 3 This is a flowchart of a scenario-based vehicle application recommendation algorithm according to an embodiment of the present invention. Detailed Implementation

[0019] The following detailed description is given with reference to the accompanying drawings. The drawings illustrate specific embodiments in which the claimed subject matter can be practiced by way of example. It should be understood that the following specific embodiments are intended to describe typical examples for illustrative purposes, but should not be construed as limiting the invention; those skilled in the art can make appropriate modifications and adjustments to the disclosed embodiments without departing from the spirit and scope of the claimed subject matter, provided they fully understand the spirit and intent of the invention.

[0020] Numerous specific details are set forth in the following detailed description to provide a thorough understanding of the various described embodiments. However, it will be apparent to those skilled in the art that the various described embodiments can be practiced without these specific details. In other instances, well-known methods, procedures, components, circuits, and networks have not been described in detail so as not to unnecessarily obscure aspects of the embodiments. Unless otherwise defined, the technical and scientific terms used herein should have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure belongs.

[0021] The embodiments of this application are exemplary implementations or examples. References to "embodiment," "one embodiment," "some embodiments," "various embodiments," or "other embodiments" in the specification mean that a particular feature, construction, or characteristic described in connection with an embodiment is included in at least some embodiments of the present technology, but not necessarily all embodiments. Various appearances of "embodiment," "one embodiment," or "some embodiments" do not necessarily refer to the same embodiment. Elements or aspects from one embodiment may be combined with elements or aspects from another embodiment.

[0022] To provide users (e.g., drivers) with a more intelligent driving experience, this invention proposes an intelligent recommendation system and method for car applications (hereinafter referred to as APPs) based on the driver's historical data and current scenario data. This invention can analyze the user's historical data using algorithms to understand the driver's preferences for using vehicle APPs, thereby proactively pushing vehicle APPs in real time and intelligently ranking them. More specifically, the system and method of this invention monitors scenario data in the current driving scenario in real time (such as, but not limited to, one or more of the following, including but not limited to: user facial expressions, fatigue and distraction information, driving duration information, current time information, and external environment information), and combines this with the user's historical APP operation data and the corresponding scenario data when operating the APPs, to push car APPs to the user in real time. Furthermore, the recommendation system and method of this invention can also adjust the recommendation results promptly based on user feedback.

[0023] Figure 1 This is a system block diagram of a scenario-based vehicle application recommendation system according to an embodiment of the present invention. In one embodiment, processor 102 can be configured to receive various data in real time from vehicle data source 104. The vehicle data source can be various systems, modules, and / or components of the vehicle, such as an MDS (driver monitoring system), various sensors, etc., as described in more detail below. Subsequently, processor 102 can be further configured to: construct a driving scenario based on the received data and monitor scenario data under the constructed driving scenario. Processor 102 can be further configured to: update the scenario data when the scenario data changes and determine whether to trigger the application recommendation function. Processor 102 can also be configured to: recommend at least one application to the user of the vehicle in response to determining that the application recommendation function has been triggered. For example, the present invention can recommend at least one application to the user of the vehicle via display 106. Display 106 can be various types of in-vehicle displays. In addition, processor 102 of the present invention can be integrated with other processing functions of the vehicle or can be a separate processor. In a further embodiment, the APP recommendation performed by processor 102 can be based on a similarity comparison between the constructed scenario data and historically stored scenario data from database 108, as described in more detail below.

[0024] Figure 2A more specific system architecture block diagram for scenario-based vehicle application recommendations is shown. In embodiments of the invention, the techniques of the invention can be primarily implemented in runtime platform 202. In a non-limiting embodiment, runtime platform 202 may be a microcomputer. In runtime platform 202, information regarding one or more of the following can be received from real-time data sources 2022 of the vehicle (e.g., relevant systems, modules, and / or components): driver status, vehicle driving status, passenger information, date, and external environment. As an example, referring to Table 1, which lists more detailed scenario data, the data that this invention can receive in real time from the real-time data source 2022 may include one or more of the following: IPS data, such as information about driver status received from, for example, a driver monitoring system, regarding driver identity (e.g., Face ID), driver mood, driver fatigue level, driver distraction level, etc.; vehicle driving data, such as information about vehicle driving status received from the vehicle's corresponding sensors via the vehicle bus, such as vehicle gear position, driving duration, vehicle speed, start driving time, end driving time, etc.; passenger data, such as information about passengers received from devices or systems such as cameras, such as whether there is someone in the front passenger seat, whether there are people in the back seat, whether there are children in the back seat, etc.; or information about the date and external environment, such as information about traffic congestion status, weekdays, holidays, road type, GPS location, navigation origin, navigation destination, etc., received from the vehicle's relevant systems, modules and / or components. It should be understood that the scenario-based vehicle application recommendation system and method of the present invention may include more or fewer scenario data than those listed above or in Table 1. Furthermore, it should be understood that the data ranges for each scenario data point in Table 1 are merely illustrative. For example, the number of rear-seat passengers may vary depending on the vehicle model; road types, holidays, etc., may also differ in different countries or regions, and so on. The range of each scenario data point can be defined according to different actual situations and needs, thereby enabling changes in scenario data to be identified.

[0025]

[0026]

[0027] Table 1

[0028] The driving scenario can be further constructed based on one or more of the scenario data received from the data source 2022. For example, a constructed driving scenario may include a set of real-time scenario data 2024. The constructed driving scenario may include one or more of the above-mentioned data information as scenario data. One or more of the above-mentioned scenario data can be flexibly selected to construct the driving scenario according to actual needs, and the present invention does not impose any limitations on this.

[0029] Subsequently, scene data within the constructed driving scenario can be monitored in real time. For example, it can be monitored whether scene data has changed. If scene data changes, it can be assumed that the user may have a need to use the app, or that their need to use the app may change. Therefore, the app recommendation function can be activated at this time.

[0030] As examples, not limitations, changes in scenario data can include: changes in user identity (e.g., a driver changing from one driver to another); changes in the driver's mood (e.g., from happy to unhappy); changes in the driver's fatigue level (e.g., from normal driving to fatigued driving); changes in vehicle driving status (e.g., from stopped to started); changes in road congestion (e.g., from smooth to congested); and changes in the driving date (e.g., whether it is a weekday or a holiday). It is understood that these are merely examples of changes in scenario data. All of the above scenario data may change, resulting in a change in the current driving scenario. At least two or more states can be defined for each scenario data point to distinguish whether a change has occurred. Alternatively, the states of each scenario data point can be divided according to the hierarchical methods commonly used in existing systems. For example, existing DMS systems can divide driver fatigue levels into different levels; when a driver transitions from one level of fatigue to another, the scenario data related to driver fatigue in this invention can be considered to have changed. Based on this principle, the states of the above scenario data points can be divided to define the boundaries of changes in scenario data. In addition, the boundaries for changes in scenario data can be defined according to actual needs.

[0031] When scene data changes, the scene data can be updated, and it can be determined whether to trigger the app recommendation function. For example, as an illustrative example, when a user's mood changes from happy to unhappy, from normal driving to fatigued driving, when the vehicle stops or starts, when the road becomes congested, or on holidays or weekdays, the user's app usage habits may differ. Therefore, when these scene data changes are detected, the app recommendation function can be triggered. In response to determining that the app recommendation function needs to be triggered, at least one app can be recommended to the vehicle's user 206. For example, the apps recommended by the system of the present invention to the user 206 can be displayed on the vehicle's display 106.

[0032] In some situations, scene data may change frequently. For example, a user's facial expressions may change rapidly, or during traffic congestion, vehicles may frequently switch between stopped and started states. In such cases, to provide a better user experience, even if changes in scene data are detected, the app recommendation function should not be triggered too frequently, as overly frequent app recommendations can be disruptive to the user. Therefore, in a preferred embodiment of the invention, a time threshold can be set. Within this time threshold, even if changes in scene data are detected, the app recommendation function may not be enabled. The time of the last recommendation can be recorded, and when the time since the last recommendation exceeds a predetermined time threshold, the app recommendation function can be triggered. This time threshold can be set according to actual needs, such as five minutes, ten minutes, twenty minutes, or half an hour. It should be understood that these listed time thresholds are merely examples, and other time thresholds can be set according to actual needs.

[0033] In the embodiments of this application, the amount of data received from various data sources 2022 is relatively large and the information is relatively redundant, which is not conducive to algorithm optimization. Therefore, the scene data included in the driving scenario constructed in this application is a simplified and representative data constructed in real time based on real-time data. In other words, the data received from various data sources 2022 can be simplified. As an illustrative example, when the vehicle system monitors the user's facial expressions, the data source 2022 can output data at a rate of approximately ten frames per second. However, the user's facial expressions may not change within these ten frames per second, or only one or two frames (or other quantities) may show a change in the user's emotions. Therefore, the embodiments of this invention can extract and simplify the received raw data, inputting only the data that has undergone substantial changes into the scenario-based car application recommendation system of this invention to construct or update the driving scenario.

[0034] In this invention, recommending apps to users relies on the user's historical app operation information and corresponding historical driving scenarios. Therefore, in a further embodiment, this invention can also collect app information. Collecting information from apps can include monitoring the app's usage status. For example, when the app's usage status changes, it can trigger the recording of a set of scenario data corresponding to that time and the status of all apps at that time, and store this data in database 204. The app's status can include whether the app is open or closed. This data stored in the database can serve as historical data corresponding to a specific user. When executing the app recommendation function, the historical data corresponding to a specific user can serve as part of the input data for the recommendation algorithm. The current scenario data can serve as another part of the input data for the recommendation algorithm. Therefore, this invention can compare the current driving scenario with the historical driving scenarios stored in the database. Furthermore, based on historical driving scenarios with high similarity and the app information used in that scenario, it can provide more accurate and personalized app recommendations to users in the current driving scenario, as described in more detail below.

[0035] A set of historical data can correspond to a set of scenario data and the corresponding APP status information. For different users (e.g., drivers), the historical data of each user can be stored separately in the database, enabling more personalized recommendations when executing the APP recommendation function. Historical data can be obtained based on the user's FACE ID. Regarding the source of user historical data, two methods can be considered: the first is to build it from a long-term accumulated historical database; the second is to directly build a smaller user database 2026 locally, which can have an upper limit on the amount of data, i.e., only retaining a certain amount of the latest data, the specific amount of data can be determined according to specific needs. A drawback of the first method is that it is time-consuming to extract and build from a large historical database every time a user switches. Therefore, embodiments of the present invention preferably consider using the second method to build the user historical database, such as... Figure 2 The example shown is a smaller user history database from 2026. This approach can significantly improve computational efficiency.

[0036] This invention takes current real-time scene data and a historical database (containing at least one or more sets of historical scene data and their corresponding APP information, such as the usage status data of all APPs in the driving scenario indicated by the set of scene data) as input, and performs APP recommendation function based on recommendation algorithm 2028. The system can further output the intelligent recommendation APP ranking results. User 206 can provide feedback based on the recommended APPs. This feedback can be received, recorded, and stored by the system to optimize the results of the next intelligent APP recommendation. For example, the next APP recommendation can be adjusted based on the feedback from user 206. As an example and not a limitation, the feedback from user 206 may include: whether user 206 ultimately selected and opened an APP in the recommendation list during the previous APP recommendation, and / or which APP(s) in the recommendation list were selected and opened, etc.

[0037] Figure 3 This is a flowchart of a scenario-based vehicle application recommendation algorithm according to an embodiment of the present invention. In general, the recommendation algorithm of the present invention may include three steps: preparing data (301), calculating similarity coefficients (302), and recommending and ranking apps (303).

[0038] In the 301 data preparation step, as described above, after the function of recommending apps to users is triggered, the current real-time scene data and the historical scene data stored in the database can be used as inputs to calculate the similarity coefficient. In embodiments of the present invention, the historical scene data stored in the database may include one or more sets of historical scene data. Furthermore, as described above, the system of the present invention can monitor the state of the app, and when the state of the app changes, it triggers the collection of the scene data at that time and all app state data at that time, and stores them in the database for later use. In embodiments of the present invention, the app state data may preferably include the app's open or closed state. By monitoring the state of the app and the corresponding scene data, the user's app usage habits can be understood. For example, as an example and not a limitation, a user may habitually open a music app to pass the time and relieve their mood during traffic jams; a user may habitually open an English learning app to improve their language skills on weekdays, and so on. These app usage habits may vary from user to user. Even for the same user, their app usage habits may change at different times (e.g., six months ago and now). Therefore, recording the app's state and its corresponding set of scenario data when the app's state changes is beneficial. This data can provide a foundation for more accurate and intelligent app recommendations later on. Furthermore, preferably, as mentioned above, the historical scenario data can be a small user database 2026 built locally to improve computational efficiency.

[0039] The data prepared in step 301 can be used as input for subsequent step 302 to calculate the similarity coefficient. Step 302, calculating the similarity coefficient, may include: defining a similarity function; selecting the first part of the scene data as first-level feature data; selecting the second part of the scene data as second-level feature data; using the defined similarity function, based on the first-level feature data, calculating the similarity between the current real-time scene data and historical scene data to filter out the top m groups of historical scene data with the highest similarity; using the similarity function, based on the second-level feature data, calculating the similarity between the current real-time scene data and the filtered top m groups of historical scene data to further filter out the top n groups of historical scene data with the highest similarity.

[0040] More specifically, in this invention, the recommendation algorithm can preferably be implemented using the idea of ​​collaborative filtering. Therefore, appropriate similarity functions known in the art can be used to calculate similarity. For example, as a non-limiting example, the similarity functions that can be used may include: cosine similarity, Euclidean distance, Pearson correlation coefficient, etc. In a preferred embodiment of this invention, the calculation of similarity is divided into two parts. First, the scene data is preprocessed and divided into two levels: first-level feature data and second-level feature data. The first-level feature data and the second-level feature data can be one or more of the above-mentioned scene data. For the first-level feature data, features that can more efficiently filter out historical scene data with higher similarity to the current driving scene can be given priority. As a non-limiting example, the solution of this invention can give priority to the most recent driving time as one of the first-level feature data. For example, users can exhibit different APP usage habits at different times, so APP usage during recent driving may be more valuable than APP usage habits in earlier periods. For example, compared to APP usage data from a month ago (or other time periods), the user's APP usage data in the most recent month may be more accurate, thus enabling more precise APP recommendations based on the latest data. Furthermore, other scene data can be selected as first-level feature data, such as facial expressions (happy or unhappy). First-level feature data is preferably data that can be quickly and accurately filtered from the database with high scene relevance or similarity, thereby reducing computational load and improving efficiency. Second-level feature data can be further refined, more precise scene data, further improving the relevance and matching degree between the filtered historical driving scenarios and the current driving scenario. Based on the above principles, one or more of the aforementioned scene data can be selected as first-level and second-level feature data according to actual needs. The preferred first-level feature data differs from the preferred second-level feature data.

[0041] The first m sets of historical scene data filtered using the first-level feature data can include any suitable number of historical scene data, such as ten, eight, six, etc. The first n sets of historical scene data filtered using the second-level feature data can also include any suitable number of historical scene data less than m, such as five, four, three, etc.

[0042] After obtaining the top n sets of historical scene data with the highest similarity, the final APP recommendation and ranking can be performed. The similarity of each application among multiple applications in the top n sets of historical scene data can be accumulated; based on the recommendation value obtained from accumulating the similarity of each application, the applications are ranked; then, APPs can be recommended to vehicle users in descending order of recommendation value. For example, taking n=5 as an example, the similarity between these 5 sets of historical scene data and the current driving scenario are a1, a2, a3, a4, and a5, respectively. Furthermore, each of these 5 sets of historical scene data contains the usage status of all APPs, and the usage status of each APP may be different. For example, in the first set of historical scene data, APP1 may be open, APP2 may be closed, and other APPs may be open or closed. Therefore, the first similarity a1 can be used as the usage value of APP1 in the first set of historical scene data, and 0 can be used as the usage value of APP2 in the first set of historical scene data. Similarly, APP1, APP2, and other APPs can also have corresponding usage values ​​in the second to fifth sets. The usage scores of each app can be summed to obtain the final recommendation score (cumulative similarity score) for that app. Apps with higher recommendation scores are placed at the top of the recommendation list, in descending order. Apps with higher recommendation scores are the ones users are most likely to use.

[0043] Therefore, the solution of this invention can analyze driver app usage preferences based on historical data and intelligently recommend apps according to different drivers and driving scenarios, bringing drivers a more comfortable driving experience. Specifically, this invention integrates rich scenario data, including: external environment, vehicle driving data, in-vehicle passenger data, and driver status data (e.g., facial expressions, fatigue, distraction, etc.), and makes relatively more accurate app recommendations based on the similarity between the current driving scenario and historical scenarios. In this invention, different users will receive different recommendations, thus achieving personalized recommendations. For new users who are driving a vehicle for the first time, they can register, and their app usage preference data in different driving scenarios can be further accumulated and stored; the first recommendation for a new user can be made according to predetermined rules or the default in-vehicle app recommendation order. Even for the same user, different app recommendations will be given in different driving scenarios. App recommendations can rely on the user's historical app operations and corresponding historical scenarios. For example, recommendations will differ depending on different emotions, different vehicle driving states, and different driving times, depending on the personal preferences reflected in the user's historical scenarios. Furthermore, the solution of this invention has good real-time performance because it monitors data from various data sources in real time. In addition, this invention can also adjust the recommendation results in a timely manner based on the user's latest operation data. The recommended content for the same user will change in real time based on the user's recent historical operations, and is not static. For example, if the user's recent preferences change, the recommendations will also make corresponding changes.

[0044] It should be understood that at least some of the functions attributable to the systems and devices described in this disclosure can be implemented as instructions on a computer-readable storage medium, such as random access memory (RAM), read-only memory (ROM), non-volatile random access memory (NVRAM), electrically erasable programmable read-only memory (EEPROM), flash memory, magnetic media, optical media, or other tangible computer-readable storage media. The computer-readable storage medium may be referred to as non-transient. The instructions are executable to support one or more aspects of the functions described in this disclosure.

[0045] Therefore, those skilled in the art can make appropriate modifications and adjustments to the embodiments specifically described above without departing from the spirit and intent of the invention. Thus, it is intended that the claimed subject matter not be limited to the specific examples disclosed, but also include all implementations falling within the scope of the appended claims and their equivalents.

Claims

1. An application recommendation method for vehicles, comprising: Data is received in real time from the vehicle's data source; Construct driving scenarios based on received data; Monitor scene data in the driving scenario; When the scene data changes, update the scene data and determine whether to trigger the application recommendation function; as well as In response to determining that the application recommendation function is triggered, recommending at least one application to the user of the vehicle includes: recommending at least one application to the user of the vehicle based on the similarity between current scene data and historical scene data in the database, wherein the historical scene data includes one or more sets of historical scene data, and each set of historical scene data includes: a set of scene data and the usage status of all corresponding applications, wherein the status of the application includes whether the application is open or closed. The recommendation of at least one application to the user of the vehicle based on the similarity between current scene data and historical scene data includes: Define a similarity function; Select the first part of the scene data as the first-level feature data; Select the second part of the scene data as the second-level feature data; Using the similarity function, based on the first-level feature data, the similarity between the current scene data and the historical scene data is calculated, so as to filter out the top m groups of historical scene data with the highest similarity. Using the similarity function, based on the second-level feature data, the similarity between the current scene data and the selected first m groups of historical scene data is calculated to further filter out the top n groups of historical scene data in terms of similarity. The similarity between the current scene data and the selected first m groups of historical scene data is used as the similarity of each application in the multiple applications under the top n groups of historical scene data whose usage status is open, and 0 is used as the similarity of each application in the multiple applications under the top n groups of historical scene data whose usage status is closed. The similarity of each application in multiple applications under the top n sets of historical scene data is accumulated; Apps are ranked based on a recommendation score derived from the sum of their similarity scores. A higher recommendation score places an app higher in the recommendation list. The recommended list is provided to the user.

2. The method as described in claim 1, characterized in that, The scenario data includes one or more of the following: driver status, vehicle driving status, passenger information, date, and external environment.

3. The method as described in claim 2, characterized in that, The driver status includes one or more of the following: driver identity, driver mood, driver fatigue level, and driver distraction level; The vehicle driving status includes one or more of the following: vehicle gear, driving duration, vehicle speed, start time of driving, and end time of driving; The passenger information includes: whether there is someone in the front passenger seat, whether there is someone in the back seat, whether there are children in the back seat, and the number of people in the back seat. The date and external environment include one or more of the following: traffic congestion, weekday, holiday, road type, GPS location, navigation origin, and navigation destination.

4. The method as described in claim 1, characterized in that, Determining whether to trigger the application recommendation function when the scene data changes further includes: Record the time of the last recommendation; When the time since the last recommendation exceeds a predetermined time threshold, the application recommendation function will be triggered.

5. The method as described in claim 1, characterized in that, This further includes collecting application usage information, including: When the state of an application changes, record the corresponding set of scene data and the state of all applications, and store the scene data and the state of all applications in the database.

6. The method as described in claim 5, characterized in that, The database is a local database that retains a fixed amount of the latest data.

7. An application recommendation system for a vehicle, the system comprising a processor configured to perform the method as described in any one of claims 1-6.

8. A computer-readable medium comprising instructions stored thereon, which, when executed, cause a processor of a vehicle to perform the method as described in any one of claims 1-6.

Images