Data analysis method and device of automatic driving system, vehicle and storage medium

Abstract

The invention provides a data analysis method and device for an automatic driving system, a vehicle and a storage medium. The method comprises the following steps: acquiring end-to-end time delay of an integral link formed by a plurality of service modules; according to the end-to-end time delay, judging whether abnormal time delay occurs in transmission of each service module or not; if yes, acquiring memory monitoring data of the automatic driving system; and according to the memory monitoring data and the exception time delay, determining the exception type of each service module. According to the invention, monitoring of data transmission and scheduling time delay of the service modules is realized on the overall architecture level of data transmission, when the number of the service modules changes, dynamic monitoring of data transmission and scheduling can be dynamically realized in real time, the monitoring is simpler and more convenient, and the monitoring efficiency is improved. And the abnormal type in the data transmission process is positioned based on two dimensions of memory monitoring data and data transmission monitoring, so that research and development personnel can conveniently carry out optimization and improvement of scheduling time delay and data transmission on abnormal time delay, and thus the accuracy of data analysis of the automatic driving system is improved.

Description

Technical Field

[0001] The present invention relates to the field of autonomous driving technology, and in particular to a data analysis method and device, a vehicle and a storage medium for an autonomous driving system. Background Art

[0002] The autonomous driving system includes multiple business modules, each of which runs on a different computing platform. How to optimize the scheduling delay and end-to-end delay is of great significance to optimizing the performance of the autonomous driving system. When performing data optimization analysis on the autonomous driving system, it is necessary to add the time stamp of receiving data and executing the corresponding action in each business module to realize the scheduling of the business modules and the detection of end-to-end delay.

[0003] However, the above method cannot dynamically monitor the latency of the business modules. When the number of business modules changes, the timestamp and execution action within each business module will change, making monitoring difficult, thereby affecting R&D personnel's optimization of the system. Summary of the invention

[0004] The present invention aims to solve at least one of the technical problems existing in the prior art.

[0005] To this end, the first purpose of the present invention is to propose a data analysis method for an autonomous driving system. The method monitors the data transmission and scheduling delays of the business modules at the overall architecture level of data transmission. When the number of business modules changes, dynamic monitoring of data transmission and scheduling can be realized in real time. The monitoring is simpler and more convenient. The abnormal types of the data transmission process are located based on the two dimensions of memory monitoring data and data transmission monitoring, which is convenient for R&D personnel to optimize and improve the scheduling delay and data transmission for abnormal delays, thereby improving the accuracy of data analysis of the autonomous driving system.

[0006] To this end, a second objective of the present invention is to provide a data analysis device for an autonomous driving system.

[0007] To this end, a third object of the present invention is to provide a vehicle.

[0008] To this end, a fourth object of the present invention is to provide a computer-readable storage medium.

[0009] In order to achieve the above-mentioned purpose, an embodiment of the first aspect of the present invention proposes a data analysis method for an autonomous driving system, wherein the autonomous driving system includes multiple business modules, and the method includes: obtaining the end-to-end delay of the multiple business modules constituting the overall link; judging whether an abnormal delay occurs in the transmission of each business module based on the end-to-end delay; if so, obtaining memory monitoring data of the autonomous driving system; determining the type of abnormality in each business module based on the memory monitoring data and the abnormal delay.

[0010] According to the data analysis method of the autonomous driving system of the embodiment of the present invention, data transmission monitoring is added at the software level to detect the overall end-to-end delay of the business module. When the end-to-end delay is abnormal, it is considered that the data transmission and scheduling delay between the business modules may be abnormal. By realizing the monitoring of the data transmission and scheduling delay of the business module at the overall architecture level of data transmission, when the number of business modules changes, dynamic monitoring of data transmission and scheduling can be realized in real time, and the monitoring is simpler and more convenient. In the process of dynamic monitoring, when the abnormal delay occurs, the memory monitoring data of the autonomous driving system is determined, and the memory data is combined with the end-to-end abnormal delay to determine the type of abnormal performance of the autonomous driving system. The type of abnormality in the data transmission process is located based on the two dimensions of memory monitoring data and data transmission monitoring, which is convenient for R&D personnel to optimize and improve the scheduling delay and data transmission for the abnormal delay, thereby improving the accuracy of data analysis of the autonomous driving system.

[0011] In some embodiments, judging whether abnormal delay occurs in transmission of the service module according to the end-to-end delay includes: determining that abnormal delay occurs in transmission of the service module when the end-to-end delay exceeds a preset delay threshold.

[0012] In some embodiments, determining the type of abnormality occurring in the business module based on the memory monitoring data and the abnormal delay includes: obtaining a first-category data ID corresponding to the abnormal delay; and determining the type of abnormality occurring in the business module based on the first-category data ID and the memory monitoring data.

[0013] In some embodiments, determining the type of abnormality of the business module based on the first category data ID and the memory monitoring data includes: determining the second category data ID corresponding to the monitoring data; and determining the type of abnormality of the business module based on the second category data ID and the first category data ID.

[0014] In some embodiments, determining the type of abnormality occurring in the business module based on the second category data ID and the first category data ID includes: determining a first timestamp corresponding to the first category data ID and a second timestamp corresponding to the second category data ID; and performing statistical analysis on the first category data ID and the second category data ID based on the first timestamp and the second timestamp to determine the type of abnormality occurring in the business module.

[0015] In some embodiments, a statistical analysis is performed on the first and second types of data IDs based on the first and second timestamps to determine the type of abnormality in the business module, including: when the first timestamp and the second timestamp are the same and the data transmission deviation corresponding to the first type of data ID and the second type of data ID exceeds a first preset data range, determining that a scheduling delay occurs in the business module; when the first timestamp and the second timestamp are the same and the data transmission deviation corresponding to the first type of data ID and the second type of data ID exceeds a second preset data range, determining that an execution delay abnormality occurs in the business module.

[0016] In some embodiments, after determining the type of abnormality in the business module, it also includes: issuing a prompt message that the autonomous driving system is abnormal.

[0017] In order to achieve the above-mentioned purpose, an embodiment of the second aspect of the present invention proposes a data analysis device for an autonomous driving system, which includes: a first acquisition module, used to obtain the end-to-end delay of the service module; a judgment module, used to judge whether an abnormal delay occurs in the transmission of the service module based on the end-to-end delay; a second acquisition module, used to obtain memory monitoring data of the autonomous driving system; and a determination module, used to determine the type of abnormality of the service module based on the memory monitoring data and the abnormal delay.

[0018] According to the data analysis device of the autonomous driving system of the embodiment of the present invention, data transmission monitoring is added at the software level to detect the overall end-to-end delay of the business module. When the end-to-end delay is abnormal, it is considered that the data transmission and scheduling delay between the business modules may be abnormal. By realizing the monitoring of the data transmission and scheduling delay of the business module at the overall architecture level of data transmission, when the number of business modules changes, dynamic monitoring of data transmission and scheduling can be realized in real time, and the monitoring is simpler and more convenient. In the process of dynamic monitoring, when the abnormal delay occurs, the memory monitoring data of the autonomous driving system is determined, and the memory data is combined with the end-to-end abnormal delay to determine the type of abnormal performance of the autonomous driving system. The type of abnormality in the data transmission process is located based on the two dimensions of memory monitoring data and data transmission monitoring, which is convenient for R&D personnel to optimize and improve the scheduling delay and data transmission for the abnormal delay, thereby improving the accuracy of data analysis of the autonomous driving system and providing support for achieving system performance optimization.

[0019] In order to achieve the above-mentioned purpose, an embodiment of the third aspect of the present invention proposes a vehicle, which includes: a data analysis device of the automatic driving system of the above-mentioned embodiment.

[0020] According to the data analysis system of the autonomous driving system of the embodiment of the present invention, data transmission monitoring is added at the software level to detect the overall end-to-end delay of the business module. When the end-to-end delay is abnormal, it is considered that the data transmission and scheduling delay between the business modules may be abnormal. By realizing the monitoring of the data transmission and scheduling delay of the business module at the overall architecture level of data transmission, when the number of business modules changes, dynamic monitoring of data transmission and scheduling can be realized in real time, and the monitoring is simpler and more convenient. In the process of dynamic monitoring, when the abnormal delay occurs, the memory monitoring data of the autonomous driving system is determined, and the memory data is combined with the end-to-end abnormal delay to determine the type of abnormal performance of the autonomous driving system. The type of abnormality in the data transmission process is located based on the two dimensions of memory monitoring data and data transmission monitoring, which is convenient for R&D personnel to optimize and improve the scheduling delay and data transmission for the abnormal delay, thereby improving the accuracy of data analysis of the autonomous driving system and providing support for achieving system performance optimization.

[0021] In order to achieve the above-mentioned objectives, an embodiment of the fourth aspect of the present invention proposes a non-temporary computer-readable storage medium, on which is a data analysis program for an autonomous driving system. When the data analysis program for an autonomous driving system is executed by a processor, the data analysis method for an autonomous driving system as described in the above-mentioned embodiment is implemented.

[0022] Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] The above and / or additional aspects and advantages of the present invention will become apparent and easily understood from the description of the embodiments in conjunction with the following drawings, in which:

[0024] Figure 1 is a schematic diagram of a data architecture of an autonomous driving system according to an embodiment of the present invention;

[0025] Figure 2 is a schematic diagram of the structure of a data executor according to an embodiment of the present invention;

[0026] Figure 3 is a flow chart of a data analysis method for an autonomous driving system according to an embodiment of the present invention;

[0027] Figure 4 is a schematic diagram of the structure of a scheduling or execution link according to an embodiment of the present invention;

[0028] Figure 5 is a schematic diagram of a structure for obtaining first-category data according to an embodiment of the present invention;

[0029] Figure 6 is a block diagram of a data analysis device for an autonomous driving system according to an embodiment of the present invention;

[0030] Figure 7 is a block diagram of a vehicle according to one embodiment of the present invention. DETAILED DESCRIPTION

[0031] Embodiments of the present invention are described in detail below. The embodiments described with reference to the accompanying drawings are exemplary. Embodiments of the present invention are described in detail below.

[0032] like Figure 1The figure is a schematic diagram of the data architecture of an automatic driving system according to an embodiment of the present invention. The tracing module is set to monitor the delay of the overall link composed of the business modules in the automatic driving system, such as the end-to-end delay of the overall link composed of the perception business module to the control module in the automatic driving system, and when it is determined that the business module transmission is abnormal according to the end-to-end delay, the information is sent to the cloud processing module. When the cloud processing module obtains the above information, the end-to-end delay in the overall link composed of the business module is analyzed to determine the type of abnormal delay in the performance abnormality in the automatic driving system, and the timestamp corresponding to the type of abnormal delay is sent to the display module, such as the Web (World Wide Web, global wide area network) interface display, or the timestamp corresponding to the type of abnormal delay is sent by email to show the type of performance of the automatic driving system to the R&D personnel, such as the scheduling delay or execution delay, so as to facilitate the R&D personnel to optimize and improve the abnormal delay and provide data support for the improvement of the performance of the automatic driving system.

[0033] Among them, the data and algorithms between business modules in the autonomous driving system are abstracted into data modules and actuators. The data information collected by the sensors and the intermediate structure data of the actuators are called data modules. The data in the data modules are all objects of vehicle-side operation data information collection.

[0034] like Figure 2 FIG. 1 is a schematic diagram of the structure of a data executor according to an embodiment of the present invention. Figure 1 and Figure 2 It can be seen that, taking the process from data perception to control as an example, there is a scheduling delay and / or execution delay in the process from the perception business module to the control business module.

[0035] Reference below Figure 3 The data analysis method of the automatic driving system according to the embodiment of the present invention is described as follows: Figure 3 As shown, the data analysis method of the autonomous driving system of an embodiment of the present invention at least includes: steps S1 to S4.

[0036] Step S1, obtaining the end-to-end delay of the overall link composed of multiple service modules.

[0037] Among them, the overall link delay is the delay in the process from perception to control of data. The delay when data goes from the perception business module to the control business module in the autonomous driving system is the end-to-end delay of the overall link. For example, taking sensor perception as an example, after sensing the obstacle ahead through sensors such as video sensors, the delay from sensing the obstacle to trajectory planning to generating control instructions to control the autonomous driving vehicle to move according to the planned trajectory is the end-to-end delay.

[0038] In an embodiment, when the data flow between the business modules in the autonomous driving system is transmitted normally, the Tracing module will record the data generation information of the business module, the data operation information of each time, the time node information of the data when transmitted on different links, the data frequency and the execution module frequency. For example, after the business sensing module senses an obstacle ahead, it sends the obstacle information to the trajectory planning business module. The trajectory planning business module plans the trajectory of the autonomous driving vehicle according to the obstacle information, and after the trajectory planning is completed, sends the planned trajectory to the control business module.

[0039] Among them, when sensing an obstacle, the data ID and timestamp formed when the data is sensed can be determined, as well as the data transmission delay when transmitting the sensed information to the trajectory planning business module and the scheduling delay from the sensed data to the start of trajectory planning; when the trajectory planning is completed, the data transmission delay when the trajectory planning information is transmitted to the control business module and the scheduling delay when the control business module controls the movement of the autonomous driving vehicle according to the control instruction can be determined. By monitoring the transmission time and execution time of each node in the business module, the start time to the end time of the transmission of the operating data from sensing obstacles to controlling the action of the autonomous driving vehicle in different business modules can be determined, so as to realize the end-to-end delay monitoring of the business modules that constitute the overall link.

[0040] like Figure 4 FIG. 1 is a schematic diagram of a scheduling or execution link structure according to an embodiment of the present invention. Figure 4 It can be seen that in an autonomous driving system, there will be multiple execution links, that is, multi-link end-to-end delays. When the end-to-end delay of the autonomous driving system is abnormal, it is believed that there may be scheduling or execution abnormalities in the end-to-end transmission link.

[0041] When the Tracing module monitors the above operating data information, it monitors the above information as a probe, that is, it monitors the end-to-end delay from the entire architecture.

[0042] Step S2: judging whether abnormal delay occurs in the transmission of each service module according to the end-to-end delay.

[0043] Among them, after the Tracing module obtains the end-to-end delay of the entire link of the service module, it detects the end-to-end delay to determine whether the end-to-end delay exceeds the preset delay threshold, for example, whether the end-to-end delay exceeds 20s. If the end-to-end delay exceeds 20s, for example, 23s, it is considered that the end-to-end delay is too long. At this time, it is considered that the transmission between the service modules may be abnormal, for example, the transmission between the perception service module to the control service is abnormal.

[0044] It is understandable that when end-to-end delay is abnormal, there may be multiple abnormal delay situations, such as abnormal scheduling delay and / or abnormal data execution delay. By detecting the end-to-end delay, it is convenient to locate the type of abnormality in the service module according to the end-to-end delay.

[0045] Step S3: If yes, obtain the memory monitoring data of the autonomous driving system.

[0046] Among them, the monitoring information of the autonomous driving system includes at least: CPU information, memory occupancy information and thread occupancy information. The monitoring module will collect the monitoring information of the autonomous driving system, such as CPU information, memory occupancy information and thread occupancy information, and the monitoring module will count the CPU information, memory occupancy information and thread occupancy information, and arrange the above information with timestamp as index.

[0047] In an embodiment, it is determined that an abnormal delay occurs in the end-to-end delay of the overall link. In order to accurately locate the abnormal delay, it is necessary to obtain the memory monitoring data of the autonomous driving system. The abnormal delay is a scheduling delay or an execution delay between business modules. After the abnormal delay occurs, the CPU information, memory occupancy information and thread occupancy information may be abnormal. Therefore, the execution sequence between the business modules is combined with the memory monitoring data of the CPU to locate the type of abnormal delay from different dimensions.

[0048] Step S4, determining the type of abnormality occurring in each service module according to the memory monitoring data and the abnormal delay.

[0049] In an embodiment, after obtaining the abnormal delay and the monitoring information of the autonomous driving system, it is necessary to analyze the above information. For example, the above information is compared according to the same timestamp. If there is a large difference between the abnormal delay and the monitoring information of the autonomous driving system at the same timestamp, it is considered that the performance of the autonomous driving system is abnormal at this timestamp, and it is necessary to conduct a comprehensive analysis of the type of abnormality. For example, at the same timestamp, the data type in the corresponding memory monitoring data of the autonomous driving system, such as abnormal resource consumption and abnormal scheduling between business modules, is aligned with the same timestamp, and the type of abnormality in the business module is specifically analyzed.

[0050] It can be understood that by analyzing abnormal delays such as vehicle-side operating data information and monitoring information of the autonomous driving system, and determining the above information according to the timestamp, the type of abnormality in the business module at the timestamp can be determined, so that it is convenient for R&D personnel to optimize and improve the abnormal delay, so as to provide support for the optimization of autonomous driving system performance, wherein the performance of the autonomous driving system includes data frequency, actuator frequency, actuator scheduling delay, overall link end-to-end delay, and system CPU, memory, temperature monitoring, etc. By optimizing the system performance, that is, optimizing the data frequency, actuator frequency, actuator scheduling delay, full-link end-to-end delay, and system CPU, memory, and temperature monitoring, the end-to-end delay and memory monitoring data are operated within the allowable range, thereby ensuring that the performance of the autonomous driving system is more stable.

[0051] According to the data analysis method of the autonomous driving system of the embodiment of the present invention, data transmission monitoring is added at the software level to detect the overall end-to-end delay of the business module. When the end-to-end delay is abnormal, it is considered that the data transmission and scheduling delay between the business modules may be abnormal. By realizing the monitoring of the data transmission and scheduling delay of the business module at the overall architecture level of data transmission, when the number of business modules changes, dynamic monitoring of data transmission and scheduling can be realized in real time, and the monitoring is simpler and more convenient. In the process of dynamic monitoring, when the abnormal delay occurs, the memory monitoring data of the autonomous driving system is determined, and the memory data is combined with the end-to-end abnormal delay to determine the type of abnormal performance of the autonomous driving system. The type of abnormality in the data transmission process is located based on the two dimensions of memory monitoring data and data transmission monitoring, which is convenient for R&D personnel to optimize and improve the scheduling delay and data transmission of the abnormal delay, thereby improving the accuracy of data analysis of the autonomous driving system and providing support for achieving system performance optimization.

[0052] In some embodiments, determining the type of abnormality in the business module based on memory monitoring data and abnormal delay includes: obtaining a first-category data ID corresponding to the abnormal delay; and determining the type of abnormality in the business module based on the first-category data ID and memory monitoring data.

[0053] In an embodiment, it is determined that an abnormal delay occurs, and a first type of data ID corresponding to the abnormal delay is obtained. The first type of data ID is, for example, data corresponding to a scheduling delay or data corresponding to an execution delay. By obtaining the corresponding data ID, the data corresponding to the scheduling delay and the execution delay are stored according to the ID, so that the corresponding delay is directly determined according to the data ID.

[0054] After obtaining the first type of data ID, the first type of data ID is compared with the memory monitoring data, for example, the first type of data ID is compared with the data ID corresponding to the memory monitoring data to determine the type of abnormality in the business module.

[0055] like Figure 5 FIG. 1 is a schematic diagram of a structure for obtaining the first type of data according to an embodiment of the present invention. Figure 4 and Figure 5 It can be seen that when data A to data E is transmitted through different links, each data message will contain a data ID and a timestamp. When passing through each business module, the input data ID and timestamp will be transparently transmitted, and the structure of the execution link will still be maintained during transparent transmission. The data information at the end of the execution link contains all upstream data IDs and timestamps, that is, data E includes the data IDs and timestamps of data A to D.

[0056] In some embodiments, determining the type of abnormality in a business module based on the first type of data ID and memory monitoring data includes: determining the second type of data ID corresponding to the monitoring data; and determining the type of abnormality in the business module based on the second type of data ID and the first type of data ID.

[0057] In an embodiment, the second type of data ID is the data ID corresponding to the CPU information, memory occupancy information and thread occupancy information, and the abnormal delay in the performance of the autonomous driving system is determined, and the first type of data ID corresponding to the abnormal delay and the second type of data ID corresponding to the monitoring data are determined, for example, the data ID corresponding to the scheduling delay and the data ID corresponding to the first timestamp and the CPU information. By comparing the data IDs at different latitudes, the type of abnormality in the business module is determined.

[0058] In some embodiments, determining the type of abnormality occurring in a business module based on a second-category data ID and a first-category data ID includes: determining a first timestamp corresponding to the first-category data ID and a second timestamp corresponding to the second-category data ID; and performing statistical analysis on the first-category data ID and the second-category data ID based on the first timestamp and the second timestamp to determine the type of abnormality occurring in the business module.

[0059] In an embodiment, when comparing a first category data ID and a second category data ID, a first timestamp corresponding to the first category data ID and a second timestamp corresponding to the second category data ID are determined, and the first category data ID and the second category data ID are analyzed corresponding to the same timestamp to determine the type of abnormality in the business module.

[0060] In some embodiments, a statistical analysis is performed on the first type of data ID and the second type of data ID based on the first timestamp and the second timestamp to determine the type of abnormality in the business module, including: when the first timestamp and the second timestamp are the same and the data transmission deviation corresponding to the first type of data ID and the second type of data ID exceeds the first preset data range, determining that the business module has a scheduling delay; when the first timestamp and the second timestamp are the same and the data transmission deviation corresponding to the first type of data ID and the second type of data ID exceeds the second preset data range, determining that the business module has an execution delay abnormality.

[0061] In an embodiment, when determining the type of abnormality of the performance of the autonomous driving system at the current timestamp, the timestamps may be aligned, that is, when the first timestamp and the second timestamp are the same, the scheduling delay or execution delay data and the target monitoring information corresponding to the first type of data ID are comprehensively analyzed to determine the type of abnormality of the business module. When analyzing the above information, it is necessary to detect in sequence according to the timestamp order.

[0062] Specifically, the first preset data range is the data range of the scheduling delay data corresponding to the preset first type of data ID and the memory data corresponding to the second type of data when the scheduling delay is normal. When the first timestamp and the second timestamp are the same and the data transmission deviation between the first type of data ID and the second type of data ID is greater than the first preset data range, it is considered that the autonomous driving system has a scheduling delay anomaly, for example, the scheduling delay from the generation of the trajectory of the trajectory planning business module to the start of execution of the above trajectory is abnormal; when the first timestamp and the second timestamp are the same and the data transmission deviation between the first type of data ID and the second type of data ID is within the first preset data range, it is considered that the autonomous driving system has no scheduling delay anomaly.

[0063] The second preset data range is the data range of the execution delay data corresponding to the first type of data ID and the memory data corresponding to the second type of data when the execution delay is normal. When the first timestamp and the second timestamp are the same and the data transmission deviation between the first type of data ID and the second type of data ID is greater than the second preset data range, it is considered that the autonomous driving system has an execution delay anomaly, for example, an anomaly occurs from the execution of the trajectory planning business module to the end of execution; when the first timestamp and the second timestamp are the same and the data transmission deviation between the first type of data ID and the second type of data ID is within the first preset data range, it is considered that the autonomous driving system has no execution delay anomaly.

[0064] In some embodiments, after determining the type of abnormality in the business module, it also includes: issuing a prompt message of the abnormality of the autonomous driving system.

[0065] In an embodiment, after determining that an abnormality occurs in the performance of the autonomous driving system, it is considered that a scheduling abnormality or an execution abnormality occurs in the performance of the autonomous driving system. At this time, a prompt message of the abnormal performance of the autonomous driving system is issued, and the prompt message is sent to a display module such as a Web interface for display, so as to show the timestamp when the system performance abnormality occurs and the abnormal delay corresponding to the timestamp to the R&D personnel, so that the R&D personnel can clearly know the system performance at the timestamp.

[0066] According to the data analysis method of the autonomous driving system of the embodiment of the present invention, data transmission monitoring is added at the software level to detect the overall end-to-end delay of the business module. When the end-to-end delay is abnormal, it is considered that the data transmission and scheduling delay between the business modules may be abnormal. By realizing the monitoring of the data transmission and scheduling delay of the business module at the overall architecture level of data transmission, when the number of business modules changes, dynamic monitoring of data transmission and scheduling can be realized in real time, and the monitoring is simpler and more convenient. In the process of dynamic monitoring, when the abnormal delay occurs, the memory monitoring data of the autonomous driving system is determined, and the memory data is combined with the end-to-end abnormal delay to determine the type of abnormal performance of the autonomous driving system. The type of abnormality in the data transmission process is located based on the two dimensions of memory monitoring data and data transmission monitoring, which is convenient for R&D personnel to optimize and improve the scheduling delay and data transmission of the abnormal delay, thereby improving the accuracy of data analysis of the autonomous driving system and providing support for achieving system performance optimization.

[0067] like Figure 6 As shown, the data analysis device 2 of the autonomous driving system of the embodiment of the present invention includes a first acquisition module 20, a judgment module 21, a second acquisition module 22 and a determination module 23, wherein the first acquisition module 20 is used to obtain the end-to-end delay of the service module; the judgment module 21 is used to determine whether an abnormal delay occurs in the transmission of the service module according to the end-to-end delay; the second acquisition module 22 is used to obtain the memory monitoring data of the autonomous driving system; the determination module 23 is used to determine the type of abnormality of the service module according to the memory monitoring data and the abnormal delay.

[0068] According to the data analysis device 2 of the automatic driving system of the embodiment of the present invention, by adding data transmission monitoring at the software level, the overall end-to-end delay of the business module is detected. When the end-to-end delay is abnormal, it is considered that the data transmission and scheduling delay between the business modules may be abnormal. By realizing the monitoring of the data transmission and scheduling delay of the business module at the overall architecture level of data transmission, when the number of business modules changes, the dynamic monitoring of data transmission and scheduling can be realized in real time, and the monitoring is simpler and more convenient. In the process of dynamic monitoring, it is determined when the abnormal delay occurs, the memory monitoring data of the automatic driving system is obtained, and the memory data is combined with the end-to-end abnormal delay to determine the type of abnormal performance of the automatic driving system. The type of abnormality in the data transmission process is located based on the two dimensions of memory monitoring data and data transmission monitoring, which is convenient for R&D personnel to optimize and improve the scheduling delay and data transmission of the abnormal delay, thereby improving the accuracy of data analysis of the automatic driving system and providing support for achieving system performance optimization.

[0069] Next, a vehicle 3 according to an embodiment of the present invention will be described.

[0070] like Figure 7 As shown, the vehicle 3 of the embodiment of the present invention includes the data analysis device 2 of the automatic driving system of the above embodiment.

[0071] According to the vehicle 3 of the autonomous driving system of the embodiment of the present invention, data transmission monitoring is added at the software level to detect the overall end-to-end delay of the business module. When the end-to-end delay is abnormal, it is considered that the data transmission and scheduling delay between the business modules may be abnormal. By realizing the monitoring of the data transmission and scheduling delay of the business module at the overall architecture level of data transmission, when the number of business modules changes, dynamic monitoring of data transmission and scheduling can be realized in real time, and the monitoring is simpler and more convenient. In the process of dynamic monitoring, when the abnormal delay occurs, the memory monitoring data of the autonomous driving system is determined, and the memory data is combined with the end-to-end abnormal delay to determine the type of abnormal performance of the autonomous driving system. The type of abnormality in the data transmission process is located based on the two dimensions of memory monitoring data and data transmission monitoring, which is convenient for R&D personnel to optimize and improve the scheduling delay and data transmission of the abnormal delay, thereby improving the accuracy of data analysis of the autonomous driving system and providing support for achieving system performance optimization.

[0072] The non-temporary computer-readable storage medium of an embodiment of the present invention, the data analysis program of the autonomous driving system on the non-temporary computer-readable storage medium, and the data analysis program of the autonomous driving system, when executed by a processor, implement the data analysis method of the autonomous driving system as mentioned in the above embodiment.

[0073] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "illustrative embodiments", "examples", "specific examples", or "some examples" means that the specific features, structures, materials, or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example.

[0074] Although the embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the claims and their equivalents.

Claims

1. A data analysis method for an autonomous driving system, characterized in that: The autonomous driving system includes multiple business modules, and the method includes: Obtaining the end-to-end delay of the overall link composed of multiple service modules; Determining whether abnormal delay occurs in the transmission of each of the service modules according to the end-to-end delay; If so, obtaining memory monitoring data of the autonomous driving system; The type of abnormality occurring in each of the service modules is determined according to the memory monitoring data and the abnormal delay.

2. The data analysis method of the automatic driving system according to claim 1, characterized in that: Judging whether abnormal delay occurs in transmission of the service module according to the end-to-end delay includes: When the end-to-end delay exceeds a preset delay threshold, it is determined that an abnormal delay occurs in the service module transmission.

3. The data analysis method of the automatic driving system according to claim 1, characterized in that: Determining the type of abnormality of the service module according to the memory monitoring data and the abnormal delay includes: Obtain the first type of data ID corresponding to the abnormal delay; The type of abnormality occurring in the business module is determined according to the first type data ID and the memory monitoring data.

4. The data analysis method of the automatic driving system according to claim 3, characterized in that: Determining the type of abnormality of the business module according to the first type of data ID and the memory monitoring data includes: Determine the second type of data ID corresponding to the monitoring data; The type of the business module abnormality is determined according to the second-category data ID and the first-category data ID.

5. The data analysis method of the automatic driving system according to claim 4, characterized in that: Determining the type of abnormality of the business module according to the second-category data ID and the first-category data ID includes: Determine a first timestamp corresponding to the first type of data ID and a second timestamp corresponding to the second type of data ID; The first category data ID and the second category data ID are statistically analyzed according to the first timestamp and the second timestamp to determine the type of abnormality in the business module.

6. The data analysis method of the automatic driving system according to claim 5, characterized in that: Performing statistical analysis on the first category data ID and the second category data ID according to the first timestamp and the second timestamp to determine the type of abnormality in the business module includes: When the first timestamp and the second timestamp are the same and the data transmission deviation corresponding to the first type data ID and the second type data ID exceeds a first preset data range, determining that a scheduling delay exception occurs in the service module; When the first timestamp and the second timestamp are the same and the data transmission deviation corresponding to the first type data ID and the second type data ID exceeds a second preset data range, it is determined that an execution delay exception occurs in the business module.

7. The data analysis method of the automatic driving system according to claim 1, characterized in that: After determining the type of abnormality in the business module, the method further includes: A prompt message indicating that the autonomous driving system is abnormal is issued.

8. A data analysis device for an automatic driving system, characterized in that: include: A first acquisition module, used to acquire the end-to-end delay of the service module; A judgment module, used to judge whether abnormal delay occurs in the transmission of the service module according to the end-to-end delay; A second acquisition module, used to acquire memory monitoring data of the autonomous driving system; A determination module is used to determine the type of abnormality that occurs in the business module based on the memory monitoring data and the abnormal delay.

9. A vehicle, characterized in that: include: A data analysis device for an autonomous driving system as claimed in claim 8.

10. A non-transitory computer-readable storage medium, characterized in that: The data analysis program of the autonomous driving system on the non-temporary computer-readable storage medium, when the data analysis of the autonomous driving system is executed by the processor, implements the data analysis method of the autonomous driving system as described in any one of claims 1-7.

Images