Method, apparatus, device and storage medium for processing data of an autonomous vehicle

Abstract

An embodiment of the present application provides a method, apparatus, device, and storage medium for processing data of an autonomous vehicle. The method can be applied to scenarios such as ports, mines, highways, or industrial parks, and includes: obtaining data to be processed through a first analysis server and caching the data to be processed; determining at least one matching processing server based on the data processing requirement information corresponding to the data to be processed; and sending the data to be processed to at least one processing server for processing respectively. The embodiment of the present application solves the problems of low debugging efficiency in training and debugging during the data flow process of collected data and low data processing efficiency, and effectively ensures the efficiency of data flow and processing.

Description

Technical Field

[0001] The embodiments of the present application relate to the technical field of data processing, and in particular, to a method, device, equipment, and storage medium for processing data of an autonomous driving vehicle. Background Art

[0002] With the emergence of diverse transportation demands, technologies such as autonomous driving and driverless driving have gradually been more widely applied. At the same time, with the continuous development of communication technologies, it is imperative to provide more efficient data transfer and more secure and effective protection measures. During the driving process, an autonomous driving vehicle continuously collects a large amount of environmental data and vehicle status data through its various sensors and monitoring devices, and collaborates with its own control unit and cloud servers, edge cloud servers, cloud processors, etc. to complete the real-time transfer and processing of these data to monitor the driving status of the vehicle itself. To ensure the efficient progress of this real-time transfer and processing process during actual driving, it is necessary to train and debug the transfer process of the collected data such as environmental data and vehicle status data required for each data processing model and function among various servers and processors during the research and development stage of the autonomous driving vehicle system.

[0003] In the prior art, the training and debugging of the transfer process of the collected data are mainly completed manually, resulting in a large amount of human participation in processes such as data parsing, handling, preprocessing, recycling, and auditing involved in the transfer process of the collected data. The debugging efficiency is relatively low, and due to human limitations, it is easy to have problems such as the transfer process of some collected data not being effectively configured and the functions of some processors or servers not being utilized, resulting in low configuration efficiency and data processing efficiency. At the same time, it is impossible to monitor vehicle abnormal behaviors in real time and effectively, adding unnecessary risks to the driving of the vehicle. Summary of the Invention

[0004] The embodiments of the present application provide a method, device, equipment, and storage medium for processing data of an autonomous driving vehicle to solve the problems of low debugging efficiency and low data processing efficiency in the training and debugging of the transfer process of the collected data.

[0005] In a first aspect, the embodiments of the present application provide a method for processing data of an autonomous driving vehicle, which is applied to a storage server. The method for processing data of an autonomous driving vehicle includes:

[0006] Obtain data to be processed through a first analysis server and cache the data to be processed;

[0007] Based on the data processing requirement information corresponding to the data to be processed, determine at least one matching processing server;

[0008] Send the data to be processed to at least one processing server for processing respectively.

[0009] It can be seen that by obtaining the data to be processed from the first analysis server and caching it, and then determining at least one matching processing server based on the data processing requirement information corresponding to the data to be processed, and sending the data to be processed to at least one processing server for processing respectively. Thus, the storage server can connect the first analysis server and multiple processing servers to jointly complete the processing requirements of the data to be processed; at the same time, the first analysis server preprocesses the data to be processed, effectively improving the readability and accessibility of the data cached in the storage server, facilitating subsequent processing. At the same time, based on the data processing requirement information of the data to be processed, the corresponding processing server can be automatically determined, directly completing the transfer of the data to be processed between the storage server and the corresponding processing server. The transfer of the data to be processed between cloud servers does not require manual participation and temporary configuration, reducing the situation where some server functions are not used, thereby effectively ensuring the efficiency of data transfer and processing.

[0010] Optionally, the storage server includes storage areas corresponding to at least one type of data. Obtaining the data to be processed from the first analysis server and caching the data to be processed includes: in response to receiving the data to be processed and its type sent by the first analysis server, the type of the data to be processed is identified by the first analysis server; based on the type, caching the data to be processed into the corresponding storage area.

[0011] It can be seen that by the first analysis server preprocessing and analyzing the data to be processed in advance, the type corresponding to the data to be processed is determined, and its corresponding data processing requirement information is determined, so as to send different types of data to be processed to the corresponding processing servers for processing based on the data processing requirement information, improving the data processing efficiency, and being able to store the corresponding type of data after being processed by the processing server in different storage areas respectively, ensuring the stable progress of the data processing process and the transfer process, and further improving the robustness of the server system for data processing.

[0012] Optionally, the data to be processed includes at least one processing step and the corresponding processing flow; determining at least one matching processing server based on the data processing requirement information corresponding to the data to be processed includes: determining the at least one processing step and the corresponding processing flow corresponding to the data to be processed based on the data processing requirement information corresponding to the data to be processed; respectively determining the processing server matching the step based on the data processing requirement information corresponding to each processing step of the data to be processed; sending the data to be processed to at least one processing server for processing respectively includes: sequentially sending the data to be processed to the processing server matching the data processing requirement information corresponding to the processing step according to the processing flow corresponding to at least one processing step.

[0013] It can be seen that according to the data processing requirement information of the data to be processed and the preset processing strategy, the processing steps corresponding to the overall data processing requirement information of the data to be processed and the processing server corresponding to each processing step can be determined. Thus, the process of automatically determining the processing server that needs to be matched according to the characteristics of the data to be processed itself can be realized, ensuring that the processing server is effectively configured, and further ensuring the data processing efficiency.

[0014] Optionally, the data processing requirement information includes one or several combinations of the following: data status requirement information, set format requirement information, label requirement information, vehicle status information, format and signal quality information.

[0015] It can be seen that by configuring a variety of different data processing requirement information and processing the data to be processed from different dimensions, it is ensured that the data to be processed is comprehensively processed, reducing the possible omission problems when manually configuring the processing strategy, ensuring the availability of the processed data, and improving the data processing efficiency.

[0016] Optionally, according to the processing flow corresponding to at least one processing step, the data to be processed is sequentially sent to the processing server that matches the data processing requirement information corresponding to the processing step, including: for each step, if the processing server that matches the data processing requirement information corresponding to the step is a data cleaning server, then the data to be processed is sent to the data cleaning server based on the processing flow, so that the data cleaning server can identify and delete the abnormal data in the step according to the data status requirement information corresponding to the step, and receive the processed step; or, if the processing server that matches the data processing requirement information corresponding to the data of the step is a data cleaning server, then the data to be processed is sent to the data cleaning server based on the processing flow, so that the data cleaning server can perform format conversion processing on the data to be processed according to the set format requirement information corresponding to the step, and receive the processed data to be processed; or, if the processing server that matches the data processing requirement information corresponding to the data of the step is a data annotation server, then the data to be processed is sent to the data annotation server based on the processing flow, so that the data annotation server can add label processing to the data to be processed according to the label requirement information corresponding to the step, and receive the processed data to be processed; or, if the processing server that matches the data processing requirement information corresponding to the data of the step is a driving data processing server, then the data to be processed is sent to the driving data processing server based on the processing flow, so that the driving data processing server can perform annotation processing on the data to be processed according to the format and signal quality information corresponding to the step, and receive the processed data to be processed; or, if the processing server that matches the data processing requirement information corresponding to the data of the step is a second analysis server, then the data to be processed is sent to the second analysis server based on the processing flow, so that the second analysis server can store the data to be processed in the form of a data sequence according to the vehicle status information corresponding to the step and the label of the data to be processed.

[0017] It can be seen that according to the different data processing requirement information of the data to be processed, the storage server can automatically send the data to be processed to the corresponding processing server, and perform processing based on the corresponding data processing requirement information. The processed data to be processed can be returned to the storage server for caching for subsequent further processing, or the processed data to be processed can be sent to servers such as the second analysis server for storage, realizing rich and diverse processing flows and transfer methods, fully meeting the data processing requirements of autonomous driving vehicles, and ensuring data processing efficiency.

[0018] Optionally, the processing server further includes a data training server for model training. After the data to be processed is respectively sent to at least one processing server for processing, it further includes: if the data processing requirement information further includes data training information, the processed data to be processed is sent to the data training server corresponding to the data training information, so that the data training server can perform model training according to the data training information and the processed data to be processed.

[0019] It can be seen that the processed data to be processed stored in the storage server can also be sent to the data training server for direct model training, without the need for manual step-by-step processing of the data for model training, significantly improving the efficiency of model training based on autonomous driving vehicle data. And since the data used for training has been pre-processed by multiple processing servers, the integrity and availability of the data are fully guaranteed, thus ensuring the data training effect.

[0020] Optionally, the processing server further includes a data scheduling server for function development scheduling. After the data to be processed is respectively sent to at least one processing server for processing, it further includes: if the data processing requirement information further includes application requirement information, the processed data to be processed is sent to the data scheduling server corresponding to the application requirement information, so that the data scheduling server can execute a data calling task according to the application requirement information and the processed data to be processed.

[0021] It can be seen that by configuring the application requirement information, the processed data to be processed can be directly provided to the server for querying and scheduling, so as to search, schedule and use various processed data according to the specific application requirements on the autonomous driving vehicle, thus completing the whole process of uploading the data to be processed from the vehicle control unit to obtaining and applying it in the slave server, solving the problem of long-term residence of data in a single server and troublesome scheduling, and improving the data flow efficiency.

[0022] In a second aspect, an embodiment of the present application provides an autonomous driving vehicle data processing device, which is applied to a storage server and includes:

[0023] A receiving module, configured to obtain the data to be processed through a first analysis server and cache the data to be processed;

[0024] A matching module, configured to determine at least one matching processing server based on the data processing requirement information corresponding to the data to be processed;

[0025] A transmission module, configured to respectively send the data to be processed to the at least one processing server for processing.

[0026] Optionally, the receiving module is specifically configured to, if there is a storage area corresponding to at least one type of data in the storage server, in response to the to-be-processed data and its type sent by the first analysis server, where the type of the to-be-processed data is identified by the first analysis server; based on the type, cache the to-be-processed data into the corresponding storage area.

[0027] Optionally, the matching module is specifically configured to, if the to-be-processed data includes at least one processing step and the corresponding processing flow, determine the at least one processing step and the corresponding processing flow corresponding to the to-be-processed data based on the data processing requirement information corresponding to the to-be-processed data; respectively determine the processing server that matches the step based on the data processing requirement information corresponding to each processing step of the to-be-processed data; the transmission module is specifically configured to, according to the processing flow corresponding to at least one processing step, sequentially send the to-be-processed data to the processing server that matches the data processing requirement information corresponding to the processing step.

[0028] Optionally, the matching module specifically includes that the data processing requirement information includes one or several combinations of the following: data status requirement information, set format requirement information, label requirement information, vehicle status information, format and signal quality information.

[0029] Optionally, the transmission module is specifically configured to, for each step, if the processing server matching the data processing requirement information corresponding to the step is a data cleaning server, send the data to be processed to the data cleaning server based on the processing flow, so that the data cleaning server can identify and delete abnormal data in the step according to the data status requirement information corresponding to the step, and receive the processed step; or, if the processing server matching the data processing requirement information corresponding to the data in the step is a data cleaning server, send the data to be processed to the data cleaning server based on the processing flow, so that the data cleaning server can perform format conversion processing on the data to be processed according to the set format requirement information corresponding to the step, and receive the processed data to be processed; or, if the processing server matching the data processing requirement information corresponding to the data in the step is a data annotation server, send the data to be processed to the data annotation server based on the processing flow, so that the data annotation server can add label processing to the data to be processed according to the label requirement information corresponding to the step, and receive the processed data to be processed; or, if the processing server matching the data processing requirement information corresponding to the data in the step is a driving data processing server, send the data to be processed to the driving data processing server based on the processing flow, so that the driving data processing server can perform annotation processing on the data to be processed according to the format and signal quality information corresponding to the step, and receive the processed data to be processed; or, if the processing server matching the data processing requirement information corresponding to the data in the step is a second analysis server, send the data to be processed to the second analysis server based on the processing flow, so that the second analysis server can store the data to be processed in the form of a data sequence according to the vehicle status information corresponding to the step and the label of the data to be processed.

[0030] Optionally, the transmission module is further configured to, if the processing server further includes a data training server for model training, after sending the data to be processed to at least one processing server for processing, if the data processing requirement information further includes data training information, send the processed data to be processed to the data training server corresponding to the data training information, so that the data training server can perform model training according to the data training information and the processed data to be processed.

[0031] Optionally, the transmission module is further configured to, if the processing server further includes a data scheduling server for function development scheduling, after sending the data to be processed to at least one processing server for processing, if the data processing requirement information further includes application requirement information, send the processed data to be processed to the data scheduling server corresponding to the application requirement information, so that the data scheduling server can execute a data calling task according to the application requirement information and the processed data to be processed.

[0032] In a third aspect, an embodiment of the present application further provides a control device, which includes:

[0033] At least one processor;

[0034] And a memory communicatively connected to the at least one processor;

[0035] Wherein, the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the control device to execute the autonomous driving vehicle data processing method corresponding to any one of the embodiments in the first aspect of the present application.

[0036] In a fourth aspect, an embodiment of the present application further provides an autonomous driving vehicle data processing system, and the autonomous driving vehicle data processing system includes:

[0037] A first analysis server, configured to receive data to be processed from a vehicle control unit and perform classification processing;

[0038] A storage server, configured to obtain the classified data to be processed from the first analysis server and send it to at least one processing server for processing based on the data processing requirement information corresponding to the data to be processed;

[0039] At least one processing server, configured to process the data to be processed based on the data processing requirement information.

[0040] It can be seen that through the combination of the first analysis server, the storage server and at least one processing server, the data generated by the autonomous driving vehicle is automatically transferred and processed in the cloud, fully ensuring the data transfer efficiency, while solving the problem of data staying in a single server for a long time, effectively ensuring that the data is processed quickly and comprehensively, and further ensuring the R & D efficiency of the autonomous driving vehicle system.

[0041] In a fifth aspect, an embodiment of the present application further provides a computer-readable storage medium, and the computer-readable storage medium stores computer-executable instructions, and when the computer-executable instructions are executed by a processor, they are used to implement any one of the autonomous driving vehicle data processing methods in the first aspect of the present application.

[0042] In a sixth aspect, an embodiment of the present application further provides a computer program product, and the program product includes computer-executable instructions, and when the computer-executable instructions are executed by a processor, they are used to implement the autonomous driving vehicle data processing method corresponding to any one of the embodiments in the first aspect of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 It is an application scenario diagram of the autonomous driving vehicle data processing method provided by an embodiment of the present application;

[0044] Figure 2Flowchart of the method for processing data of an autonomous vehicle provided by an embodiment of the present application;

[0045] Figure 3a Flowchart of the method for processing data of an autonomous vehicle provided by another embodiment of the present application;

[0046] Figure 3b For Figure 3a Flowchart of the method for determining a processing server provided in the illustrated embodiment;

[0047] Figure 4 Schematic structural diagram of an apparatus for processing data of an autonomous vehicle provided by another embodiment of the present application;

[0048] Figure 5 Schematic structural diagram of a control device provided by another embodiment of the present application;

[0049] Figure 6a Schematic structural diagram of a system for processing data of an autonomous vehicle provided by another embodiment of the present application;

[0050] Figure 6b For Figure 6a Schematic diagram of an application scenario for storing data in the cloud provided in the illustrated embodiment;

[0051] Figure 6c For Figure 6a Schematic diagram of an application scenario for sorting and storing data in the cloud provided in the illustrated embodiment;

[0052] Figure 6d For Figure 6a Schematic diagram of an application scenario for using data in the cloud for application development provided in the illustrated embodiment. Detailed implementation manners

[0053] Here, exemplary embodiments will be described in detail, and examples thereof are shown in the drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The implementation manners described in the following exemplary embodiments do not represent all implementation manners consistent with the embodiments of the present application. On the contrary, they are merely examples of devices and methods consistent with some aspects of the embodiments of the present application as detailed in the appended claims.

[0054] The technical solutions of the embodiments of the present application and how the technical solutions of the embodiments of the present application solve the above technical problems will be described in detail below with specific embodiments. The following several specific embodiments may be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of the present application will be described below with reference to the drawings.

[0055] With the emergence of diversified transportation needs, technologies such as autonomous driving and unmanned driving are gradually being more widely used. During the driving process, autonomous vehicles will continuously collect a large amount of environmental data and vehicle status data through their own sensors and monitoring equipment. These environmental data and vehicle status data will flow between the vehicle's own control unit and different cloud servers through different methods and processes to complete real-time data processing and control the vehicle's own driving status. In order to process these different types of data, different types of processing servers are usually configured during the research and development stage of autonomous vehicles to process these data in corresponding ways.

[0056] However, due to the complex types of data involved in autonomous vehicles and the processing requirements of different data, there are usually multiple different data flow processes at the same time during the research and development and application stages. When some servers or businesses need to be updated, the business corresponding data or the server corresponding process is prone to data loss, and after multiple updates, some servers may not be fully utilized and some functions may require repeated manual configuration by personnel, especially those involving data analysis, transportation, preprocessing, recycling, and review. After the server and business are updated, a large amount of manpower is usually required to debug the server and data flow process, and the debugging efficiency is low. Due to manpower limitations, it is easy for the flow process of some data to not be effectively configured, and the functions of some processors or servers to not be used. The configuration efficiency and data processing efficiency are low, which leads to insufficient system safety of autonomous vehicles.

[0057] In order to solve the above problems, an embodiment of the present application provides a method for processing data of an autonomous driving vehicle, which realizes the automatic flow and processing of data generated by the autonomous driving vehicle through the combination of different servers, ensuring that the data corresponding to the autonomous driving vehicle can always be effectively processed, thereby effectively improving the system safety of the autonomous driving vehicle.

[0058] Figure 1 This is an application scenario diagram of the autonomous driving vehicle data processing method provided in the embodiment of the present application. Figure 1 As shown, during the autonomous driving vehicle data processing process, the vehicle control unit 100 of the vehicle sends the unprocessed data that needs to be processed in the cloud to the cloud server 110, so that each cloud server 110 performs flow processing based on the unprocessed data, thereby completing the autonomous driving vehicle data processing process.

[0059] It should be noted that Figure 1 In the scenario shown, only one vehicle control unit and one cloud server are used as an example for illustration, but the embodiments of the present application are not limited to this, that is, the number of vehicle control units and cloud servers can be arbitrary.

[0060] The data processing method for autonomous driving vehicles provided by the present application will be described in detail below through specific embodiments. It should be noted that the following several specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments.

[0061] Figure 2 The following is a flowchart of the data processing method for autonomous driving vehicles provided by an embodiment of the present application. The data processing method for autonomous driving vehicles provided by this embodiment is applied to a storage server, such as Figure 2 shown, and it includes but is not limited to the following steps:

[0062] Step S201: Obtain the data to be processed through the first analysis server and cache the data to be processed.

[0063] Specifically, the storage server is used to cache the data to be processed uploaded by the autonomous driving vehicle, and as a heatstroke unit in the cloud, cache the data to be processed, and distribute the data to be processed to the corresponding processing server for processing (such as cleaning the data, eliminating errors, annotating, storing records, etc.), then receive the processed data from the processing server (such as receiving the annotated or specified operation data), and send it to the destination corresponding to the data to be processed (such as the vehicle end or the server for storage).

[0064] Thereby, it is convenient to decouple data storage and data processing in the cloud, and it is convenient to update different servers separately without affecting data flow (because the data flow is no longer controlled by the server for data processing), which not only ensures the efficiency of data processing, but also ensures the stability and reliability in the data flow process.

[0065] The first analysis server is generally a distributed cloud data analysis server, which is used to obtain the data to be processed that needs to be processed in the cloud from the vehicle control unit of the autonomous driving vehicle, analyze and process the data, and then transmit it to the storage server.

[0066] The vehicle control unit of the autonomous driving vehicle will directly process the data that only needs to be processed locally according to the pre-configuration, and upload the data that needs to be processed by the cloud server to the first analysis server for processing through the cloud server. This part of the data is the data to be processed.

[0067] In some embodiments, the processing of the data to be processed by the first analysis server may be classification processing. For example, according to the purpose, it can be determined whether the data needs to be stored or the data to be returned to the vehicle terminal. It can also be distinguished according to the corresponding process of the data to be processed, depending on the different processes required and the corresponding processing servers. It can also be distinguished according to different data sources, such as different types of environmental data collected (such as positioning data, radar sensor data, captured video image data) and vehicle status data (such as driving speed, engine speed, etc.).

[0068] In some embodiments, the processing of the data to be processed by the first analysis server may preprocess the data, that is, determine the data processing requirement information based on the type of the data to be processed. The data processing requirement information corresponding to each type of data to be processed can be preconfigured by developers. When the first analysis server receives the data to be processed, it can determine its type according to the message of the data to be processed, and then further configure the corresponding data processing requirement information and send it to the storage server to complete the transfer and processing of the data to be processed according to the data processing requirement information.

[0069] Step S202: Based on the data processing requirement information corresponding to the data to be processed, determine at least one matching processing server.

[0070] Specifically, the data processing requirement information is used to record the processing methods required for the data to be processed, such as methods like adding annotations, modifying formats, normalizing, etc., as well as the destinations for sending, such as returning to the vehicle terminal, sending to a database server for storage, a server for querying, etc. Through the data processing requirement information, the transfer of the data to be processed can be decoupled from the specific processing server. As long as the data processing requirement information remains unchanged, even if the processing server is updated, it will not affect the normal transfer of the data to be processed. Similarly, when the data to be processed needs to be changed, by adjusting its corresponding data processing requirement information, it can ensure that after data iteration and adjustment, the data transfer can still be effectively completed.

[0071] According to the processing method in the data processing requirement information, the storage server can directly determine the processing server corresponding to this processing method, so as to send the data to be processed to this processing server for processing.

[0072] In some embodiments, when there are multiple processing methods in the data processing requirement information, a priority order of multiple processing methods is configured in the storage server to send them to different processing servers for processing in sequence.

[0073] In some embodiments, the data processing requirement information corresponding to different types of data to be processed records the processing order of different processing methods (for example, for the same tag addition processing, it may be required to add a signal quality tag first and then adjust the format. At this time, the format of the signal quality tag will also be adjusted, but it may also be the opposite, in which case the format of the signal quality tag does not need to be adjusted), so that the storage server can sequentially send it to different processing servers for processing based on this processing order.

[0074] In some embodiments, the data processing requirement information is pre-configured by the user and is allocated and added by the first analysis server when receiving the data to be processed (added to the message corresponding to the data to be processed). By pre-configuring the data processing requirement information by the user, personalized configuration and scheduling of the data flow process are realized, which is convenient for improving the efficiency of data processing and flow.

[0075] In some embodiments, the data processing requirement information can also be added by the storage server to the data to be processed based on the classification to which the data to be processed belongs when receiving the data to be processed classified by the first analysis server, so as to transfer the data to be processed based on the data processing requirement information.

[0076] In some embodiments, the data processing requirement information can also be directly configured in the storage server instead of being added to the data to be processed. At this time, the storage server sends the data to be processed to the corresponding processing server for processing based on the configured data processing requirement information, and sends the corresponding data processing requirement information to the corresponding processing server, rather than adding all the data processing requirement information corresponding to the data to be processed to the message of the data to be processed.

[0077] In some embodiments, the data processing requirement information can also be configured in different processing servers. When the processing server needs to obtain the corresponding type of data to be processed from the storage server, it sends the data processing requirement information including the type of data to be processed to the storage server, so that the storage server can send the corresponding data to be processed to the corresponding processing server based on the received data processing requirement information.

[0078] Step S203: Send the data to be processed to at least one processing server for processing.

[0079] Specifically, if the data processing requirement information records the sending destination of the data to be processed, and the processing server is not the destination, the processing server will return the processed data to the storage server for reprocessing or send it to the corresponding destination.

[0080] In some embodiments, the storage server stores the data at various stages of processing (such as unprocessed data to be processed, processed data to be sent, and data being processed) in different partitions, thereby facilitating the flow and scheduling of the data to be processed and improving the efficiency and reliability of data scheduling.

[0081] The method for processing autonomous driving vehicle data provided by the embodiments of the present application obtains the data to be processed from the first analysis server and caches it, and then determines at least one matching processing server based on the data processing requirement information corresponding to the data to be processed, and sends the data to be processed to at least one processing server for processing respectively. Thus, the storage server can connect the first analysis server and multiple processing servers to jointly complete the processing requirements of the data to be processed; at the same time, the first analysis server preprocesses the data to be processed, effectively improving the readability and accessibility of the data cached in the storage server, facilitating subsequent processing. At the same time, through the data processing requirement information of the data to be processed, the corresponding processing server can be automatically determined, directly completing the transfer of the data to be processed between the storage server and the corresponding processing server. The transfer of the data to be processed between the cloud servers does not require manual participation and temporary configuration, reducing the situation where some server functions are not used, thereby effectively ensuring the efficiency of data transfer and processing.

[0082] Figure 3a It is a flowchart of a method for processing autonomous driving vehicle data provided by an embodiment of the present application. As Figure 3a shown, the method for processing autonomous driving vehicle data provided in this embodiment includes the following steps:

[0083] Step S301: Respond to the data to be processed and its type sent by the first analysis server.

[0084] Among them, the type of the data to be processed is obtained by the first analysis server through identification.

[0085] Specifically, when receiving the data to be processed, the first analysis server can determine its corresponding type by analyzing the message of the data to be processed, so as to send the data to be processed to the corresponding storage area of the storage server according to its type for classification processing.

[0086] In some embodiments, the first analysis server is configured with tools such as an Internet of Things data receiving end, a data analysis service suite, and a distributed search analysis engine to receive the data to be processed (such as data on pressure changes in the engine that require cloud computing and analysis) from the vehicle control unit interface of the autonomous driving vehicle through the Internet of Things data receiving end, and realize data connection between the vehicle and the cloud, so as to transmit the data to be processed to the storage server. The storage server stores the data in the form of a data lake and completes the data flow; through the combination of the data analysis service suite and the distributed search analysis engine, the classification analysis of the data to be processed is completed, and the results are transmitted to the storage server.

[0087] Step S302: based on the type, cache the data to be processed into a corresponding storage area.

[0088] The storage server includes a storage area corresponding to at least one type of data.

[0089] Specifically, the storage server will save the data to be processed in a corresponding location (such as a database form) according to the type of data to be processed, and allocate corresponding data processing requirement information according to the type of data to be processed (such as each database form corresponds to a data processing requirement information file to record the data processing requirement information of the corresponding type of data to be processed) for subsequent targeted processing.

[0090] In some embodiments, a dedicated regular data partition is provided in the storage server to store the to-be-processed data received from the first analysis server, and other data partitions are also provided to store the to-be-processed data at different processing stages.

[0091] In some embodiments, the storage area of ​​the storage server includes a regular data partition, a low-quality data partition, a parsed data partition, a processed data partition, and a marked information data partition.

[0092] Specifically, the regular data partition and the low-quality data partition are used to receive classified data to be processed sent from the first analysis server. At this time, in addition to distinguishing the types of data to be processed, the first analysis server will also distinguish them into regular data and low-quality data according to the overall quality of the corresponding types of data (such as data with missing items or abnormal values ​​exceeding a set proportion, it will be considered as low-quality data), and send them to the corresponding data partitions respectively. Before the data in the low-quality data partition is further processed according to the data processing requirement information, it is necessary to perform data cleaning, completion and other operations to eliminate low-quality data parts, or complete them to ensure the normal progress of subsequent processing procedures.

[0093] The parsed data partition and the processed data partition are used to store data at different processing stages. For example, the processed data partition stores data that has not been fully processed (such as wheel speed data whose numerical format has only been adjusted but no labels have been added yet), and the parsed data partition is used to store data that has been processed (such as engine fuel consumption data with weather and temperature labels added).

[0094] The annotation information data partition is used to store data that needs to be annotated, or separate information for adding annotations (such as weather, coordinates, temperature and humidity, etc.).

[0095] The storage server stores and schedules the data to be processed, decoupling the storage service from the data processing according to the service, so as to independently design the data storage and data processing, ensuring both the high efficiency of data processing and providing an important guarantee for the timeliness of input and output during data flow.

[0096] Especially for the processing network and the stability of the output end of the remote transmission data when vehicle control data is sent to the cloud, there is a certain requirement. In addition, the storage server can receive the data to be processed as soon as possible and write it in time to ensure the non-loss of data, thereby ensuring the complete operation of the process. The cloud cluster method based on the storage server has strong scalability for the business to be expanded.

[0097] Step S303: Based on the data processing requirement information corresponding to the data to be processed, determine at least one processing step and the corresponding processing flow corresponding to the data to be processed.

[0098] Specifically, the storage server will cooperate with open-source workflow management tools to achieve the scheduling of the data processing flow based on the data processing requirement information. Open-source workflow management tools (such as existing management tools like osworkflow and jbpm) can determine the processing methods required for the data to be processed (such as format correction, adding annotations, etc.) according to the data processing requirement information, and determine the corresponding processing steps (such as format correction first, then adding annotations) and processing flow (such as the first step is format correction, and the second step is to add annotations after receiving) based on the processing methods.

[0099] In some embodiments, the data processing requirement information includes the processing order of each processing method, so that the open-source workflow management tool can determine the processing steps and processing flow based on the processing order. At this time, the open-source workflow management tool will sequentially extract the corresponding data to be processed from the corresponding storage area according to the processing flow and transmit it to different processing servers to complete the corresponding processing flow and realize the data flow.

[0100] Furthermore, the data processing requirement information includes one or several of the following combinations:

[0101] Data status requirement information, setting format requirement information, label requirement information, vehicle status information, format and signal quality information.

[0102] Specifically, the data status requirement information, that is, the information related to the required data status, such as the need to find abnormal data (such as the wheel hub temperature data, and the need to find the jump value that suddenly changes from positive to negative), or the need to eliminate some data in advance, or find the extreme value data for subsequent processing.

[0103] The setting format requirement information, that is, the information related to the required data format, such as the format requirements for digital-related data, or the need to convert the data in the message into data in a specific format (such as a specific number of spaces are included between messages).

[0104] The label requirement information, that is, the information related to the labels that need to be added, such as the need to add labels such as positioning, time, weather, etc.

[0105] The vehicle status information, that is, the information related to the vehicle status data that needs to be ensured to be included. For example, if the tire pressure data of a certain wheel of the vehicle is lacking in the data to be processed, it is necessary to improve this data based on the data collected by other sensors.

[0106] The format and signal quality information, that is, the format version information and signal quality information that need to be added to the data to be processed (such as vehicle positioning information, driving speed information, etc., and signal quality information is required to ensure the confidence level), such as the version of the driving data and the signal quality at the time of collection, etc. This type of information may not be stored in the data to be processed, but requires the corresponding processing server to obtain the corresponding data to add to the data to be processed.

[0107] Step S304: Based on the data processing requirement information corresponding to each processing step of the data to be processed, determine the processing server that matches the step.

[0108] Specifically, usually each processing step will have a corresponding processing server to process based on the corresponding data processing requirement information through this processing server.

[0109] In some embodiments, there may be multiple processing servers that perform the same type of function. For example, servers that perform annotation in different ways (such as point cloud annotation and image annotation can be completed by different processing servers), and there may also be multiple processing servers that perform the same function (such as there may be multiple processing servers for model training, corresponding to different models).

[0110] Further, as Figure 3b shown, it is a flowchart of the method for determining the processing server, which includes the following steps:

[0111] Step S3041: If the processing server matching the data processing requirement information corresponding to the step is a data cleaning server, send the data to be processed to the data cleaning server based on the processing flow, so that the data cleaning server can identify and delete the abnormal data in the step according to the data status requirement information corresponding to the step, and receive the processed step.

[0112] Specifically, the data cleaning server is used to process data deletion and format conversion. When it is necessary to find abnormal data in the data (such as deleting the values that become 0 in the cumulative mileage data of vehicle driving) and add formats, the data cleaning server needs to process these data to be processed.

[0113] In some embodiments, the data cleaning server provides a data visualization tool, a marking tool, and a database tool. The data visualization tool can facilitate managers to view the data processed therein; the marking tool is used to identify and find abnormal data or data that needs to be formatted for processing; the database tool is used to perform data deletion processing or format conversion processing.

[0114] Step S3042: If the processing server matching the data processing requirement information corresponding to the step is a data cleaning server, send the data to be processed to the data cleaning server based on the processing flow, so that the data cleaning server can perform format conversion processing on the data to be processed according to the set format requirement information corresponding to the step, and receive the processed data to be processed.

[0115] Specifically, the data cleaning server can receive low-quality data and unstructured data sent by the storage server (sometimes including a modification request for database information to correspondingly adjust and modify the data to be processed, such as unifying the two formats representing negative values, -1.0 and negative 1.0).

[0116] Step S3043: If the processing server matching the data processing requirement information corresponding to the step is a data annotation server, send the data to be processed to the data annotation server based on the processing flow, so that the data annotation server can add label processing to the data to be processed according to the label requirement information corresponding to the step, and receive the processed data to be processed.

[0117] Specifically, the data annotation server is used to identify, annotate, etc. the data to be processed based on the data processing requirement information.

[0118] In some embodiments, the data annotation server provides a point cloud recognition and annotation tool, an image recognition and annotation tool, and a fusion annotation tool. The point cloud recognition and annotation tool can recognize and annotate point cloud data in the image-like data to be processed. The image recognition and annotation tool can recognize and annotate information such as objects and image features in the image / video-like data to be processed and perform annotations. The fusion annotation tool can provide recognition and annotation of some other relevant information (such as image quality, video length, position relative to the vehicle, etc.).

[0119] Through the combination of various tools in the data annotation server, the information recognition, annotation, and other processing of the data to be processed can be completed.

[0120] In some embodiments, the data annotation server can also obtain information such as GPS information, timestamps, and weather through tools such as data parsing services, weather acquisition applications, and environmental information databases, add corresponding tags to the data to be processed, enrich the data content, and thus obtain parsed data or semi-processed data to be processed (at this time, the data to be processed is usually semi-structured data).

[0121] Step S3044: If the processing server matching the data processing requirement information corresponding to the step is a driving data processing server, then send the data to be processed to the driving data processing server based on the processing flow, so that the driving data processing server can perform annotation processing on the data to be processed according to the format and signal quality information corresponding to the step, and receive the processed data to be processed.

[0122] Specifically, the driving data processing server is also a processing server used to enrich and increase the data to be processed.

[0123] In some embodiments, the driving data processing server includes a container orchestration engine (such as the EKS tool, Elastic Kubernetes Service, a commonly used existing container orchestration tool) to receive the semi-processed data, perform further image processing or signal depth verification, and add signal quality tags based on the signal depth verification results (for example, when the vehicle is driving in a tunnel, the vehicle status data and environmental data transmitted back may be distorted due to the influence of the tunnel, so it is necessary to determine the corresponding signal quality), or output the data after image processing. Through the container orchestration engine, the processing units arranged in the driving data processing server are controlled to adjust the processing efficiency to ensure the efficiency of image processing and signal depth verification.

[0124] In some embodiments, the driving data processing server includes a driving and traffic processor for recording information related to driving data. Through the format of the dynamic content file in the driving and traffic processor, the required metadata can be extracted from the data to be processed such as vehicle state data and returned to the storage server for storage, or sent to the second analysis server for subsequent search, analysis, etc.

[0125] Step S3045: If the processing server matching the data processing requirement information of the corresponding data in the step is the second analysis server, then send the data to be processed to the second analysis server based on the processing flow, so that the second analysis server stores the data to be processed in the form of a data sequence according to the vehicle state information corresponding to the step and the label of the data to be processed.

[0126] Specifically, the second analysis server is used to further analyze, process, and / or store the processed or semi-processed data to be processed. For example, based on the data labels added in step S3043 and / or step S3044, analyze the relationship between the data to be processed to re-store the data to be processed in an association relationship based on the label, or perform further feature analysis based on the label (such as combining the environmental data and vehicle state data after adding the label to calculate an integrated index to evaluate the vehicle's own state).

[0127] In some embodiments, the second analysis server includes a graph database for storing data, an ELT management server for scheduling data for querying and processing (ELT means extract, load, and transform, that is, extraction, loading, and transformation operations), and a dynamic database for calling and processing data. Through the cooperation of each tool, the analysis, processing, scheduling, and storage of the received data to be processed are realized, such as the recognition, analysis, sorting, and storage of environmental images / videos collected by the image sensors on the vehicle.

[0128] By using the graph database service, a storage data sequence corresponding to the processed data based on information such as labels can be constructed and queried. The data in this storage data sequence is highly interconnected based on information such as labels, and the ETL management service can be used for the data integration service based on the serverless architecture within the second analysis server, and a data catalog function can be provided to improve the query efficiency of the data.

[0129] In some embodiments, the second analysis server may further include a database analysis service, a RESTful distributed search engine, and an analysis engine (REST stands for Representational State Transfer in English, which means representational state transfer. A RESTful distributed search engine and analysis engine can better process and schedule cloud data according to data processing requirement information without frequent interaction with the vehicle control unit of an autonomous vehicle, improving processing efficiency), and a metadata format processor that cooperates with the database analysis service to perform internal format processing.

[0130] Step S305: According to the processing flow corresponding to at least one processing step, sequentially send the data to be processed to the processing server that matches the data processing requirement information corresponding to the processing step.

[0131] Specifically, according to the processing flow, the storage server will sequentially send the data to be processed and the data processing requirement information to the processing server corresponding to each processing flow, so that the processing server can process the data based on the data processing requirement information.

[0132] Step S306: If the data processing requirement information further includes data training information, send the processed data to be processed to the data training server corresponding to the data training information for the data training server to perform model training based on the data training information and the processed data to be processed.

[0133] Among them, the processing server further includes a data training server for model training.

[0134] Specifically, after completing the processing of the data to be processed through the foregoing steps, usable data, that is, the processed data to be processed, can be obtained. The data at this time can be parsed data (i.e., data that has completed processing) or structured data (i.e., data that meets the format, structure, and other characteristics required for subsequent applications). Thus, the processed data to be processed can be used in scenarios such as model training, query and invocation, and application development.

[0135] If it is necessary to use the processed data to be processed in the scenario of model training, usually the corresponding data processing requirement information will include data training information, so that the storage server can send the corresponding processed data to be processed to the data training server for training based on the requirement information to perform model training work (such as training a model for judging whether there is a pedestrian in front of the vehicle from the video data collected by an image sensor).

[0136] In some embodiments, the data training information for model training may not be recorded by the storage server and sent to the data training server. Instead, when the data training server needs to obtain data samples, it sends data training information to the storage server to obtain specified processed data to be processed as sample data for model training.

[0137] In some embodiments, the data training server includes a tagging tool for constructing a training data set (to further tabulate data that needs to be machine-learned), a machine learning function set for model training, and can also provide a Web-based unified visualization interface so that data developers can cooperate with the data training server through a terminal to quickly prepare, construct, train, and deploy high-quality machine learning models.

[0138] Step S307: If the data processing requirement information further includes application requirement information, send the processed data to be processed to the data scheduling server corresponding to the application requirement information for the data scheduling server to execute a data call task according to the application requirement information and the processed data to be processed.

[0139] Among them, the processing server further includes a data scheduling server for function development scheduling.

[0140] Specifically, if the processed data to be processed needs to be used for regular queries, function development, or application calls (such as facilitating remote monitoring of vehicle driving status, location, etc.), there is a need to send the data in the storage server to the data scheduling server for subsequent calls. At this time, the storage server usually sends the corresponding data to the data scheduling server corresponding to the application requirement information based on the processed data to be processed corresponding to the application requirement information.

[0141] In some embodiments, the application requirement information for determining the data to be sent by the storage server can be sent by the data scheduling server to the storage server to obtain specified processed data to be processed and complete subsequent calls.

[0142] In some embodiments, the data scheduling server includes controls such as a graph database hosting service for facilitating data query, an event-driven service for scheduling data to execute application functions (such as generating a tire pressure alarm based on monitored tire pressure data and sending it to a remote user terminal), and a code resource calculation management service for cooperating with function development (such as recording the code for the seat heating function being developed and the seat temperature data at the same time) to meet different scheduling requirements of the data to be processed, and simplifies the development of applications related to querying / operating the graph database by handling connection tasks with data sources such as an event-driven and automatic code operation resource management calculation service platform.

[0143] The data processing method for an autonomous vehicle provided by an embodiment of this application caches the data to be processed in a corresponding storage area by responding to the data to be processed and its type sent by the first analysis server and based on the type. Then, based on the data processing requirement information corresponding to the data to be processed, at least one processing step and the corresponding processing flow corresponding to the data to be processed are determined, and the processing server matching the step is determined. Subsequently, the data to be processed is sequentially sent to the processing server that matches the data processing requirement information corresponding to the processing step. Then, the processed data to be processed is sent to the data training server corresponding to the data training information or sent to the data scheduling server corresponding to the application requirement information. Thus, the full-process automated transfer and scheduling processing of the data to be processed from the vehicle control unit to the cloud are realized, and the sending and receiving of the data to be processed can be automatically completed according to different types and states of the data to be processed. The characteristics of each server are fully utilized, the computing resources in the cloud are integrated, the configuration utilization rate is improved, the operations of each server can be non-interfering or non-coupled, the operation and maintenance workload of personnel is significantly reduced, and the overall resource utilization rate and system robustness are improved.

[0144] Figure 4 It is a schematic structural diagram of an autonomous vehicle data processing device provided by an embodiment of this application. As Figure 4 shown, the autonomous vehicle data processing device 400 is applied to a storage server and includes a receiving module 410, a matching module 420, and a transmission module 430. Among them:

[0145] The receiving module 410 is configured to obtain the data to be processed through the first analysis server and cache the data to be processed;

[0146] The matching module 420 is configured to determine at least one matching processing server based on the data processing requirement information corresponding to the data to be processed;

[0147] The transmission module 430 is configured to separately send the data to be processed to the at least one processing server for processing.

[0148] Optionally, the receiving module 410 is specifically configured to, if the storage server includes a storage area corresponding to at least one type of data, respond to the data to be processed and its type sent by the first analysis server, and the type of the data to be processed is identified by the first analysis server; based on the type, cache the data to be processed in the corresponding storage area.

[0149] Optionally, the matching module 420 is specifically configured to, if the data to be processed includes at least one processing step and a corresponding processing flow, determine the at least one processing step and the corresponding processing flow corresponding to the data to be processed based on the data processing requirement information corresponding to the data to be processed; respectively determine the processing server that matches the step based on the data processing requirement information corresponding to each processing step of the data to be processed. The transmission module 430 is specifically configured to sequentially send the data to be processed to the processing server that matches the data processing requirement information corresponding to the processing step according to the processing flow corresponding to at least one processing step.

[0150] Optionally, the matching module 420 specifically includes that the data processing requirement information includes one or several combinations of the following: data status requirement information, set format requirement information, label requirement information, vehicle status information, format and signal quality information.

[0151] Optionally, the transmission module 430 is specifically configured to, for each step, if the processing server that matches the data processing requirement information corresponding to the step is a data cleaning server, send the data to be processed to the data cleaning server based on the processing flow, so that the data cleaning server can identify and delete abnormal data in the step according to the data status requirement information corresponding to the step, and receive the processed step; or, if the processing server that matches the data processing requirement information corresponding to the data of the step is a data cleaning server, send the data to be processed to the data cleaning server based on the processing flow, so that the data cleaning server can perform format conversion processing on the data to be processed according to the set format requirement information corresponding to the step, and receive the processed data to be processed; or, if the processing server that matches the data processing requirement information corresponding to the data of the step is a data annotation server, send the data to be processed to the data annotation server based on the processing flow, so that the data annotation server can add label processing to the data to be processed according to the label requirement information corresponding to the step, and receive the processed data to be processed; or, if the processing server that matches the data processing requirement information corresponding to the data of the step is a driving data processing server, send the data to be processed to the driving data processing server based on the processing flow, so that the driving data processing server can perform annotation processing on the data to be processed according to the format and signal quality information corresponding to the step, and receive the processed data to be processed; or, if the processing server that matches the data processing requirement information corresponding to the data of the step is a second analysis server, send the data to be processed to the second analysis server based on the processing flow, so that the second analysis server can store the data to be processed in the form of a data sequence according to the vehicle status information corresponding to the step and the label of the data to be processed.

[0152] Optionally, the transmission module 430 is further configured to, if the processing server further includes a data training server for model training, after sending the data to be processed to at least one processing server for processing respectively, if the data processing requirement information further includes data training information, send the processed data to be processed to the data training server corresponding to the data training information, so that the data training server can perform model training according to the data training information and the processed data to be processed.

[0153] Optionally, the transmission module 430 is further configured to, if the processing server further includes a data scheduling server for function development scheduling, after sending the data to be processed to at least one processing server for processing respectively, if the data processing requirement information further includes application requirement information, send the processed data to be processed to the data scheduling server corresponding to the application requirement information, so that the data scheduling server can execute a data calling task according to the application requirement information and the processed data to be processed.

[0154] In this embodiment, the automatic driving vehicle data processing device can, through the combination of each module, solve the problems of low debugging efficiency in training and debugging during the data flow process of data collection and low data processing efficiency in the prior art, improve the overall resource utilization rate of the cloud server, and enhance the robustness of the cloud system.

[0155] Figure 5 The following is a schematic structural diagram of a control device provided by an embodiment of the present application, as Figure 5 shown. The control device 500 includes: a memory 510 and a processor 520.

[0156] Among them, the memory 510 stores a computer program that can be executed by at least one processor 520. The computer program is executed by at least one processor 520, so that the control device can implement the automatic driving vehicle data processing method provided in any one of the above embodiments.

[0157] Among them, the memory 510 and the processor 520 can be connected through a bus 530.

[0158] For relevant descriptions, reference can be made to the corresponding descriptions and effects in the method embodiments, which will not be elaborated here.

[0159] Figure 6a The following is a schematic structural diagram of an automatic driving vehicle data processing system provided by an embodiment of the present application, as Figure 6a shown. The automatic driving vehicle data processing system 600 includes:

[0160] A first analysis server 610, configured to receive data to be processed from the vehicle control unit and perform classification processing;

[0161] A storage server 620, configured to obtain the to-be-processed data after classification processing from the first analysis server and send it to at least one processing server for processing based on the data processing requirement information corresponding to the to-be-processed data;

[0162] At least one processing server 630, configured to process the to-be-processed data based on the data processing requirement information.

[0163] In some embodiments, the processing server 630 includes a data cleaning server 631, a data annotation server 632, a driving data processing server 633, a second analysis server 634, a data training server 635, and a data scheduling server 636. The functions and uses of the above-mentioned processing server 630 can refer to Figure 3a the corresponding descriptions in the illustrated embodiments, which will not be elaborated here.

[0164] The following is an example of the transfer process of the to-be-processed data between servers in a specific application:

[0165] As Figure 6b shown, it is a schematic diagram of an application scenario for storing data in the cloud. In an application scenario such as recording data such as the driving mileage and driving time of a vehicle, only the first analysis server 610 needs to send the to-be-processed data for preliminary analysis to the storage server 620 for storage. Combining Figure 3a with the illustrated embodiments, the conventional data partition in the storage server 620 is used at this time.

[0166] As Figure 6c shown, it is a schematic diagram of an application scenario for organizing and storing data in the cloud. If it is necessary to clean and adjust the format of the data in the storage server, only the storage server 620 (combining Figure 3a with the illustrated embodiments, the low-quality data partition is used at this time) needs to send the data to the data cleaning server 631 and receive the processed to-be-processed data returned (and store it in the parsed data partition).

[0167] As Figure 6d shown, it is a schematic diagram of an application scenario for using data in the cloud for application development. In an application scenario such as using engine pressure change data for the development of functions related to engine leakage detection, the storage server 620 (combining Figure 3a with the illustrated embodiments, the conventional data partition, the processed data partition, and the parsed data partition are used at this time) needs to send the data to the data annotation server 632 and the driving data processing server 633 for processing to determine the relative relationship between the corresponding engine pressure data and the pressure data in the environment, and then send the relative relationship data to the data scheduling server 636 to develop and test the corresponding application functions based on the relative relationship data.

[0168] The autonomous vehicle data system provided by the embodiments of the present application combines a first analysis server, a storage server, and at least one processing server to achieve automatic cloud transfer and automatic processing of the data generated by autonomous vehicles in the cloud, fully ensuring the data transfer efficiency, while solving the problem of data staying in a single server for a long time, effectively ensuring that the data is processed quickly and comprehensively, and thus ensuring the R & D efficiency of the autonomous vehicle system.

[0169] An embodiment of the embodiments of the present application provides a computer-readable storage medium, on which a computer program is stored, and the computer program is executed by a processor to implement the Figures 2 to 3a autonomous vehicle data processing method of any corresponding embodiment.

[0170] Among them, the computer-readable storage medium may be ROM, random access memory (RAM), CD-ROM, magnetic tape, floppy disk, and optical data storage device, etc.

[0171] An embodiment of the embodiments of the present application provides a computer program product, which includes computer execution instructions, and when the computer execution instructions are executed by a processor, they are used to implement the Figures 2 to 3a autonomous vehicle data processing method of any corresponding embodiment.

[0172] In several embodiments provided by the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of modules is only a logical function division. In actual implementation, there may be other division methods. For example, multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed mutual coupling or direct coupling or communication connection may be through some interfaces, and the indirect coupling or communication connection of devices or modules may be in an electrical, mechanical or other form.

[0173] Those skilled in the art will easily think of other implementation schemes of the present application after considering the specification and practicing the disclosure here. The present application aims to cover any variations, uses, or adaptive changes of the present application. These variations, uses, or adaptive changes follow the general principles of the present application and include the common general knowledge or conventional technical means in the technical field not disclosed in the present application. The specification and embodiments are only regarded as exemplary, and the true scope of the present application is pointed out by the claims.

[0174] It should be understood that the present application is not limited to the exact structure already described and shown in the drawings, and various modifications and changes can be made without departing from its scope.

Claims

1. A method for processing autonomous vehicle data, characterized in that, applied to a storage server, the method for processing autonomous vehicle data includes: obtaining data to be processed through a first analysis server and caching the data to be processed; based on the data processing requirement information corresponding to the data to be processed, determining at least one matching processing server, where the data processing requirement information includes one or several combinations of the following: data status requirement information, set format requirement information, label requirement information, vehicle status information, format and signal quality information; sending the data to be processed to the at least one processing server for processing respectively; the storage server includes storage areas corresponding to at least one type of data, and obtaining the data to be processed through the first analysis server and caching the data to be processed includes: in response to the data to be processed and its type sent by the first analysis server, the type of the data to be processed is identified by the first analysis server; based on the type, caching the data to be processed into the corresponding storage area; the data to be processed includes at least one processing step and the corresponding processing flow; the determining at least one matching processing server based on the data processing requirement information corresponding to the data to be processed includes: based on the data processing requirement information corresponding to the data to be processed, determining the at least one processing step and the corresponding processing flow corresponding to the data to be processed; respectively determining the processing server matching the step based on the data processing requirement information corresponding to each processing step of the data to be processed; the sending the data to be processed to the at least one processing server for processing respectively includes: according to the processing flow corresponding to the at least one processing step, sequentially sending the data to be processed to the processing server matching the data processing requirement information corresponding to the processing step.

2. The method for processing autonomous vehicle data according to claim 1, characterized in that, the sequentially sending the data to be processed to the processing server matching the data processing requirement information corresponding to the processing step includes: for each step, if the processing server matching the data processing requirement information corresponding to the step is a data cleaning server, then sending the data to be processed to the data cleaning server based on the processing flow, so that the data cleaning server can identify and delete abnormal data in the step according to the data status requirement information corresponding to the step, and receive the processed step; or, if the processing server matching the data processing requirement information corresponding to the data of the step is a data cleaning server, then sending the data to be processed to the data cleaning server based on the processing flow, so that the data cleaning server can perform format conversion processing on the data to be processed according to the set format requirement information corresponding to the step, and receive the processed data to be processed; Alternatively, if the processing server that matches the data processing requirement information of the data corresponding to the step is a data annotation server, then send the data to be processed to the data annotation server based on the processing flow, so that the data annotation server can add label processing to the data to be processed according to the label requirement information corresponding to the step, and receive the processed data to be processed; Alternatively, if the processing server that matches the data processing requirement information of the data corresponding to the step is a driving data processing server, then send the data to be processed to the driving data processing server based on the processing flow, so that the driving data processing server can perform annotation processing on the data to be processed according to the format and signal quality information corresponding to the step, and receive the processed data to be processed; Alternatively, if the processing server that matches the data processing requirement information of the data corresponding to the step is a second analysis server, then send the data to be processed to the second analysis server based on the processing flow, so that the second analysis server can store the data to be processed in the form of a data sequence according to the vehicle state information corresponding to the step and the label of the data to be processed.

3. The method for processing autonomous vehicle data according to claim 1 or 2, wherein, the processing server further includes a data training server for model training, after sending the data to be processed to the at least one processing server for processing respectively, it further includes: if the data processing requirement information further includes data training information, then send the processed data to be processed to the data training server corresponding to the data training information, so that the data training server can perform model training according to the data training information and the processed data to be processed.

4. The method for processing autonomous vehicle data according to claim 1 or 2, wherein, the processing server further includes a data scheduling server for function development scheduling, after sending the data to be processed to the at least one processing server for processing respectively, it further includes: if the data processing requirement information further includes application requirement information, then send the processed data to be processed to the data scheduling server corresponding to the application requirement information, so that the data scheduling server can execute a data call task according to the application requirement information and the processed data to be processed.

5. An autonomous vehicle data processing device, wherein, applied to a storage server, including: a receiving module, configured to obtain the data to be processed through a first analysis server and cache the data to be processed; a matching module, configured to determine at least one matching processing server based on the data processing requirement information corresponding to the data to be processed, where the data processing requirement information includes one or several combinations of the following: data status requirement information, set format requirement information, label requirement information, vehicle state information, format and signal quality information; a transmission module, configured to send the data to be processed to the at least one processing server for processing; The storage server includes a storage area corresponding to at least one type of data. The receiving module is specifically configured to, if the storage server includes a storage area corresponding to at least one type of data, in response to the received to-be-processed data and its type sent by the first analysis server, the type of the to-be-processed data is obtained by the first analysis server; based on the type, cache the to-be-processed data into the corresponding storage area; The to-be-processed data includes at least one processing step and a corresponding processing flow; the matching module is specifically configured to, if the to-be-processed data includes at least one processing step and a corresponding processing flow, based on the data processing requirement information corresponding to the to-be-processed data, determine the at least one processing step and the corresponding processing flow corresponding to the to-be-processed data; respectively based on the data processing requirement information corresponding to each processing step of the to-be-processed data, determine the processing server that matches the step; the transmission module is specifically configured to, according to the processing flow corresponding to at least one processing step, sequentially send the to-be-processed data to the processing server that matches the data processing requirement information corresponding to the processing step.

6. A control device, characterized in that, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein, the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to cause the control device to execute the autonomous vehicle data processing method according to any one of claims 1 to 4.

7. An autonomous vehicle data processing system, the autonomous vehicle data processing system is used to implement the autonomous vehicle data processing method according to any one of claims 1-4, characterized in that, comprising: a first analysis server, configured to receive to-be-processed data from the vehicle control unit and perform classification processing; a storage server, configured to obtain the classified to-be-processed data from the first analysis server and send it to at least one processing server for processing based on the data processing requirement information corresponding to the to-be-processed data; at least one processing server, configured to process the to-be-processed data based on the data processing requirement information.

Images