Autonomous driving data collection method, device, electronic device and readable medium

Abstract

The present application relates to an autonomous driving data collection method, device, electronic device and readable medium. The method includes: during the autonomous driving process of a vehicle, when an abnormality in the autonomous driving function of the vehicle is detected, obtaining a description of the current problem of the vehicle; determining target recording parameters that match the current problem description from a pre-built key-value pair mapping table; the target recording parameters include a target sensor and a target recording duration; determining a first time period based on the target recording duration and the triggering time of the abnormality in the autonomous driving function; and obtaining target sensor data collected by the target sensor within the first time period. The solution provided by the present application can solve the problems of high storage and transmission pressure and low efficiency in extracting key information caused by recording full data for all time periods or a certain time period in related technologies.

Description

Technical Field

[0001] The present application relates to the technical field of autonomous driving data processing, and in particular to autonomous driving data collection methods, devices, electronic devices and readable media. Background Art

[0002] During operation and testing, autonomous vehicles need to collect large amounts of data for model optimization and version iteration to improve autonomous driving capabilities; or to trace back data when a failure occurs to analyze the cause of the failure and optimize response measures.

[0003] The related technology records the full amount of data for the entire time period or records the full amount of data for a fixed time period when manual takeover is performed, and then stores it on a designated hard disk, and then uploads it in batches to the cloud for analysis. However, this method has many problems: first, a large amount of invalid data will be retained, resulting in a waste of storage resources; second, the large amount of data will not only increase the burden of recording, but the excessive storage pressure on the hard disk may also cause data frame loss, thereby missing key information; third, the return and data processing are time-consuming, affecting the timeliness of analysis. Summary of the Invention

[0004] In order to solve or partially solve the problems existing in the related art, the present application provides an autonomous driving data collection method, device, electronic device and readable medium, which can solve the problems in the related art of high storage and transmission pressure and low efficiency of key information extraction caused by full data recording for all time periods or a certain time period.

[0005] In a first aspect, the present application provides a method for collecting autonomous driving data, the method comprising:

[0006] During the automatic driving process of the vehicle, when an abnormality of the automatic driving function of the vehicle is detected, obtaining a description of the current problem of the vehicle;

[0007] Determining target recording parameters that match the current problem description from a pre-built key-value pair mapping table; the target recording parameters include a target sensor and a target recording duration;

[0008] determining a first time period according to the target recording duration and the triggering time of the abnormality of the autonomous driving function;

[0009] Target sensor data collected by the target sensor during the first time period is acquired.

[0010] In one embodiment, the method further comprises:

[0011] storing the target sensor data in a hard disk;

[0012] The target sensor data stored in the hard disk is used as the autonomous driving data to be analyzed, so as to upload the autonomous driving data to the cloud for analysis.

[0013] In one embodiment, the key-value pair mapping table includes a plurality of preset question keywords and a designated sensor and a preset recording duration corresponding to each of the preset question keywords; and determining the target recording parameters that match the current question description from the pre-constructed key-value pair mapping table includes:

[0014] Extract target problem keywords from the current problem description;

[0015] From the key-value pair mapping table, a designated sensor that matches the target question keyword is determined as a target sensor, and a preset recording duration that matches the target question keyword is determined as a target recording duration.

[0016] In one embodiment, the target recording parameters further include a target log file and a target interception duration, and the key-value pair mapping table further includes a designated log file and a preset interception duration corresponding to each of the preset question keywords;

[0017] After extracting target question keywords from the current question description, the method further includes:

[0018] From the key-value pair mapping table, a designated log file that matches the target question keyword is determined as a target log file, and a preset interception duration that matches the target question keyword is determined as a target interception duration.

[0019] In one embodiment, the method further comprises:

[0020] determining a second time period according to the target interception time and the triggering time of the abnormality of the autonomous driving function;

[0021] Obtain target log data collected by the target log file during the second time period.

[0022] In one embodiment, the using the target sensor data stored on the hard disk as the autonomous driving data to be analyzed further includes:

[0023] The target log data and the target sensor data stored in the hard disk are used as the autonomous driving data to be analyzed.

[0024] In one embodiment, the target recording duration includes a first forward recording duration and a total recording duration; and determining the first time period based on the target recording duration and the triggering time of the abnormal autonomous driving function includes:

[0025] Taking the triggering time of the abnormal autonomous driving function as a benchmark, extending the first forward time forward to obtain a recording starting point;

[0026] Taking the recording start point as a reference, the total recording time is extended backward to obtain the recording end point;

[0027] The time period between the recording start point and the recording end point is determined as a first time period.

[0028] In one embodiment, the target interception time includes a second forward time and a total interception time; and determining the second time period based on the target interception time and the triggering time of the abnormal autonomous driving function includes:

[0029] Taking the triggering time of the abnormal autonomous driving function as a benchmark, extending the second forward traceback time forward to obtain an interception starting point;

[0030] Taking the interception starting point as a reference, the total interception time is extended backward to obtain the interception end point;

[0031] A time period between the interception start point and the interception end point is determined as a second time period.

[0032] In one embodiment, the acquiring target sensor data collected by the target sensor during the first time period includes:

[0033] Target sensor data collected by the target sensor during the first time period is extracted from the annular buffer area; wherein the annular buffer area is used to continuously and rollingly cache sensor data collected by multiple sensors of the vehicle during a preset buffer time period, and the preset buffer time period includes the first time period.

[0034] In one embodiment, storing the target sensor data in a hard disk includes:

[0035] The target sensor data is encrypted, and the encrypted target sensor data is stored in a hard disk.

[0036] A second aspect of the present application provides an autonomous driving data collection device, the device comprising:

[0037] A current problem description acquisition module is used to acquire a current problem description of the vehicle when an abnormality in the automatic driving function of the vehicle is detected during the automatic driving process of the vehicle;

[0038] A target recording parameter determination module is used to determine target recording parameters that match the current problem description from a pre-built key-value pair mapping table; the target recording parameters include a target sensor and a target recording duration;

[0039] a first time period determination module, configured to determine a first time period according to the target recording duration and a triggering time of an abnormality in the autonomous driving function;

[0040] The target sensor data acquisition module is used to acquire the target sensor data collected by the target sensor within the first time period.

[0041] A third aspect of the present application provides an electronic device, including:

[0042] processor; and

[0043] The memory stores executable codes thereon, and when the executable codes are executed by the processor, the processor is caused to execute the method described above.

[0044] A fourth aspect of the present application provides a computer-readable storage medium having executable code stored thereon. When the executable code is executed by a processor of an electronic device, the processor is caused to execute the method described above.

[0045] The technical solution provided by this application may have the following beneficial effects:

[0046] The solution provided by this application is to obtain the current problem description of the vehicle when an abnormality in the vehicle's autonomous driving function is detected during the vehicle's autonomous driving process; determine the target recording parameters that match the current problem description from a pre-constructed key-value mapping table; the target recording parameters include the target sensor and the target recording duration; determine a first time period based on the target recording duration and the trigger time of the autonomous driving function abnormality; and obtain the target sensor data collected by the target sensor during the first time period. This application pre-constructs a key-value mapping table so that the target sensor and the target recording duration that match the current problem description are determined based on the key-value mapping table. In this way, the recording object is only the target sensor, not all sensors, thereby reducing the data volume. The first time period is determined based on the target recording duration and the trigger time of the autonomous driving function abnormality. In this way, only the data of the first time period needs to be recorded, and there is no need to record data for the entire time period or a fixed time period, thereby further reducing the data volume and greatly alleviating the hard disk storage pressure. In addition, the reduction in data volume can shorten the cloud upload time, thereby enabling real-time backhaul and accelerating analysis time.

[0047] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] The above and other objects, features and advantages of the present application will become more apparent by describing in more detail the exemplary embodiments of the present application in conjunction with the accompanying drawings, wherein the same reference numerals generally represent the same components in the exemplary embodiments of the present application.

[0049] Figure 1 1 is a flow chart of an autonomous driving data collection method according to an embodiment of the present application;

[0050] Figure 2 is another flowchart of the autonomous driving data collection method shown in an embodiment of the present application;

[0051] Figure 3 1 is a schematic structural diagram of an autonomous driving data acquisition device according to an embodiment of the present application;

[0052] Figure 4 It is a structural diagram of an electronic device shown in an embodiment of the present application. DETAILED DESCRIPTION

[0053] The following describes embodiments of the present application in more detail with reference to the accompanying drawings. Although the accompanying drawings illustrate embodiments of the present application, it should be understood that the present application can be implemented in various forms and should not be limited by the embodiments described herein. Rather, these embodiments are provided to make the present application more thorough and complete, and to fully convey the scope of the present application to those skilled in the art.

[0054] The terms used in this application are for the purpose of describing specific embodiments only and are not intended to limit this application. As used in this application and the appended claims, the singular forms "a," "an," "the," and "the" are intended to include the plural forms, unless the context clearly indicates otherwise. It should also be understood that the term "and / or" as used herein refers to and encompasses any and all possible combinations of one or more of the associated listed items.

[0055] It should be understood that although the terms "first", "second", "third", etc. may be used in this application to describe various information, this information should not be limited to these terms. These terms are only used to distinguish information of the same type from each other. For example, without departing from the scope of this application, the first information may also be referred to as the second information, and similarly, the second information may also be referred to as the first information. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of this application, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0056] The relevant technologies include two recording methods. One is to record the full amount of data for the entire period of time. For example, the vehicle records the sensor data collected by all sensors in real time during the autonomous driving process. The other is to record the full amount of data for a certain period of time. For example, the vehicle caches the sensor data collected by all sensors in real time during the autonomous driving process. When manual control is taken over, the system will trigger the recording of the cached sensor data for a fixed period of time. However, regardless of the recording method, the relevant technology still adds a large number of hard disks to store the full amount of data, and then uses the hard disk as a storage medium for relevant personnel to analyze. This method has many problems: First, a large amount of invalid data will be retained, resulting in a waste of storage resources; second, the large amount of data will not only increase the burden of recording, but the excessive storage pressure of the hard disk may also cause data frame loss, thereby missing key information; third, the backhaul and data processing are time-consuming, affecting the timeliness of analysis.

[0057] Since all sensors work together to generate a large amount of data and increase in volume, how to accurately collect lightweight autonomous driving data to alleviate storage and transmission pressures and quickly transfer it to improve analysis efficiency is a technical challenge that needs to be solved urgently.

[0058] In response to the above problems, an embodiment of the present application provides an autonomous driving data collection method, which pre-constructs a key-value pair mapping table to determine the target sensor and target recording duration that match the current problem description based on the key-value pair mapping table. In this way, the recording object is only the target sensor, rather than all sensors, thereby reducing the data volume, and determines the first time period based on the target recording duration and the trigger time of the autonomous driving function abnormality. In this way, only the data of the first time period needs to be recorded, and there is no need to record data for the entire time period or a fixed time period, thereby further reducing the data volume, thereby greatly alleviating the hard disk storage pressure. In addition, the reduction in data volume can shorten the cloud upload time, so real-time feedback can be achieved, accelerating the analysis timeliness.

[0059] The technical solutions of the embodiments of the present application are described in detail below with reference to the accompanying drawings.

[0060] Figure 1 It is a flow chart of the autonomous driving data collection method shown in an embodiment of the present application.

[0061] See also Figure 1 , the autonomous driving data collection method of the present application includes:

[0062] S110, during the vehicle's automatic driving process, when it is detected that the vehicle's automatic driving function is abnormal, obtain a description of the vehicle's current problem.

[0063] In the embodiment of the present application, it can be applied to the collection equipment, and the collection equipment can specifically be an autonomous driving system or a VCU (Vehicle Control Unit) or an ECU (Electronic Control Unit), and the embodiment of the present application does not limit this.

[0064] In one implementation, a vehicle drives autonomously on a set route (such as a test route or an operating route). During this process, the driver observes the vehicle's driving conditions and surrounding environment in real time. When the driver senses that danger is about to occur, the driver can trigger a takeover operation for the vehicle by stepping on the brake pedal or manipulating the steering wheel. Based on the takeover operation, the acquisition device determines that the autonomous driving function is abnormal, and then switches the vehicle from autonomous driving mode to manual driving mode in response to the takeover operation, and can obtain a description of the current problem of the vehicle.

[0065] In another implementation, the vehicle autonomously drives along a set route. During this process, the chassis system detects in real time whether the vehicle is experiencing any abnormal conditions. If so, the chassis system generates feedback data and sends it to a collection device. Based on this feedback data, the collection device determines that the autonomous driving function is abnormal and then obtains a description of the vehicle's current problem. It should be noted that this scenario could, for example, involve the autonomous driving function malfunctioning without manual intervention, or serious traffic accidents such as collisions and spontaneous combustion involving autonomous vehicles from certain manufacturers reported in the news. In such scenarios, pre-incident autonomous driving data collection and feedback analysis are also necessary.

[0066] Among them, the current problem description is used to characterize the current problem that the vehicle encounters during the autonomous driving process, such as the risk of collision between the vehicle and the vehicle in front, or the situation where the vehicle draws a dragon.

[0067] S120 , determining target recording parameters that match the current problem description from a pre-built key-value pair mapping table; the target recording parameters include a target sensor and a target recording duration.

[0068] In an embodiment of the present application, a key-value pair mapping table can be pre-built. In a specific implementation, various historical problem descriptions are first collected, so that based on the keywords contained in each historical problem description, multiple preset problem keywords with different problem types can be configured. For example, the preset problem keywords can include at least one of collision risk, "dragon drawing," "failed lane change," "no obstacle avoidance," and "no deceleration." The recording object corresponding to each preset problem keyword is then determined. The recording object can be a designated sensor. As an example, assuming the preset problem keyword is "dragon drawing," since "dragon drawing" is related to planning and control sensors, the recording object corresponding to "dragon drawing" can be determined to be a planning and control sensor, so that the planning and control sensor can be used as the designated sensor for "dragon drawing." As another example, assuming the preset problem keyword is "no obstacle avoidance," since "no obstacle avoidance" is related to obstacle avoidance sensors, the recording object corresponding to "no obstacle avoidance" can be determined to be an obstacle avoidance sensor, so that the obstacle avoidance sensor can be used as the designated sensor for "no obstacle avoidance." The preset recording duration required for each preset problem keyword is then configured. After obtaining each preset question keyword and its corresponding designated sensor and preset recording time, a one-to-many mapping relationship between each preset question keyword and its corresponding designated sensor and preset recording time can be constructed to obtain a key-value pair mapping table.

[0069] In actual applications, after obtaining the current problem description of the vehicle, the acquisition device can determine the preset problem keyword that matches the current problem description from the above-mentioned key-value pair mapping table, and then determine the designated sensor corresponding to the preset problem keyword as the target sensor, and determine the preset recording time corresponding to the preset problem keyword as the target recording time. The target sensor and target recording time are the target recording parameters that match the current problem description.

[0070] S130: Determine a first time period based on the target recording time and the triggering time of the abnormal autonomous driving function.

[0071] The acquisition device can determine the trigger time of the abnormality of the autonomous driving function based on the time point when the takeover operation is detected, or can determine the trigger time of the abnormality of the autonomous driving function based on the time point when the feedback data is received, so as to determine the first time period according to the trigger time of the abnormality of the autonomous driving function and the target recording time, and the first time period is the recording time period of the target sensor.

[0072] S140: Acquire target sensor data collected by the target sensor within a first time period.

[0073] The acquisition device can acquire the target sensor data collected by the target sensor during the first time period so as to subsequently store the target sensor data on a hard disk, thereby completing the recording of the target sensor data. The hard disk is a device configured within the vehicle for long-term data and information storage. The hard disk can include at least one of an HDD (Hard Disk Drive) and an SSD (Solid State Drive).

[0074] The hard disk is a long-term storage device of the vehicle. Even when the vehicle is powered off, the data in the hard disk can be permanently saved. Therefore, after the embodiment of the present application stores the target sensor data in the hard disk, the target sensor data can be extracted from the hard disk at any time during the testing or operation phase. For example, the target sensor data can be extracted from the hard disk of the vehicle immediately after the vehicle completes recording, or the target sensor data can be extracted from the hard disk of the vehicle after the vehicle is powered off, or the corresponding target sensor data can be extracted from the hard disk of each vehicle after some or all vehicles have completed recording. The embodiment of the present application does not limit this.

[0075] The target sensor data extracted from the hard disk can be used as autonomous driving data to be analyzed, and then uploaded to the cloud and stored to trigger reminders to relevant personnel so that they can analyze the autonomous driving data in a timely manner.

[0076] As can be seen, the embodiment of the present application configures corresponding designated sensors and preset recording durations for different preset problem keywords. Therefore, the embodiment of the present application mainly reduces the data volume from two aspects: the number of sensors and the recording duration. For example, the designated sensors corresponding to "drawing a dragon" are sensors A, B, and C, and the preset recording duration corresponding to "drawing a dragon" is x seconds. The designated sensors corresponding to "not bypassing obstacles" are sensors D, E, F, and G, and the preset recording duration corresponding to "not bypassing obstacles" is y seconds. If the current problem description is "this vehicle has a dragon-drawing situation", on the one hand, the recording objects at this time only include sensors A, B, and C, and do not include sensors D, E, F, G and other sensors, thereby reducing the data volume by reducing the number of useless sensors; on the other hand, the recording duration at this time is x seconds, not y seconds, nor a fixed period (e.g., regardless of the type of problem the vehicle has, it records for z seconds), nor the entire period (e.g., real-time recording during autonomous driving), thereby further reducing the data volume by the currently required recording duration. Therefore, compared with the data volume in the related art, the data volume of the embodiment of the present application can be greatly reduced, for example, by 50%. Due to the reduction in the number of orders of magnitude, the pressure on hard disk storage is greatly alleviated, and the cloud upload time is shortened, so real-time feedback can be achieved and analysis time is accelerated.

[0077] As can be seen from this example, the solution provided by this application, during the process of autonomous driving of a vehicle, when an abnormality in the autonomous driving function of the vehicle is detected, obtains the current problem description of the vehicle; determines the target recording parameters that match the current problem description from a pre-built key-value mapping table; the target recording parameters include the target sensor and the target recording duration; determines a first time period based on the target recording duration and the trigger time of the autonomous driving function abnormality; and obtains the target sensor data collected by the target sensor during the first time period. This application pre-builds a key-value mapping table so that the target sensor and the target recording duration that match the current problem description are determined based on the key-value mapping table. In this way, the recording object is only the target sensor, not all sensors, thereby reducing the amount of data. The first time period is determined based on the target recording duration and the trigger time of the autonomous driving function abnormality. In this way, only the data of the first time period needs to be recorded, without recording data for the entire time period or a fixed time period, thereby further reducing the amount of data, thereby greatly alleviating the pressure on hard disk storage. In addition, the reduction in data volume can shorten the upload time to the cloud, thereby enabling real-time backhaul and accelerating analysis time.

[0078] Figure 2 This is another flow chart of the autonomous driving data collection method shown in an embodiment of the present application.

[0079] See also Figure 2 , the autonomous driving data collection method of the present application includes:

[0080] S210, during the vehicle's automatic driving process, when it is detected that the vehicle's automatic driving function is abnormal, obtain a description of the vehicle's current problem.

[0081] In one implementation, a vehicle drives autonomously on a set route (such as a test route or an operating route). During this process, the driver observes the vehicle's driving conditions and surrounding environment in real time. When the driver senses that danger is about to occur, the driver can trigger a takeover operation for the vehicle by stepping on the brake pedal or manipulating the steering wheel. Based on the takeover operation, the acquisition device determines that the autonomous driving function is abnormal, and then switches the vehicle from autonomous driving mode to manual driving mode in response to the takeover operation, and can obtain a description of the current problem of the vehicle.

[0082] In another implementation, the vehicle performs autonomous driving on a set route. During this process, the chassis system detects in real time whether the vehicle is in an abnormal state. If the vehicle is in an abnormal state, the chassis system generates feedback data and sends the feedback data to the collection device. Based on the feedback data, the collection device determines that the autonomous driving function is abnormal and then obtains a description of the current problem of the vehicle.

[0083] The current problem description is used to characterize the current problem that the vehicle is experiencing during autonomous driving, such as the risk of collision between the vehicle and the vehicle in front, or the situation where the vehicle is drawing a dragon.

[0084] In one embodiment, when an abnormality in the autonomous driving function of a vehicle is detected, obtaining a description of the current problem of the vehicle may include:

[0085] When a takeover operation for the vehicle is detected, multiple preset question keywords are displayed on the vehicle's central control screen; the current problem description of the vehicle transmitted by the central control screen is received; wherein the current problem description is generated by the central control screen based on the target problem keyword selected by the driver from multiple preset problem keywords.

[0086] The embodiment of the present application can pre-collect various historical problem descriptions so as to configure multiple preset problem keywords with different problem types according to the keywords contained in each historical problem description, wherein the preset problem keywords may include at least one of collision risk, drawing a dragon, failure to change lanes at the right time, not bypassing obstacles, and not slowing down.

[0087] In actual applications, when the driver's takeover operation for the vehicle is detected, the acquisition device can display the above-mentioned multiple preset problem keywords on the vehicle's central control screen, so that the driver can select any preset problem keyword from these preset problem keywords. In response to the selection operation, the central control screen determines the selected preset problem keyword as the target problem keyword, and then generates a description of the vehicle's current problem based on the target problem keyword, and then transmits the current problem description to the acquisition device. For example, assuming that the driver selects "drawing a dragon" from preset problem keywords such as "collision risk", "drawing a dragon", "failure to change lanes at an appropriate time", "not bypassing obstacles", and "not slowing down", then "drawing a dragon" can be used as the target problem keyword. The central control screen generates a description of the vehicle's current problem based on "drawing a dragon". The current problem description can be, for example, "a dragon drawing situation occurs in this vehicle", and the central control screen transmits the current problem description to the acquisition device.

[0088] In addition, the embodiments of the present application can also obtain the current problem description of the vehicle through other methods, for example, it can receive and recognize the target problem keywords input by the driver's voice, and then generate the current problem description of the vehicle based on the target problem keywords.

[0089] In another embodiment, when an abnormality in the autonomous driving function of a vehicle is detected, obtaining a description of the current problem of the vehicle may include:

[0090] When feedback data from the chassis system is received, a current problem description of the vehicle is generated based on the feedback data.

[0091] In actual applications, when the chassis system detects that the vehicle is in an abnormal state, it sends the current driving data of the vehicle as feedback data to the acquisition device. The acquisition device determines the current problem of the vehicle based on the current driving data, and then generates a current problem description of the vehicle based on the determined problem. For example, the current driving data includes the relative position of the front vehicle relative to the vehicle, and the relative speed of the front vehicle relative to the vehicle. The acquisition device determines that the distance between the vehicle and the front vehicle is too small based on the relative position and relative speed, and there is a risk of collision. Therefore, the acquisition device generates the current problem description of the vehicle based on the determined problem. The current problem description can be, for example, "There is a risk of collision between the vehicle and the vehicle in front."

[0092] S220 , determining target recording parameters that match the current problem description from a pre-built key-value pair mapping table; the target recording parameters include a target sensor, a target recording duration, a target log file, and a target capture duration.

[0093] This embodiment of the present application can pre-build a key-value mapping table. In a specific implementation, this embodiment separately determines the recording object corresponding to each of the preset question keywords. The recording object can be a designated sensor and its log file (log file). As an example, assuming the preset question keyword is "drawing a dragon," since "drawing a dragon" is related to planning and control sensors, the recording object corresponding to "drawing a dragon" can be determined to be a planning and control sensor and its log file. Therefore, the planning and control sensor is designated as the designated sensor for "drawing a dragon," and the log file of the planning and control sensor is designated as the designated log file for "drawing a dragon." As another example, assuming the preset question keyword is "not bypassing obstacles," since "not bypassing obstacles" is related to obstacle avoidance sensors, the recording object corresponding to "not bypassing obstacles" can be determined to be an obstacle avoidance sensor and its log file. Therefore, the obstacle avoidance sensor is designated as the designated sensor for "not bypassing obstacles," and the log file of the obstacle avoidance sensor is designated as the designated log file for "not bypassing obstacles." The preset recording duration and preset capture duration are then configured for each preset question keyword. After obtaining each preset question keyword and its corresponding specified sensor, preset recording time, specified log file and preset interception time, a one-to-many mapping relationship between each preset question keyword and its corresponding specified sensor, preset recording time, specified log file and preset interception time can be constructed respectively, thereby constructing a key-value pair mapping table.

[0094] In one example, the format of the key-value mapping table is as follows:

[0095] {

[0096] 'Drawing Dragon': {

[0097] 'msg': {

[0098] 'Sensor 1 Name': ['Previous Recording Duration', 'Total Recording Duration'],

[0099] 'Sensor 2 Name': ['Previous Recording Duration', 'Total Recording Duration'],

[0100] 'Sensor 3 Name': ['Previous Recording Duration', 'Total Recording Duration']

[0101] },

[0102] 'log': {

[0103] 'Sensor 1 log name': ['Previous duration', 'Total intercept duration'],

[0104] 'Sensor 2 log name': ['Previous duration', 'Total intercept duration'],

[0105] 'Sensor 3 log name': ['Previous duration', 'Total intercept duration']

[0106] }

[0107] },

[0108] 'No Obstacles': {

[0109] 'msg': {

[0110] 'Sensor 1 Name': ['Previous Recording Duration', 'Total Recording Duration'],

[0111] 'Sensor 2 Name': ['Previous Recording Duration', 'Total Recording Duration'],

[0112] 'Sensor 3 Name': ['Previous Recording Duration', 'Total Recording Duration']

[0113] },

[0114] 'log': {

[0115] 'Sensor 1 log name': ['Previous duration', 'Total intercept duration'],

[0116] 'Sensor 2 log name': ['Previous duration', 'Total intercept duration'],

[0117] 'Sensor 3 log name': ['Previous duration', 'Total intercept duration']

[0118] }

[0119] }

[0120] }

[0121] In actual applications, after obtaining the current problem description of the vehicle, the acquisition device can determine the preset problem keyword that matches the current problem description from the above-mentioned key-value pair mapping table, and then determine the designated sensor corresponding to the preset problem keyword as the target sensor, and the preset recording time corresponding to the preset problem keyword as the target recording time, and the designated log file corresponding to the preset problem keyword as the target log file, and the preset interception time corresponding to the preset problem keyword as the target interception time. The target sensor, target recording time, target log file and target interception time are the target recording parameters that match the current problem description.

[0122] It should be noted that in the above key-value mapping table, Sensor 1, Sensor 2, and Sensor 3 are sensors arranged in order. For example, Sensor 1, Sensor 2, and Sensor 3 are the first, second, and third sensors, respectively. Different preset question keywords usually correspond to different designated sensors. For example, the designated sensor corresponding to "drawing a dragon" is usually a planning and control sensor, while the designated sensor corresponding to "not bypassing obstacles" is usually an obstacle avoidance sensor. Therefore, in the above key-value mapping table, the names of Sensor 1, Sensor 2, and Sensor 3 corresponding to "drawing a dragon" all point to planning and control sensors, while the names of Sensor 1, Sensor 2, and Sensor 3 corresponding to "not bypassing obstacles" all point to obstacle avoidance sensors.

[0123] In one embodiment, a key-value pair mapping table includes a plurality of preset question keywords and a designated sensor, a preset recording duration, a designated log file, and a preset capture duration corresponding to each preset question keyword; determining target recording parameters that match the current question description from the pre-built key-value pair mapping table may include:

[0124] Extract target question keywords from the current question description; determine the specified sensor that matches the target question keyword as the target sensor from the key-value pair mapping table, and determine the preset recording time that matches the target question keyword as the target recording time, and determine the specified log file that matches the target question keyword as the target log file, and determine the preset interception time that matches the target question keyword as the target interception time.

[0125] As shown in the format of the key-value pair mapping table above, the key-value pair mapping table may include multiple preset question keywords, such as "drawing a dragon" and "not bypassing obstacles". The embodiment of the present application may expand the preset question keywords according to actual conditions. The key-value pair mapping table may also include the designated sensors and preset recording duration corresponding to each preset question keyword. For example, the designated sensors corresponding to "drawing a dragon" are sensors A, B, and C. Sensors A, B, and C are specifically planning and control sensors. For another example, the designated sensors corresponding to "not bypassing obstacles" are sensors D, E, F, and G. Sensors D, E, F, and G are specifically obstacle avoidance sensors. The key-value pair mapping table may also include the designated log files and preset interception duration corresponding to each preset question keyword. For example, the designated log files corresponding to "drawing a dragon" are the log files logA of sensor A, logB of sensor B, and logC of sensor C. For another example, the designated log files corresponding to "not bypassing obstacles" are the log files logD of sensor D, logE of sensor E, logF of sensor F, and logG of sensor G.

[0126] In actual applications, the acquisition device can extract the target problem keywords from the current problem description, and then determine the specified sensor that matches the target problem keyword as the target sensor from the above-mentioned key-value pair mapping table, and determine the preset recording time that matches the target problem keyword as the target recording time, and determine the specified log file that matches the target problem keyword as the target log file, and determine the preset interception time that matches the target problem keyword as the target interception time. The target sensor, target recording time, target log file and target interception time are the target recording parameters that match the current problem description.

[0127] As an example, assuming that the current problem description is "a dragon-drawing situation occurs on this vehicle", the target problem keyword extracted by the acquisition device from the current problem description is "dragon-drawing", and then from the above key-value pair mapping table, sensors A, B, and C that match "dragon-drawing" are all determined as target sensors, and the preset recording time of sensor A, the preset recording time of sensor B, and the preset recording time of sensor C are all determined as target recording time, and the log file logA of sensor A, the log file logB of sensor B, and the log file logC of sensor C are all determined as target log files, and the preset interception time of log file logA, the preset interception time of log file logB, and the preset interception time of log file logC are all determined as target interception time. These target sensors, target recording time, target log files, and target interception time are the target recording parameters that match "a dragon-drawing situation occurs on this vehicle".

[0128] As another example, assuming that the current problem description is "the vehicle encounters a situation where it cannot bypass obstacles", the target problem keyword extracted by the acquisition device from the current problem description is "no obstacle bypassing", and then from the above key-value pair mapping table, sensors D, E, F, and G that match "no obstacle bypassing" are all determined as target sensors, and the preset recording time of sensor D, the preset recording time of sensor E, the preset recording time of sensor F, and the preset recording time of sensor G are all determined as target recording time, and the log file logD of sensor D, the log file logE of sensor E, the log file logF of sensor F, and the log file logG of sensor G are all determined as target log files, and the preset interception time of log file logD, the preset interception time of log file logE, the preset interception time of log file logF, and the preset interception time of log file logG are all determined as target interception time. These target sensors, target recording time, target log files, and target interception time are the target recording parameters that match "the vehicle encounters a situation where it cannot bypass obstacles".

[0129] S230, determining a first time period according to the target recording time and the triggering time of the abnormal automatic driving function, and determining a second time period according to the target interception time and the triggering time of the abnormal automatic driving function.

[0130] The acquisition device can determine the triggering time of the abnormality of the autonomous driving function based on the time point when the takeover operation is detected, or can determine the triggering time of the abnormality of the autonomous driving function based on the time point when the feedback data is received, so as to determine the first time period according to the triggering time of the abnormality of the autonomous driving function and the target recording duration, and determine the second time period according to the triggering time of the abnormality of the autonomous driving function and the target interception duration, wherein the first time period is the recording time period of the target sensor, and the second time period is the interception time period of the target log file.

[0131] In one embodiment, the target recording duration includes a first forward recording duration and a total recording duration; and determining the first time period based on the target recording duration and the triggering time of the abnormal autonomous driving function may include:

[0132] Taking the triggering time of the abnormal automatic driving function as a benchmark, extend the first forward time forward to obtain the recording starting point; taking the recording starting point as a benchmark, extend the total recording time backward to obtain the recording end point; determine the time period between the recording start point and the recording end point as the first time period.

[0133] As shown in the format of the key-value pair mapping table above, the preset recording duration and the preset interception duration both include the forward-tracking duration. For ease of distinction, the embodiment of the present application may define the forward-tracking duration included in the preset recording duration as the first forward-tracking duration. Therefore, the target recording duration may include the first forward-tracking duration and the total recording duration, wherein the total recording duration may cover the first forward-tracking duration.

[0134] The first forward time period refers to a preset period of time before the triggering time of the autonomous driving function anomaly. Therefore, the acquisition device can extend the first forward time period forward, based on the triggering time of the autonomous driving function anomaly, to obtain the recording start point. The target sensor data may include vehicle status data and surrounding environment data. Therefore, the target sensor data for the period between the recording start point and the triggering time can reflect the vehicle's status and surrounding environment before the problem occurred. Then, the acquisition device extends the total recording duration backward, based on the recording start point, to obtain the recording end point. The target sensor data for the period between the triggering time and the recording end point can reflect the vehicle's status and surrounding environment after the problem occurred (e.g., after the vehicle was taken over by the driver). Therefore, the period between the recording start point and the recording end point is determined as the first time period. The target sensor data for the first time period can fully reflect the vehicle's status and surrounding environment from before the problem occurred to after the problem occurred.

[0135] In one example, assuming that the target question keyword is "drawing a dragon", the corresponding first forward time length can be 15s, and the total recording time can be 30s. If the trigger time of the abnormal autonomous driving function is 9:30:00, the recording start point is 9:29:45 and the recording end point is 9:30:15, so the first time period is 9:29:45~9:30:15.

[0136] In one embodiment, the target interception time includes the second forward time and the total interception time; and determining the second time period based on the target interception time and the triggering time of the abnormal autonomous driving function may include:

[0137] Taking the triggering time of the abnormal automatic driving function as the benchmark, the second forward time is extended forward to obtain the interception starting point; taking the interception starting point as the benchmark, the total interception time is extended backward to obtain the interception end point; the time period between the interception starting point and the interception end point is determined as the second time period.

[0138] As shown in the format of the key-value pair mapping table above, the preset recording time and the preset interception time both include the forward-tracking time. For ease of distinction, the embodiment of the present application can define the forward-tracking time included in the preset interception time as the second forward-tracking time. Therefore, the target interception time can include the second forward-tracking time and the total interception time, wherein the total interception time can cover the second forward-tracking time.

[0139] Similarly, the second forward time period refers to a preset period of time before the triggering time of the autonomous driving function anomaly. Therefore, the collection device can extend the second forward time period forward, based on the triggering time of the autonomous driving function anomaly, to obtain the interception starting point. The target log data is used to analyze the detailed conditions of the target sensor and typically includes the event type and event description. Therefore, the target log data for the period between the interception starting point and the triggering time can reflect the detailed conditions of the target sensor before the problem occurred. Then, the collection device extends the interception total time period backward, based on the interception starting point, to obtain the interception ending point. The target log data for the period between the triggering time and the interception ending point can reflect the detailed conditions of the target sensor after the problem occurred (e.g., after the vehicle was taken over by the driver). Therefore, the period between the interception starting point and the interception ending point is determined as the second time period. The target log data for the second time period can fully reflect the detailed conditions of the target sensor from before to after the problem occurred.

[0140] In one example, assuming that the target question keyword is "drawing a dragon", the corresponding second forward tracking time can be 1 minute, and the total interception time can be 3 minutes. If the trigger time of the abnormal autonomous driving function is 9:30:00, the interception starting point is 9:29:00 and the interception end point is 9:32:00, so the second time period is 9:29:00~9:32:00.

[0141] It should be noted that, for the same preset question keyword, the preset recording time (i.e., the first look-forward time and the total recording time) of different designated sensors may be the same. For example, for "drawing a dragon", the first look-forward time of sensor A, the first look-forward time of sensor B, and the first look-forward time of sensor C may all be 15s, and the total recording time of sensor A, the total recording time of sensor B, and the total recording time of sensor C may all be 30s; or, for the same preset question keyword, the preset recording time (i.e., the first look-forward time and the total recording time) of different designated sensors may be different. For example, for "not bypassing obstacles", sensor D and sensor E are configured at the front of the vehicle, and sensor F and sensor G are configured at the rear of the vehicle. Generally speaking, during the automatic driving process of the vehicle, the obstacle avoidance sensor at the front of the vehicle can play a more effective role. Therefore, in the embodiment of the present application, the preset recording time of sensor D and the preset recording time of sensor E may be set to be greater than the preset recording time of sensor F and the preset recording time of sensor G.

[0142] It should be noted that, for the same preset problem keyword, the preset interception time (i.e., the second forward tracing time and the total interception time) of different designated log files can be the same. For example, for "drawing a dragon", the second forward tracing time of log file logA, the second forward tracing time of log file logB, and the second forward tracing time of log file logC can all be 1 minute, and the total interception time of log file logA, the total interception time of log file logB, and the total interception time of log file logC can all be 3 minutes; or, for the same preset problem keyword, the preset interception time (i.e., the second forward tracing time and the total interception time) of different designated log files can be different. For example, for "not bypassing obstacles", the embodiment of the present application can set the preset interception time of log file logD and the preset interception time of log file logE to be greater than the preset interception time of log file logF and the preset interception time of log file logG.

[0143] It should be noted that for the same designated sensor with the same preset question keyword, the first lookback duration and the second lookback duration can be the same or different. Furthermore, for the same designated sensor with the same preset question keyword, the total recording duration and the total capture duration can be the same or different.

[0144] S240 , acquiring target sensor data collected by the target sensor in the first time period, storing the target sensor data in a hard disk, and acquiring target log data collected by the target log file in the second time period.

[0145] The acquisition device may record target sensor data collected by the target sensor during a first time period. In a specific implementation, the acquisition device may obtain the target sensor data collected by the target sensor during the first time period and then store the target sensor data in a hard disk. The hard disk may be a type of at least one of an HDD and an SSD.

[0146] Since the target log data is recorded in the system disk in real time in the form of a target log file, the target log data does not need to be recorded to the hard disk. The collection device directly extracts the target log data collected in the second time period from the target log file from the system disk.

[0147] In one embodiment, acquiring target sensor data collected by the target sensor within a first time period and storing the target sensor data in a hard disk may include:

[0148] Target sensor data collected by the target sensor within a first time period is extracted from the annular buffer area; wherein the annular buffer area is used to continuously and rollingly cache sensor data collected by multiple sensors of the vehicle within a preset cache time period, and the preset cache time period includes the first time period; the target sensor data is encrypted, and the encrypted target sensor data is stored in the hard disk.

[0149] The embodiment of the present application can pre-construct a circular buffer area, then configure a preset cache duration of the circular buffer area, and then extend the preset cache duration forward based on the current time to obtain the cache starting point, and then use the time period between the cache starting point and the cache end point (i.e., the current time) as the preset cache time period. For example, the preset cache duration can be 3 minutes, then the preset cache time period is the most recent 3 minutes, so the circular buffer area can continuously scroll and cache the sensor data collected by all sensors of the vehicle in the most recent 3 minutes, wherein the preset cache time period includes the first time period, so the embodiment of the present application can extract the target sensor data collected by the target sensor in the first time period from the circular buffer area. For example, assuming that the trigger time of the abnormal autonomous driving function is 9:30:00, the first forward tracking time is 15s, the total recording time is 30s, the current time is 9:32:00, and the preset cache time is 3min, then the preset cache time period is 9:29:00~9:32:00, and the first time period is 9:29:45~9:30:15. Therefore, the annular buffer area currently caches the sensor data collected by all sensors of the vehicle between 9:29:00~9:32:00. Therefore, the acquisition device can extract the target sensor data collected by the target sensor between 9:29:45~9:30:15 from the annular buffer area.

[0150] After extracting the target sensor data, since the target sensor data stored on the hard disk needs to be uploaded to the cloud later, in order to ensure the security of the data upload, the acquisition device can introduce an encryption algorithm when the target sensor data is written to the disk. The encryption algorithm can be the AES (Advanced Encryption Standard) algorithm in CTR (Counter Mode) mode.

[0151] In a specific implementation, the acquisition device can first set an initial value for the counter. This initial value is typically a random number whose length is generally the same as the block size of the AES algorithm. For example, if the block size of the AES algorithm is typically 128 bits, the initial value of the counter is also 128 bits. A symmetric encryption key is then used, which remains unchanged throughout the encryption and decryption process. When encrypting each block of plaintext, the initial value of the counter is incremented block by block. This increment can be a simple addition of 1. Each target sensor data can be considered a block of plaintext. For example, the target sensor data of sensor A can be considered the first block of plaintext, the target sensor data of sensor B can be considered the second block of plaintext, and the target sensor data of sensor C can be considered the third block of plaintext. Alternatively, the acquisition device can divide each target sensor data into multiple sub-data, each of which can be considered a block of plaintext. For example, the acquisition device divides the target sensor data of sensor A into sub-data A1, A2, and A3. Sub-data A1 can be considered the first block of plaintext, sub-data A2 can be considered the second block of plaintext, and sub-data A3 can be considered the third block of plaintext. Each block of plaintext corresponds to a unique counter value. The acquisition device can input the current counter value into the AES encryption function to obtain the encrypted counter value. It then performs an exclusive-or (XOR) operation on the encrypted counter value with the current plaintext block to be encrypted. The XOR operation is based on the rule that identical values ​​are 0 and different values ​​are 1. This operation generates the corresponding ciphertext block. The acquisition device repeats the above steps of incrementing the counter, encrypting the counter value, and performing the XOR operation for each subsequent plaintext block until all plaintext blocks are encrypted, ultimately obtaining the entire ciphertext (encrypted target sensor data).

[0152] Afterwards, the acquisition device can store the entire ciphertext (encrypted target sensor data) to the hard disk.

[0153] In addition, the embodiment of the present application can also encrypt the target log data. The encryption algorithm can be the AES algorithm in the CTR mode mentioned above. The specific encryption process is the same as the encryption process of the target sensor data mentioned above, and will not be repeated here.

[0154] S250 , using the target log data and the target sensor data stored in the hard disk as autonomous driving data to be analyzed, and uploading the autonomous driving data to the cloud for analysis.

[0155] The hard disk is a long-term storage device of the vehicle. Even when the vehicle is powered off, the data in the hard disk can be permanently saved. Therefore, after the embodiment of the present application stores the encrypted target sensor data in the hard disk, the encrypted target sensor data can be extracted from the hard disk at any time during the testing or operation phase.

[0156] The target log data extracted from the system disk and the encrypted target sensor data extracted from the hard disk can both be used as autonomous driving data to be analyzed, and then uploaded to the cloud and stored to trigger reminders to relevant personnel so that they can analyze the autonomous driving data in a timely manner.

[0157] As can be seen from this example, the solution provided by this application pre-builds a key-value pair mapping table so that the target sensor and target recording time that match the current problem description can be determined based on the key-value pair mapping table. In this way, the recording object is only the target sensor, not all sensors, thereby reducing the data level, and the first time period is determined based on the target recording time and the trigger time of the abnormal autonomous driving function. In this way, only the data of the first time period needs to be recorded, and there is no need to record the data of the full time period or fixed time period, thereby further reducing the data level, thereby greatly alleviating the hard disk storage pressure. In addition, the reduction in data level can shorten the cloud upload time, so real-time backhaul can be achieved, speeding up the analysis time. Furthermore, this application can improve the analysis accuracy by introducing target log data, and the target log data is only the log data of the second time period recorded by the target log file that matches the current problem description, rather than the log data of the full time period or fixed time period recorded by all log files. Therefore, it can ensure that the analysis accuracy can be improved even when the data level is small.

[0158] Furthermore, the present application introduces an encryption algorithm when the target sensor data is written to the disk, and also introduces an encryption algorithm after the target log data is extracted from the system disk, which can provide security protection for the subsequent process of uploading the target sensor data and target log data to the cloud.

[0159] Corresponding to the aforementioned application function implementation method embodiment, the present application also provides an autonomous driving data acquisition device, an electronic device, a computer-readable storage medium and corresponding embodiments.

[0160] Figure 3 It is a structural diagram of the autonomous driving data acquisition device shown in an embodiment of the present application.

[0161] See also Figure 3 , the present application provides an autonomous driving data collection device, which may include:

[0162] The current problem description acquisition module 310 is used to acquire the current problem description of the vehicle when an abnormality in the automatic driving function of the vehicle is detected during the automatic driving process of the vehicle;

[0163] The target recording parameter determination module 320 is used to determine the target recording parameters that match the current problem description from a pre-built key-value pair mapping table; the target recording parameters include the target sensor and the target recording duration;

[0164] A first time period determination module 330 is configured to determine a first time period based on a target recording duration and a triggering time of an abnormality in the autonomous driving function;

[0165] The target sensor data acquisition module 340 is configured to acquire target sensor data acquired by the target sensor within a first time period.

[0166] In one embodiment, the device may further include:

[0167] A target sensor data storage module, used for storing target sensor data into a hard disk;

[0168] The autonomous driving data upload module is used to use the target sensor data stored on the hard disk as the autonomous driving data to be analyzed, so as to upload the autonomous driving data to the cloud for analysis.

[0169] In one embodiment, the key-value pair mapping table includes a plurality of preset question keywords and a designated sensor and a preset recording duration corresponding to each preset question keyword; the target recording parameter determination module 320 may include:

[0170] Target question keyword extraction submodule, used to extract target question keywords from the current question description;

[0171] The first determination submodule of the target recording parameters is used to determine the designated sensor that matches the target question keyword as the target sensor from the key-value pair mapping table, and to determine the preset recording duration that matches the target question keyword as the target recording duration.

[0172] In one embodiment, the target recording parameters further include a target log file and a target interception duration, and the key-value pair mapping table further includes a designated log file and a preset interception duration corresponding to each preset question keyword;

[0173] After extracting target question keywords from the current question description, the target recording parameter determination module 320 may further include:

[0174] The second determination submodule of the target recording parameters is used to determine the specified log file matching the target question keyword as the target log file from the key-value pair mapping table, and to determine the preset interception duration matching the target question keyword as the target interception duration.

[0175] In one embodiment, the device may further include:

[0176] a second time period determination module, configured to determine the second time period based on the target interception duration and the triggering time of the abnormality of the automatic driving function;

[0177] The target log data acquisition module is used to acquire the target log data collected from the target log file within the second time period.

[0178] In one embodiment, the autonomous driving data uploading module may also be used to:

[0179] The target log data and the target sensor data stored on the hard disk are used as the autonomous driving data to be analyzed.

[0180] In one embodiment, the target recording duration includes the first forward recording duration and the total recording duration; the first time period determination module 330 may include:

[0181] The first extension submodule is configured to extend the first forward time length forward based on the triggering time of the abnormal autonomous driving function to obtain a recording starting point;

[0182] The second extension submodule is used to extend the total recording time backward based on the recording start point to obtain the recording end point;

[0183] The first time period definition submodule is configured to determine the time period between the recording start point and the recording end point as the first time period.

[0184] In one embodiment, the target interception duration includes the second forward-looking duration and the total interception duration; and the second time period determination module may include:

[0185] The third extension submodule is used to extend the second forward traceback time forward based on the trigger time of the abnormal autonomous driving function to obtain the interception starting point;

[0186] The fourth extension submodule is used to extend the total interception time backward based on the interception starting point to obtain the interception end point;

[0187] The second time period definition submodule is used to determine the time period between the interception start point and the interception end point as the second time period.

[0188] In one embodiment, the target sensor data acquisition module 340 may include:

[0189] The target sensor data extraction submodule is used to extract the target sensor data collected by the target sensor within the first time period from the annular buffer area; wherein the annular buffer area is used to continuously and rollingly cache the sensor data collected by multiple sensors of the vehicle within a preset buffer time period, and the preset buffer time period includes the first time period.

[0190] In one embodiment, the target sensor data storage module may include:

[0191] The target sensor data storage submodule is used to encrypt the target sensor data and store the encrypted target sensor data in the hard disk.

[0192] In one embodiment, the current problem description obtaining module 310 may include:

[0193] A preset question keyword display submodule is used to display multiple preset question keywords on the vehicle's central control screen when a takeover operation for the vehicle is detected;

[0194] The current problem description receiving submodule is used to receive the current problem description of the vehicle transmitted by the central control screen; wherein, the current problem description is generated by the central control screen based on the target problem keyword selected by the driver from multiple preset problem keywords.

[0195] In another embodiment, the current problem description obtaining module 310 may include:

[0196] The current problem description generating submodule is used to generate a current problem description of the vehicle according to the feedback information received from the chassis system.

[0197] As can be seen from this example, the solution provided by this application, when an abnormality in the vehicle's autonomous driving function is detected during the vehicle's autonomous driving process, obtains the vehicle's current problem description; determines target recording parameters that match the current problem description from a pre-built key-value mapping table; the target recording parameters include a target sensor and a target recording duration; determines a first time period based on the target recording duration and the trigger time of the autonomous driving function abnormality; and obtains target sensor data collected by the target sensor during the first time period. This application pre-builds a key-value mapping table so that the target sensor and target recording duration that match the current problem description are determined based on the key-value mapping table. In this way, the recording object is only the target sensor, not all sensors, thereby reducing the data volume. The first time period is determined based on the target recording duration and the trigger time of the autonomous driving function abnormality. In this way, only the data of the first time period needs to be recorded, and there is no need to record data for the entire time period or a fixed time period, thereby further reducing the data volume and greatly alleviating the hard disk storage pressure. In addition, the reduction in data volume can shorten the cloud upload time, thereby enabling real-time backhaul and accelerating analysis time.

[0198] Regarding the apparatus in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated again here.

[0199] Figure 4 It is a structural diagram of an electronic device shown in an embodiment of the present application.

[0200] See also Figure 4 , the electronic device 400 includes a memory 410 and a processor 420.

[0201] The processor 420 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor.

[0202] Memory 410 may include various types of storage units, such as system memory, read-only memory (ROM), and permanent storage. ROM may store static data or instructions required by processor 420 or other computer modules. Permanent storage may be a readable and writable storage device. Permanent storage may be a non-volatile storage device that retains stored instructions and data even when the computer is powered off. In some embodiments, the permanent storage device utilizes a mass storage device (e.g., a magnetic or optical disk, flash memory). In other embodiments, the permanent storage device may be a removable storage device (e.g., a floppy disk, optical drive). System memory may be a readable and writable storage device or a volatile readable and writable storage device, such as dynamic random access memory (DRAM). System memory may store some or all instructions and data required by the processor during operation. Furthermore, memory 410 may include any combination of computer-readable storage media, including various types of semiconductor memory chips (e.g., DRAM, SRAM, SDRAM, flash memory, programmable read-only memory), as well as magnetic disks and / or optical disks. In some embodiments, the memory 410 may include a readable and / or writable removable storage device, such as a compact disc (CD), a read-only digital versatile disc (e.g., DVD-ROM, double-layer DVD-ROM), a read-only Blu-ray disc, an ultra-density optical disc, a flash memory card (e.g., SD card, mini SD card, Micro-SD card, etc.), a magnetic floppy disk, etc. Computer-readable storage media do not contain carrier waves and transient electronic signals transmitted wirelessly or wired.

[0203] The memory 410 stores executable codes. When the executable codes are processed by the processor 420 , the processor 420 may execute part or all of the above-mentioned methods.

[0204] In addition, the method according to the present application may also be implemented as a computer program or a computer program product, which includes computer program code instructions for executing some or all of the steps in the above method of the present application.

[0205] Alternatively, the present application can also be implemented as a computer-readable storage medium (or non-transitory machine-readable storage medium or machine-readable storage medium) on which executable code (or computer program or computer instruction code) is stored. When the executable code (or computer program or computer instruction code) is executed by a processor of an electronic device (or server, etc.), the processor executes part or all of the steps of the above-mentioned method according to the present application.

[0206] The embodiments of the present application have been described above. The above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is selected to best explain the principles of the embodiments, their practical applications, or improvements to the technology in the market, or to enable other persons skilled in the art to understand the embodiments disclosed herein.

Claims

1. A method for collecting data for autonomous driving, characterized in that: The method comprises: During the automatic driving process of the vehicle, when an abnormality of the automatic driving function of the vehicle is detected, obtaining a description of the current problem of the vehicle; Determining target recording parameters that match the current problem description from a pre-built key-value pair mapping table; the target recording parameters include a target sensor and a target recording duration; determining a first time period according to the target recording duration and the triggering time of the abnormality of the autonomous driving function; Acquiring target sensor data collected by the target sensor within the first time period; The key-value pair mapping table includes a plurality of preset question keywords and a designated sensor and a preset recording duration corresponding to each of the preset question keywords; determining the target recording parameters that match the current question description from the pre-constructed key-value pair mapping table includes: extracting the target question keyword from the current question description; determining the designated sensor that matches the target question keyword as the target sensor from the key-value pair mapping table, and determining the preset recording duration that matches the target question keyword as the target recording duration; The target recording parameters also include a target log file and a target interception time, and the key-value pair mapping table also includes a designated log file and a preset interception time corresponding to each of the preset question keywords; After extracting target question keywords from the current question description, the method further includes: From the key-value pair mapping table, a designated log file that matches the target question keyword is determined as a target log file, and a preset interception duration that matches the target question keyword is determined as a target interception duration.

2. The method according to claim 1, characterized in that The method further comprises: storing the target sensor data in a hard disk; The target sensor data stored in the hard disk is used as the autonomous driving data to be analyzed, so as to upload the autonomous driving data to the cloud for analysis.

3. The method according to claim 2, characterized in that The method further comprises: determining a second time period according to the target interception time and the triggering time of the abnormality of the autonomous driving function; Obtain target log data collected by the target log file during the second time period.

4. The method according to claim 3, characterized in that The method further includes: using the target sensor data stored in the hard disk as the autonomous driving data to be analyzed; The target log data and the target sensor data stored in the hard disk are used as the autonomous driving data to be analyzed.

5. The method according to claim 1, wherein The target recording duration includes the first forward recording duration and the total recording duration; The determining the first time period according to the target recording time and the triggering time of the abnormality of the automatic driving function includes: Taking the triggering time of the abnormal autonomous driving function as a benchmark, extending the first forward time forward to obtain a recording starting point; Taking the recording start point as a reference, the total recording time is extended backward to obtain the recording end point; The time period between the recording start point and the recording end point is determined as a first time period.

6. The method according to claim 3, characterized in that The target interception time includes a second forward time and a total interception time; and determining the second time period according to the target interception time and the triggering time of the abnormal autonomous driving function includes: Taking the triggering time of the abnormal autonomous driving function as a benchmark, extending the second forward traceback time forward to obtain an interception starting point; Taking the interception starting point as a reference, the total interception time is extended backward to obtain the interception end point; A time period between the interception start point and the interception end point is determined as a second time period.

7. The method according to claim 1, characterized in that The acquiring target sensor data collected by the target sensor within the first time period includes: Target sensor data collected by the target sensor during the first time period is extracted from the annular buffer area; wherein the annular buffer area is used to continuously and rollingly cache sensor data collected by multiple sensors of the vehicle during a preset buffer time period, and the preset buffer time period includes the first time period.

8. The method according to claim 2, characterized in that The storing the target sensor data in a hard disk includes: The target sensor data is encrypted, and the encrypted target sensor data is stored in a hard disk.

9. An automatic driving data collection device, characterized in that: The device comprises: A current problem description acquisition module is used to acquire a current problem description of the vehicle when an abnormality in the automatic driving function of the vehicle is detected during the automatic driving process of the vehicle; A target recording parameter determination module is used to determine target recording parameters that match the current problem description from a pre-built key-value pair mapping table; the target recording parameters include a target sensor and a target recording duration; a first time period determination module, configured to determine a first time period according to the target recording duration and a triggering time of an abnormality in the autonomous driving function; a target sensor data acquisition module, configured to acquire target sensor data collected by the target sensor within the first time period; The key-value pair mapping table includes a plurality of preset question keywords and a designated sensor and a preset recording duration corresponding to each of the preset question keywords; the target recording parameter determination module includes: A target question keyword extraction submodule is used to extract target question keywords from the current question description; a first target recording parameter determination submodule, configured to determine, from the key-value pair mapping table, a designated sensor that matches the target question keyword as a target sensor, and to determine a preset recording duration that matches the target question keyword as a target recording duration; The target recording parameters also include a target log file and a target interception time, and the key-value pair mapping table also includes a designated log file and a preset interception time corresponding to each of the preset question keywords; After extracting target question keywords from the current question description, the target recording parameter determination module further includes: The second determination submodule of the target recording parameters is used to determine the specified log file that matches the target question keyword as the target log file from the key-value pair mapping table, and to determine the preset interception time that matches the target question keyword as the target interception time.

10. An electronic device, characterized in that: include: processor; as well as A memory having executable codes stored thereon, which, when executed by the processor, causes the processor to perform the method according to any one of claims 1 to 8.

11. A computer-readable storage medium having executable codes stored thereon, wherein when the executable codes are executed by a processor of an electronic device, the processor is caused to execute the method according to any one of claims 1 to 8.

Images