A vehicle data processing method
By deploying basic triggers on the vehicle side and automatically classifying them in the cloud, we solve the existing technology problems of heavy vehicle data collection burden and inability to differentiate between scenarios, and achieve a flexible and efficient data collection method.
Patent Information
- Application Number
- CN202111561591.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2021-12-20
- Publication Date
- 2025-01-21
- Estimated Expiration
- 2041-12-20
AI Technical Summary
The existing vehicle data collection methods generally store all data generated by the vehicle and upload it to the server, resulting in a large burden of vehicle data collection and the inability to distinguish vehicle data in scenarios, and low flexibility.
Deploy several basic triggers on the vehicle side, trigger these triggers based on set scenarios, query scenario deployment information, evaluate vehicle data, and upload data and trigger call status information to the cloud when conditions are met. The cloud classifies vehicle data into a scene library based on scene deployment information.
By deploying basic triggers on the car side and automatic classification on the cloud, a data collection method that combines the car side and the cloud is realized, reducing the burden on the car side and improving the flexibility of data collection.
Smart Images
Figure CN114333101B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of vehicle data collection, and in particular to a vehicle data processing method. Background Art
[0002] With the rapid development of automobile intelligence, the automatic driving of automobiles has attracted more and more attention. However, the current automatic driving of automobiles still has great safety risks and low reliability. Automobile data collection can provide a large amount of data support for automatic driving technology, which has a vital impact on the reliability and safety of automatic driving. At present, vehicle data collection is generally divided into two stages, one is that the vehicle stores the data generated by itself on the local hard disk, and the other is to regularly upload the local hard disk of the vehicle to the server. However, the existing vehicle data collection method is to generally store all the data generated by the vehicle and upload it to the server. On the one hand, the burden of vehicle data collection is heavy, and on the other hand, it is impossible to distinguish the scene of vehicle data, and the flexibility of data collection is low. Summary of the invention
[0003] An embodiment of the present invention provides a vehicle data processing method, which can reduce the burden on the vehicle side and improve the flexibility of data collection.
[0004] An embodiment of the present invention provides a vehicle data processing method, wherein a plurality of basic triggers are deployed on a vehicle side; the method comprises:
[0005] When the vehicle side runs into a set scenario and causes at least one of the basic triggers to be triggered, querying the scenario deployment information;
[0006] The vehicle end evaluates the vehicle data acquired in the set scenario according to the scenario deployment information, and when the data upload condition is met, uploads the vehicle data and the basic trigger call status information to the cloud;
[0007] The cloud classifies the vehicle data into a scene library according to the scene deployment information.
[0008] As an improvement of the above solution, the vehicle end evaluates the vehicle data acquired in the set scenario according to the scenario deployment information, including:
[0009] The vehicle end scores the vehicle data obtained under the set scenario according to the scenario deployment information; when the score of the vehicle data is greater than a preset threshold, it is determined that the data upload condition is met.
[0010] As an improvement of the above solution, a plurality of the basic triggers form a plurality of basic trigger groups; the scenario deployment information carries the calling status of the basic triggers in each of the basic trigger groups;
[0011] Then, the vehicle end scores the vehicle data acquired in the set scenario according to the scenario deployment information, including:
[0012] The vehicle end determines the trigger quantity of the basic triggers called in each of the basic trigger groups according to the scenario deployment information;
[0013] The vehicle end scores the vehicle data according to the trigger quantity of the basic triggers called in each of the basic trigger groups and the preset weights corresponding to the basic trigger groups.
[0014] As an improvement of the above solution, the scoring of the vehicle data according to the trigger quantity of the called basic triggers in each of the basic trigger groups and the preset weight corresponding to the basic trigger group includes:
[0015] According to the formula Q = ∑w j ·x j Calculating a total score of the vehicle data;
[0016] Among them, w j represents the preset weight of the jth basic trigger group, x j Indicates the number of triggers of the called basic trigger in the jth basic trigger group;
[0017] The total score of the vehicle data is normalized to obtain a score of the vehicle data.
[0018] As an improvement of the above solution, a scene library classifier and multiple scene libraries are deployed on the cloud;
[0019] Then, the cloud classifies the vehicle data into a scene library according to the scene deployment information, including:
[0020] The cloud sends the received scene deployment information and the vehicle data to the scene library classifier;
[0021] The scenario library classifier queries the calling status of each basic trigger in the basic trigger group according to the scenario deployment information;
[0022] The scenario library classifier classifies the vehicle data into a corresponding scenario library according to the calling status of the basic triggers in each of the basic trigger groups.
[0023] As an improvement of the above solution, the scenario deployment information includes character strings of each of the basic trigger groups; one character in the character string of each basic trigger group corresponds to a calling state of a basic trigger.
[0024] As an improvement of the above solution, the character string is a binary code.
[0025] As an improvement of the above solution, the vehicle side deploys a plurality of high-order scene triggers, each of which corresponds to a scene;
[0026] The method further comprises:
[0027] When the vehicle side runs to a set scene, the basic triggers in each of the basic trigger groups are called and triggered according to the scene deployment information corresponding to the set scene;
[0028] Among them, the scene deployment information is determined according to the basic triggers that need to be called by the corresponding high-order scene triggers in each of the basic trigger groups.
[0029] As an improvement of the above solution, the method further includes:
[0030] The vehicle end collects vehicle data within a first set time before the basic trigger is triggered and a second set time after the trigger is triggered.
[0031] As an improvement of the above-mentioned solution, the basic trigger group includes: a perception module basic trigger group, a positioning module basic trigger group, a planning module basic trigger group, a control module basic trigger group and a vehicle-road collaboration module basic trigger group.
[0032] As an improvement of the above solution, the scenario library classifier queries the calling status of each basic trigger in the basic trigger group according to the scenario deployment information, including:
[0033] The scenario library classifier queries the status of each code element of the binary code of each basic trigger group in the scenario deployment information;
[0034] When the state of the i-th code element of the binary code of the basic trigger group is 1, determining that the basic trigger corresponding to the i-th code element in the basic trigger group is called;
[0035] When the state of the i-th code element of the binary code of the basic trigger group is 0, it is determined that the basic trigger corresponding to the j-th code element in the basic trigger group has not been called; wherein j=1, 2, ..., n; n represents the number of basic triggers in the corresponding basic trigger group.
[0036] Compared with the prior art, the beneficial effects of the embodiments of the present invention are as follows: by deploying a number of basic triggers on the vehicle side; when the vehicle side runs into a set scenario causing at least one of the basic triggers to be triggered, the scenario deployment information is queried, and then the vehicle data acquired under the set scenario is evaluated according to the scenario deployment information, and when the data upload conditions are met, the vehicle data and the basic trigger call status information are uploaded to the cloud; the cloud classifies the vehicle data into a scenario library according to the scenario deployment information; the embodiments of the present invention deploy basic triggers on the vehicle side, trigger the upload of vehicle data and automatic classification on the cloud based on the call triggering conditions of the basic triggers, thereby realizing a data collection method combining the vehicle side and the cloud side. Compared with the prior art method in which the vehicle uploads all data to the cloud in general, this method can effectively reduce the burden on the vehicle side and improve the flexibility of data collection. BRIEF DESCRIPTION OF THE DRAWINGS
[0037] In order to more clearly illustrate the technical solution of the present invention, the drawings required for use in the implementation mode will be briefly introduced below. Obviously, the drawings described below are only some implementation modes of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0038] Figure 1 is a flow chart of a vehicle data processing method provided by an embodiment of the present invention;
[0039] Figure 2 It is a schematic diagram of the layout of the trigger library provided in an embodiment of the present invention. DETAILED DESCRIPTION
[0040] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0041] See also Figure 1 , is a flow chart of a vehicle data processing method provided by an embodiment of the present invention, wherein several basic triggers are deployed on the vehicle side; the method comprises:
[0042] S11: When the vehicle side runs into a set scenario and causes at least one of the basic triggers to be triggered, querying the scenario deployment information;
[0043] Exemplarily, the set scene can be an APA parking scene, a low-speed / high-speed road scene, an avoidance scene, an intersection scene, etc. A high-level scene trigger can be formed by combining the basic triggers in different ways, wherein each set scene is deployed with a high-level scene trigger. The scene deployment information records the basic triggers that need to be called for the corresponding set scene, which can be pre-set and stored locally on the vehicle side to facilitate the subsequent call status query of the basic trigger.
[0044] In an embodiment of the present invention, different combinations are performed based on the deployed basic triggers to form high-order scenario triggers for different scenarios, which are used to cope with various complex autonomous driving scenarios. There is no need to customize a single trigger for each autonomous driving scenario. All development tasks will be concentrated on the basic triggers. When there is no need to add a new basic trigger, a new high-order scenario trigger can be customized in a very short time, thereby improving development efficiency. At the same time, since all high-order scenario triggers are only related to the basic triggers, the vehicle side only needs to deploy the basic triggers, which can reduce the vehicle side resource consumption to a certain extent.
[0045] S12: the vehicle side evaluates the vehicle data acquired in the set scenario according to the scenario deployment information, and when the data upload condition is met, uploads the vehicle data and the basic trigger calling status information to the cloud;
[0046] Furthermore, the method further comprises:
[0047] The vehicle end collects vehicle data within a first set time before the basic trigger is triggered and a second set time after the trigger is triggered.
[0048] For example, the vehicle end collects vehicle data within 10 seconds before and after the trigger is triggered.
[0049] S13: The cloud classifies the vehicle data into a scene library according to the scene deployment information.
[0050] In the first step, a scene library classifier and multiple scene libraries are deployed on the cloud;
[0051] Then, the cloud classifies the vehicle data into a scene library according to the scene deployment information, including:
[0052] The cloud sends the received scene deployment information and the vehicle data to the scene library classifier;
[0053] The scenario library classifier queries the calling status of each basic trigger in the basic trigger group according to the scenario deployment information;
[0054] The scenario library classifier classifies the vehicle data into a corresponding scenario library according to the calling status of the basic triggers in each of the basic trigger groups.
[0055] In an embodiment of the present invention, basic triggers are deployed on the vehicle side, and the vehicle data is uploaded to the cloud based on the calling trigger conditions of the basic triggers. The cloud side queries the calling status of the basic triggers in each basic trigger group based on the scene deployment information, thereby determining which basic triggers have been called by the vehicle data, and automatically classifying them into corresponding scene libraries, thereby achieving the purpose of automatic data classification and reducing the complexity of the cloud scene library classifier. Compared with the prior art method in which the vehicle uploads all data to the cloud in a general manner, it can effectively reduce the burden on the vehicle side and improve the flexibility of data collection.
[0056] In an optional embodiment, the vehicle end evaluates the vehicle data acquired in the set scenario according to the scenario deployment information, including:
[0057] The vehicle end scores the vehicle data acquired in the set scenario according to the scenario deployment information;
[0058] When the score of the vehicle data is greater than a preset threshold, it is determined that the data uploading condition is met.
[0059] Exemplarily, data uploading is triggered only when the score of the vehicle data is greater than a preset threshold, otherwise data uploading is not triggered, thereby avoiding uploading low-value vehicle data to the cloud and improving the quality of vehicle data collected in the cloud.
[0060] In an optional embodiment, a plurality of the basic triggers form a plurality of basic trigger groups; the scenario deployment information carries the calling status of the basic triggers in each of the basic trigger groups;
[0061] The scenario deployment information includes character strings of each of the basic trigger groups; one character in the character string of each basic trigger group corresponds to the calling state of a basic trigger.
[0062] Furthermore, the character string is a binary code.
[0063] The basic trigger group includes: a perception module basic trigger group, a positioning module basic trigger group, a planning module basic trigger group, a control module basic trigger group and a vehicle-road collaboration module basic trigger group.
[0064] Furthermore, the vehicle side deploys a plurality of high-level scene triggers, each of which corresponds to a scene;
[0065] The method further comprises:
[0066] When the vehicle side runs to a set scene, the basic triggers in each of the basic trigger groups are called and triggered according to the scene deployment information corresponding to the set scene;
[0067] Among them, the scene deployment information is determined according to the basic triggers that need to be called by the corresponding high-order scene triggers in each of the basic trigger groups.
[0068] like Figure 2 As shown in the figure, the basic triggers deployed on the vehicle side are stored in the basic trigger library and are divided into five basic trigger groups according to the functional modules of the car, such as perception, positioning, planning, control, and vehicle-road collaboration. Each type of basic trigger group contains various basic triggers under the corresponding functional module, such as the GPS signal trigger of the positioning module. Each basic trigger group is managed by a specific binary code to manage which basic triggers under the corresponding basic trigger group need to be called. All binary codes will be determined according to the basic triggers that need to be called by each basic trigger group when designing high-level scenario triggers in the early stage, and will remain unchanged.
[0069] Take the triggering of a high-order scenario trigger as an example: query the binary codes of each basic trigger group, each basic trigger in each basic trigger cluster will have a sequential number, when the binary code of a basic trigger group is 100110101, counting from right to left, each number represents whether the basic trigger with the corresponding number in the basic trigger group is called, 1 represents being called, and 0 represents not being called; for example: 100110101 means that the basic triggers numbered 1, 4, 5, 7, and 9 in the basic trigger group are called.
[0070] Among them, each basic trigger group can be based on To determine the code element corresponding to the basic trigger P, and then based on Generate binary code; where f n (trigger p) is a function that generates binary code, and n represents the number of basic triggers in the basic trigger group.
[0071] Then, the vehicle end scores the vehicle data acquired in the set scenario according to the scenario deployment information, including:
[0072] The vehicle end determines the trigger quantity of the basic triggers called in each of the basic trigger groups according to the scenario deployment information:
[0073] The vehicle end scores the vehicle data according to the trigger quantity of the basic triggers called in each of the basic trigger groups and the preset weights corresponding to the basic trigger groups.
[0074] In an optional embodiment, scoring the vehicle data according to the trigger quantity of the called basic triggers in each of the basic trigger groups and the preset weight corresponding to the basic trigger group includes:
[0075] According to the formula Q = ∑w j ·x j Calculating a total score of the vehicle data;
[0076] Among them, w j represents the preset weight of the jth basic trigger group, x j Indicates the number of triggers of the called basic trigger in the jth basic trigger group;
[0077] The total score of the vehicle data is normalized to obtain a score of the vehicle data.
[0078] The preset weight of each basic trigger group can be calculated based on the ratio of the basic triggers in the corresponding basic trigger group to all the basic triggers deployed on the vehicle side. The total score of the vehicle data can be normalized using a sigmoid function. It should be noted that the sigmoid function belongs to the prior art and will not be described in detail here.
[0079] In an ideal situation, when the vehicle runs to the set scene, the basic triggers required to be called in the corresponding high-level scene triggers should be called; however, in the scene, some basic triggers may not be triggered because the real scene is too complicated. In order to collect as much valuable data as possible, the collected vehicle data is scored on the vehicle side. When the normalized score is greater than the preset threshold, the vehicle side triggers the data upload mechanism. It should be noted that the preset threshold can be customized by the user according to actual conditions, for example, set to 0.9. When the score of vehicle data is higher than 0.9, the vehicle data is considered valuable and is triggered to be uploaded. Through the scoring of vehicle data, valuable vehicle data can be collected as much as possible, reducing unnecessary data collection and uploading.
[0080] In an optional embodiment, the scenario library classifier queries the calling status of each basic trigger in the basic trigger group according to the scenario deployment information, including:
[0081] The scenario library classifier queries the status of each code element of the binary code of each basic trigger group in the scenario deployment information;
[0082] When the state of the i-th code element of the binary code of the basic trigger group is 1, determining that the basic trigger corresponding to the i-th code element in the basic trigger group is called;
[0083] When the state of the i-th code element of the binary code of the basic trigger group is 0, it is determined that the basic trigger corresponding to the i-th code element in the basic trigger group has not been called; wherein, i=1, 2, ..., n; n represents the number of basic triggers in the corresponding basic trigger group.
[0084] In the embodiment of the present invention, after the vehicle data is uploaded to the cloud, the scene library classifier in the scene library will automatically classify the uploaded vehicle data according to the calling situation of the basic trigger, for example, automatically classify it into the corresponding scene library according to the sensor category corresponding to the vehicle data, the scene it is in, the recording time, etc. The same set of vehicle data will be marked by both the low-level and high-level scene library systems, providing a large amount of data for the future development of high-level autonomous driving technology.
[0085] Compared with the prior art, the beneficial effects of the embodiments of the present invention are:
[0086] 1. By deploying basic triggers on the vehicle side, the vehicle data upload and automatic classification on the cloud are triggered based on the calling trigger conditions of the basic triggers, thus realizing a data collection method combining the vehicle side and the cloud side. Compared with the existing technology in which the vehicle uploads all data to the cloud in general, it can effectively reduce the burden on the vehicle side and improve the flexibility of data collection.
[0087] 2. There is no need to customize a single trigger for each autonomous driving scenario. All development tasks will be concentrated on the basic triggers. When there is no need to add new basic triggers, a new high-level scenario trigger can be customized in a very short time, improving development efficiency.
[0088] 3. Since all high-level scenario triggers are only related to basic triggers, the vehicle side only needs to deploy basic triggers, which can reduce the resource consumption of the vehicle side to a certain extent.
[0089] 4. The cloud-based scenario library classifier can automatically classify vehicle data based on the calling conditions of basic triggers, reducing the complexity of the classifier.
[0090] The above is a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, many improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications are also considered to be within the scope of protection of the present invention.
Claims
1. A vehicle data processing method, characterized in that: Several basic triggers are deployed on the vehicle side; the method includes: When the vehicle side runs to a set scene and causes at least one of the basic triggers to be triggered, the scene deployment information is queried, and the scene deployment information records the basic triggers that need to be called for the corresponding set scene; The vehicle end evaluates the vehicle data acquired in the set scenario according to the information of the basic trigger called in the scenario deployment information, and uploads the vehicle data and the basic trigger calling status information to the cloud when the data uploading condition is met; The cloud classifies the vehicle data into a scene library according to the scene deployment information.
2. The vehicle data processing method according to claim 1, characterized in that: The vehicle end evaluates the vehicle data acquired in the set scenario according to the scenario deployment information, including: The vehicle end scores the vehicle data acquired in the set scenario according to the scenario deployment information; When the score of the vehicle data is greater than a preset threshold, it is determined that the data uploading condition is met.
3. The vehicle data processing method according to claim 2, characterized in that: A plurality of the basic triggers form a plurality of basic trigger groups; the scenario deployment information carries the calling status of the basic triggers in each of the basic trigger groups; Then, the vehicle end scores the vehicle data acquired in the set scenario according to the scenario deployment information, including: The vehicle end determines the trigger quantity of the basic triggers called in each of the basic trigger groups according to the scenario deployment information; The vehicle end scores the vehicle data according to the trigger quantity of the basic triggers called in each of the basic trigger groups and the preset weights corresponding to the basic trigger groups.
4. The vehicle data processing method according to claim 3, characterized in that: Scoring the vehicle data according to the trigger quantity of the called basic triggers in each of the basic trigger groups and the preset weight corresponding to the basic trigger group includes: According to the formula Q = ∑w j ·x j Calculating a total score of the vehicle data; Among them, w j represents the preset weight of the jth basic trigger group, x j Indicates the number of triggers of the called basic trigger in the jth basic trigger group; The total score of the vehicle data is normalized to obtain a score of the vehicle data.
5. The vehicle data processing method according to claim 3, characterized in that: The cloud is deployed with a scene library classifier and multiple scene libraries; Then, the cloud classifies the vehicle data into a scene library according to the scene deployment information, including: The cloud sends the received scene deployment information and the vehicle data to the scene library classifier; The scenario library classifier queries the calling status of each basic trigger in the basic trigger group according to the scenario deployment information; The scenario library classifier classifies the vehicle data into a corresponding scenario library according to the calling status of the basic triggers in each of the basic trigger groups.
6. The vehicle data processing method according to claim 5, characterized in that: The scenario deployment information includes character strings of each of the basic trigger groups; one character in the character string of each basic trigger group corresponds to the calling state of a basic trigger.
7. The vehicle data processing method according to claim 6, characterized in that: The character string is a binary code.
8. The vehicle data processing method according to claim 7, characterized in that: The vehicle side deploys a plurality of high-level scene triggers, each of which corresponds to a scene; The method further comprises: When the vehicle side runs to a set scene, the basic triggers in each of the basic trigger groups are called and triggered according to the scene deployment information corresponding to the set scene; Among them, the scene deployment information is determined according to the basic triggers that need to be called by the corresponding high-order scene triggers in each of the basic trigger groups.
9. The vehicle data processing method according to claim 8, characterized in that: The method also includes: the vehicle end collects vehicle data within a first set time before the basic trigger is triggered to a second set time after the trigger is triggered.
10. The vehicle data processing method according to claim 3, characterized in that: The basic trigger group includes: a perception module basic trigger group, a positioning module basic trigger group, a planning module basic trigger group, a control module basic trigger group and a vehicle-road collaboration module basic trigger group.
11. The vehicle data processing method according to claim 7, characterized in that: The scenario library classifier queries the calling status of each basic trigger in the basic trigger group according to the scenario deployment information, including: The scenario library classifier queries the status of each code element of the binary code of each basic trigger group in the scenario deployment information; When the state of the i-th code element of the binary code of the basic trigger group is 1, determining that the basic trigger corresponding to the i-th code element in the basic trigger group is called; When the state of the i-th code element of the binary code of the basic trigger group is 0, it is determined that the basic trigger corresponding to the i-th code element in the basic trigger group has not been called; wherein, i=1, 2, ..., n; n represents the number of basic triggers in the corresponding basic trigger group.
Citation Information
Patent Citations
Vehicle data acquisition method and device
CN112164155A
Data acquisition method and device
CN112562114A