CN115766861A - Automatic driving data closed loop system, method and medium for vehicle software upgrade - Google Patents
Automatic driving data closed loop system, method and medium for vehicle software upgrade Download PDFInfo
- Publication number
- CN115766861A CN115766861A CN202211431339.8A CN202211431339A CN115766861A CN 115766861 A CN115766861 A CN 115766861A CN 202211431339 A CN202211431339 A CN 202211431339A CN 115766861 A CN115766861 A CN 115766861A
- Authority
- CN
- China
- Prior art keywords
- data
- model
- module
- vehicle
- training
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 53
- 238000012549 training Methods 0.000 claims abstract description 178
- 238000012360 testing method Methods 0.000 claims abstract description 95
- 238000004088 simulation Methods 0.000 claims abstract description 67
- 238000012795 verification Methods 0.000 claims abstract description 65
- 238000004422 calculation algorithm Methods 0.000 claims description 71
- 238000011156 evaluation Methods 0.000 claims description 23
- 238000000605 extraction Methods 0.000 claims description 23
- 238000013500 data storage Methods 0.000 claims description 18
- 238000011161 development Methods 0.000 claims description 15
- 238000003860 storage Methods 0.000 claims description 15
- 238000002372 labelling Methods 0.000 claims description 14
- 238000009826 distribution Methods 0.000 claims description 9
- 238000013507 mapping Methods 0.000 claims description 7
- 238000011157 data evaluation Methods 0.000 claims description 4
- 238000007418 data mining Methods 0.000 claims description 4
- 238000013079 data visualisation Methods 0.000 claims description 4
- 238000013480 data collection Methods 0.000 abstract description 10
- 230000008569 process Effects 0.000 description 20
- 238000012545 processing Methods 0.000 description 12
- 238000010586 diagram Methods 0.000 description 11
- 238000004590 computer program Methods 0.000 description 9
- 230000006870 function Effects 0.000 description 8
- 238000004140 cleaning Methods 0.000 description 7
- 230000008676 import Effects 0.000 description 6
- 238000004891 communication Methods 0.000 description 5
- 238000012544 monitoring process Methods 0.000 description 5
- 238000006243 chemical reaction Methods 0.000 description 4
- 238000000586 desensitisation Methods 0.000 description 4
- 230000008901 benefit Effects 0.000 description 3
- 238000003745 diagnosis Methods 0.000 description 3
- 238000007726 management method Methods 0.000 description 3
- 238000004519 manufacturing process Methods 0.000 description 3
- 230000004044 response Effects 0.000 description 3
- 230000000007 visual effect Effects 0.000 description 3
- 230000005540 biological transmission Effects 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 238000010276 construction Methods 0.000 description 2
- 238000007405 data analysis Methods 0.000 description 2
- 238000013461 design Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 238000013439 planning Methods 0.000 description 2
- 238000003491 array Methods 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 238000012550 audit Methods 0.000 description 1
- 230000006835 compression Effects 0.000 description 1
- 238000007906 compression Methods 0.000 description 1
- 238000013499 data model Methods 0.000 description 1
- 238000009795 derivation Methods 0.000 description 1
- 230000003203 everyday effect Effects 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 230000008447 perception Effects 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 230000001737 promoting effect Effects 0.000 description 1
- 238000011158 quantitative evaluation Methods 0.000 description 1
- 238000012827 research and development Methods 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 230000001953 sensory effect Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012384 transportation and delivery Methods 0.000 description 1
- 238000010200 validation analysis Methods 0.000 description 1
- 238000012800 visualization Methods 0.000 description 1
Images
Landscapes
- Stored Programmes (AREA)
Abstract
An autopilot data closed loop system, method and medium for vehicle software upgrade is disclosed. The system comprises: the method comprises the steps that vehicle end data corresponding to the intelligent vehicle are collected through a vehicle end data collection module, a data service module, a model training module, a simulation verification module and an intelligent system upgrading module are deployed on a cloud platform, a target model of which a model test result accords with a preset model verification standard is obtained based on the vehicle end data, and then the target model is sent to the intelligent vehicle so as to control the intelligent vehicle to carry out software upgrading. According to the technical scheme, the data service module, the model training module, the simulation verification module and the intelligent system upgrading module are deployed on the cloud platform, so that vehicle-side data can be universally circulated in the modules, the data circulation speed can be increased, the uploading and downloading times of the data can be reduced, and the network bandwidth and the flow pressure can be reduced.
Description
Technical Field
The present invention relates to the field of autopilot technology, and more particularly, to an autopilot data closed loop system, method, and medium for vehicle software upgrade.
Background
Data is a brand new production element in the digital era, and the rapid development of the artificial intelligence industry is being promoted due to the fusion of the data, an algorithm and calculation power. In the process of landing application of intelligent vehicle software, a large number of extreme scenes can be met, and the extreme scenes are the main factors influencing the safety of the intelligent vehicle. The method solves the problem of vehicle safety in an extreme scene, and can not leave the support of a large amount of data, so that an automatic driving data closed-loop system is constructed to efficiently acquire and utilize data, and the data circulation speed is improved, thereby becoming a key element for promoting the iterative upgrade of intelligent vehicle software.
At present, an existing automatic driving data closed-loop system generally comprises a plurality of sub-modules, such as an algorithm development platform, a scene extraction platform, a data annotation platform, an algorithm training platform, a simulation test platform, a client and the like, wherein the sub-modules are deployed on a local server. However, for the prior art, in the aspect of the application range, the data closed-loop process is limited to a certain small link such as vehicle-road cooperation, data acquisition, algorithm training, simulation test and the like, in different small links, data are heterogeneous and cannot be universally circulated, the operation of the whole automatic driving data closed-loop system cannot be run through, and the speed of the intelligent vehicle software landing application is limited. In addition, in the aspect of environment deployment, model development and training are performed on the local server, and a collected data set of the cloud needs to be frequently uploaded and downloaded, so that network traffic and bandwidth are greatly wasted.
Disclosure of Invention
The invention provides an automatic driving data closed-loop system, a method and a medium for vehicle software upgrading, which can enable vehicle-end data to automatically complete data format conversion in each module of a cloud platform, and the vehicle-end data is universally circulated in each module, so that the data circulation speed can be increased, the uploading and downloading times of the data can be reduced, and the network bandwidth and the flow pressure can be reduced.
According to one aspect of the invention, an automatic driving data closed-loop system for vehicle software upgrading is provided, which comprises a data acquisition module deployed on an intelligent vehicle, a data service module, a model training module, a simulation verification module and an intelligent system upgrading module, wherein the data service module, the model training module, the simulation verification module and the intelligent system upgrading module are deployed on a cloud platform;
the data acquisition module is connected with the data service module and used for acquiring vehicle-end data corresponding to the intelligent vehicle according to a current acquisition rule and sending the vehicle-end data to the data service module;
the data service module is respectively connected with the model training module and the simulation verification module and is used for carrying out data labeling on the vehicle-end data to obtain a training data set and sending the training data set to the model training module; carrying out scene extraction on the vehicle end data to obtain a scene data set, and sending the scene data set to the simulation verification module;
the model training module is connected with the simulation verification module and used for acquiring an initial algorithm model, training the initial algorithm model based on the training data set, acquiring a trained target model and sending the target model to the simulation verification module;
the simulation verification module is connected with the intelligent system upgrading module and used for carrying out model test on the target model based on the scene data set and sending the target model to the intelligent system upgrading module when detecting that a model test result meets a preset model verification standard;
and the intelligent system upgrading module is used for sending the target model to the intelligent vehicle so as to control the intelligent vehicle to carry out software upgrading.
Optionally, the data service module is further configured to, when it is detected that the current data acquisition requirement is updated, obtain an updated acquisition rule according to the updated data acquisition requirement, and send the updated acquisition rule to the data acquisition module;
the data acquisition module is further configured to obtain updated vehicle-side data corresponding to the intelligent vehicle according to the updated acquisition rule, and send the updated vehicle-side data to the data service module.
Optionally, the data service module is further configured to perform at least one of data cleaning, data diagnosis, data mining, data visualization, data retrieval, data evaluation, and data auditing.
Optionally, the data service module includes a data storage unit;
the data storage unit is used for storing the training data set and the scene data set.
Optionally, the model training module includes a data set extraction unit, an algorithm development unit, a task training unit and a model factory unit;
the data set extraction unit is connected with the task training unit and is used for extracting a training data set from the data storage unit and sending the training data set to the task training unit;
the algorithm development unit is connected with the task training unit and used for responding to a configuration instruction of an initial algorithm model, acquiring the initial algorithm model and sending the initial algorithm model to the task training unit;
the task training unit is connected with the model factory unit and used for responding to a configuration instruction of a model training task, generating a model training task, training the initial algorithm model based on the training data set according to the model training task, acquiring a trained target model and sending the target model to the model factory unit;
the model factory unit is used for responding to a version configuration instruction corresponding to a target model, acquiring version information corresponding to the target model, generating a mapping relation between the target model and the version information, and storing the mapping relation between the target model and the version information in a preset database.
Optionally, the intelligent system upgrade module includes an upgrade control gateway and a file download gateway;
the upgrade control gateway is connected with the file download gateway and is used for acquiring version information corresponding to the target model from a preset database when a software upgrade instruction corresponding to the intelligent vehicle is detected; acquiring current version information corresponding to the intelligent vehicle, generating an upgrade file distribution instruction corresponding to the intelligent vehicle when detecting that the current version information is different from the version information corresponding to the target model, and sending the upgrade file distribution instruction corresponding to the intelligent vehicle to the file download gateway;
and the file downloading gateway is used for sending the target model to the intelligent vehicle when receiving an upgrading file distribution instruction corresponding to the intelligent vehicle so as to control the intelligent vehicle to carry out software upgrading.
According to another aspect of the present invention, there is provided an automatic driving data closed-loop method for vehicle software upgrade, which is applied to an automatic driving data closed-loop system for vehicle software upgrade according to any embodiment of the present invention, and includes:
the method comprises the steps that vehicle end data corresponding to an intelligent vehicle are obtained through a data acquisition module according to a current acquisition rule, data labeling is conducted on the vehicle end data through a data service module to obtain a training data set, and scene extraction is conducted on the vehicle end data to obtain a scene data set;
acquiring an initial algorithm model through a model training module, training the initial algorithm model based on the training data set, and acquiring a trained target model;
and carrying out model test on the target model based on the scene data set through a simulation verification module, and sending the target model to the intelligent vehicle through an intelligent system upgrading module when a model test result is detected to meet a preset model verification standard so as to control the intelligent vehicle to carry out software upgrading.
According to another aspect of the present invention, there is provided a computer readable storage medium storing computer instructions for causing a processor to implement an autopilot data closed loop method for vehicle software upgrade according to any one of the embodiments of the present invention when executed.
According to the technical scheme of the embodiment of the invention, vehicle-end data corresponding to the intelligent vehicle is obtained through a data acquisition module according to a current acquisition rule, data labeling is carried out on the vehicle-end data through a data service module so as to obtain a training data set, and scene extraction is carried out on the vehicle-end data so as to obtain a scene data set; then, obtaining an initial algorithm model through a model training module, training the initial algorithm model based on a training data set, and obtaining a trained target model; furthermore, model testing is carried out on the target model through the simulation verification module based on the scene data set, when the fact that the model testing result meets the preset model verification standard is detected, the target model is sent to the intelligent vehicle through the intelligent system upgrading module to control the intelligent vehicle to carry out software upgrading, the data service module, the model training module, the simulation verification module and the intelligent system upgrading module are deployed on the cloud platform, data format conversion of vehicle-end data can be automatically completed in each module of the cloud platform, the vehicle-end data can be universally circulated in each module, the data circulation speed can be accelerated, the uploading and downloading times of the data can be reduced, and the network bandwidth and the flow pressure can be reduced.
It should be understood that the statements in this section do not necessarily identify key or critical features of the embodiments of the present invention, nor do they necessarily limit the scope of the invention. Other features of the present invention will become apparent from the following description.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings needed to be used in the description of the embodiments will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without creative efforts.
FIG. 1A is a schematic structural diagram of an automatic driving data closed-loop system for vehicle software upgrade according to an embodiment of the present invention;
fig. 1B is a schematic flow chart of a closed loop for data acquisition according to an embodiment of the present invention;
fig. 1C is a schematic structural diagram of a data service module according to an embodiment of the present invention;
fig. 1D is a schematic structural diagram of another data service module according to an embodiment of the present invention;
FIG. 1E is a schematic structural diagram of a model training module according to an embodiment of the present invention;
FIG. 1F is a schematic diagram of another model training module according to an embodiment of the present invention;
fig. 1G is a schematic structural diagram of an intelligent system upgrade module according to an embodiment of the present invention;
FIG. 1H is a schematic structural diagram of another intelligent system upgrade module according to an embodiment of the present invention;
FIG. 1I is a schematic diagram of a deployment structure of another automatic driving data closed-loop system for vehicle software upgrading according to an embodiment of the present invention;
FIG. 1J is a schematic flow chart of an autonomous driving data closed loop according to an embodiment of the present invention;
FIG. 2A is a flowchart of an automated driving data closed-loop method for vehicle software upgrade according to a second embodiment of the present invention;
FIG. 2B is a schematic flow chart of a model test according to the second embodiment of the present invention;
FIG. 2C is a schematic flow chart of model development and testing according to a second embodiment of the present invention;
fig. 2D is a schematic diagram of a full-link closed loop according to the second embodiment of the present invention.
Detailed Description
In order to make the technical solutions of the present invention better understood, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
It should be noted that the terms "first," "second," "target," and the like in the description and claims of the present invention and in the above-described drawings are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used is interchangeable under appropriate circumstances such that the embodiments of the invention described herein are capable of operation in sequences other than those illustrated or described herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
Example one
Fig. 1A is a schematic structural diagram of an automatic driving data closed-loop system for vehicle software upgrade, which is provided in an embodiment of the present invention, and the automatic driving data closed-loop system 100 for vehicle software upgrade may include a data acquisition module 111 disposed in an intelligent vehicle 110, and a data service module 121, a model training module 122, a simulation verification module 123, and an intelligent system upgrade module 124 disposed in a cloud platform 120. The intelligent vehicle 110 may be a vehicle installed with intelligent vehicle software.
The data acquisition module 111 is connected to the data service module 121, and is configured to acquire vehicle-side data corresponding to the intelligent vehicle 110 according to a current acquisition rule, and send the vehicle-side data to the data service module 121. The current collection rule may be a current collection rule for vehicle-side data, for example, a collection period, specific scene data, sensor data, and the like, and may be issued by the data service module 121. The data types of the vehicle-end data can comprise image data, point cloud data, video data, bus data, log data and the like. The data acquisition module 111 may transmit the vehicle-side data to the data service module 121 by using wireless transmission, hard disk express, and Web/App (Application program) import.
Specifically, the data collection module 111 may collect data on various types of sensors, such as a laser radar sensor, a millimeter wave radar sensor, a camera, and the like, deployed on the smart vehicle 110 and on the bus by using a data collection application program to obtain vehicle-side data, and may upload the vehicle-side data to the data service module 121 of the cloud-side platform 120 according to the data type and a preset data type priority.
The data service module 121 is respectively connected to the model training module 122 and the simulation verification module 123, and is configured to perform data tagging on the vehicle-end data to obtain a training data set, and send the training data set to the model training module 122; extracting scenes from the vehicle-end data to obtain a scene data set, and sending the scene data set to the simulation verification module 123;
the data service module 121 may be configured to receive the vehicle-side data sent by the data acquisition module 111, and perform operations such as classification, storage, processing, and extraction on the vehicle-side data, so as to provide data support for other modules.
In this embodiment, the data service module 121 may first perform preprocessing on the vehicle-side data, such as data parsing, unpacking, desensitization, cleaning, and the like, and perform data labeling and scene extraction on the preprocessed vehicle-side data to generate a training data set used by a training model and a scene data set used by simulation verification. Further, the training data set may be sent to the model training module 122, and the scenario data set may be sent to the simulation verification module 123.
It should be noted that the data service module 121 may be preconfigured with a data upload real-time monitoring program, and after the data acquisition module 111 finishes uploading all data every day, the program script may automatically trigger data tagging and scene extraction to finish processing of uploaded data, and generate a tagged data set required by model training and a scene data set required by simulation verification. All data processing can be automatically completed by an algorithm program, and after the data processing is completed to generate a data set, model increment training and testing can also be automatically completed by a script, namely, the whole process from data acquisition and processing to data increment training and verification can be completed manually without any intervention.
The advantage of above-mentioned setting lies in, through the source of backstage real time monitoring data to the whole process that automatic triggering data processing, data model trained to model simulation verification not only can liberate manpower resources, can also accelerate data circulation speed, thereby can fall to the ground for intelligent vehicle 110's volume production operation provides more powerful guarantee.
The model training module 122 is connected to the simulation verification module 123, and is configured to obtain an initial algorithm model, train the initial algorithm model based on the training data set, obtain a trained target model, and send the target model to the simulation verification module 123.
The initial algorithm model may include an algorithm model that is constructed based on a machine learning algorithm and does not have any function, or may be a history model obtained through previous training, and the history model may be stored in the data center of the data service module 121.
In this embodiment, the model training module 122 may create an initial algorithm model by algorithm creation, import, and algorithm version configuration. The initial algorithm model may then be supervised trained using a training data set to obtain a trained target model, and the target model is sent to the simulation verification module 123. Alternatively, the historical model and the historical training data set corresponding to the historical model may be searched from the data service module 121, and then, the historical model may be iteratively trained together based on the current training data set and the historical training data set to obtain the target model.
The simulation verification module 123 is connected to the intelligent system upgrade module 124, and is configured to perform a model test on the target model based on the scene data set, and send the target model to the intelligent system upgrade module 124 when it is detected that a model test result meets a preset model verification standard.
It should be noted that, under data driving, before issuing the generated target model to a real vehicle for testing, performing simulation verification on the target model is also a crucial link. The simulation verification can realize the construction of working conditions as required and improve the coverage of a test scene, thereby improving the safety factor of real vehicle test, reducing the time and cost of intelligent driving system test and regression, and improving the quantitative evaluation level of the algorithm version based on large-scale test results.
In this embodiment, the simulation verification module 123 may be configured to detect the availability and reliability of the target model; for example, a series of model tests, such as model-on-loop, software-on-loop, hardware-on-loop, vehicle-on-loop, etc., may be performed on the target model based on the scene data set to verify whether the model is capable of adapting to different driving scenarios. If the target model successfully passes a series of model tests, the model test results may be determined to meet the preset model validation criteria, and the target model may be sent to the intelligent system upgrade module 124. The model test can be composed of a simulation test, a virtual-real combination test and a test evaluation.
Optionally, after the model test is completed, the simulation verification module 123 may also automatically generate a test report and a model evaluation report corresponding to the target model according to the model test result. Wherein, the report template of the test report and the model evaluation report can be preset.
The advantage of the above-mentioned setting lies in, based on the simulation verification module 123 of deployment at high in the clouds platform 120, can be through using emulation technique, establish out each type scene of vehicle driving fast, even including dangerous driving scene to can avoid the problem that the collection cycle that leads to by factors such as weather, vehicle in the real vehicle collection faces is long, the scene collection security is poor.
In the embodiment, different from the prior art in which a training and simulation platform is built on a local server, the model training module 122 and the simulation verification module 123 are deployed on the cloud platform 120, so that on one hand, the problem that the local storage space is insufficient due to the fact that a data set is increasingly huge in the operation of the intelligent vehicle 110 can be solved; on the other hand, the occupation of network bandwidth and the consumption of network flow in the process of frequently downloading and uploading data between the cloud and the local can be reduced.
The intelligent system upgrade module 124 is configured to send the target model to the intelligent vehicle 110, so as to control the intelligent vehicle 110 to perform software upgrade.
In this embodiment, the intelligent system upgrading module 124 may issue the target model to the intelligent vehicle 110 after the target model is evaluated, so as to control the intelligent vehicle 110 to perform remote upgrading of the intelligent vehicle software.
The advantage of the above arrangement is that through the intelligent system upgrading module 124, after the intelligent driving vehicle is operated in a large scale in the later period, the intelligent vehicle 110 can be subjected to the upgrading iterative management of the models and the software in batches, and the safety and the reliability of the intelligent vehicle can be continuously improved.
In this embodiment, a full-flow data closed loop from data acquisition, data processing, model training, simulation testing to intelligent vehicle software upgrading can be realized, which is different from a small closed loop system in which a data closed loop is concentrated on a certain aspect in the prior art. Secondly, from the perspective of data driving, the data is connected in series with the whole system, so that the data circulation speed can be increased, the problem of 'data long tail' faced by the intelligent vehicle 110 is solved, the coverage of the intelligent vehicle 110 on complex scenes, multi-limit scenes and problem scenes can be improved, and the safety and the reliability of the intelligent vehicle 110 can be improved.
According to the technical scheme of the embodiment of the invention, vehicle-end data corresponding to the intelligent vehicle is obtained through a data acquisition module according to a current acquisition rule, data labeling is carried out on the vehicle-end data through a data service module so as to obtain a training data set, and scene extraction is carried out on the vehicle-end data so as to obtain a scene data set; then, obtaining an initial algorithm model through a model training module, training the initial algorithm model based on a training data set, and obtaining a trained target model; furthermore, model testing is carried out on the target model through the simulation verification module based on the scene data set, when the fact that the model testing result meets the preset model verification standard is detected, the target model is sent to the intelligent vehicle through the intelligent system upgrading module to control the intelligent vehicle to carry out software upgrading, the data service module, the model training module, the simulation verification module and the intelligent system upgrading module are deployed on the cloud platform, data format conversion of vehicle-end data can be automatically completed in each module of the cloud platform, the vehicle-end data can be universally circulated in each module, the data circulation speed can be accelerated, the uploading and downloading times of the data can be reduced, and the network bandwidth and the flow pressure can be reduced.
Optionally, the data service module 121 is further configured to, when it is detected that the current data acquisition requirement is updated, obtain an updated acquisition rule according to the updated data acquisition requirement, and send the updated acquisition rule to the data acquisition module 111;
the data collection module 111 is further configured to obtain updated vehicle-side data corresponding to the intelligent vehicle 110 according to the update collection rule, and send the updated vehicle-side data to the data service module 121.
In this embodiment, the whole flow of the automatic driving data closed loop may be divided into a plurality of sub-data closed loops according to each sub-link, and typically, the whole flow of the automatic driving data closed loop may include a data acquisition closed loop. As shown in fig. 1B, first, the intelligent vehicle 110 may collect vehicle sensor data and part of scene tag information through the vehicle-side deployed data collection module 111, and may upload the collected data to the cloud deployed data service module 121 through the vehicle ethernet. The data service module 121 may then store and calculate the vehicle-side data uploaded by the data acquisition module 111 (e.g., data cleaning, data desensitization, etc.), and apply the calculated vehicle-side data to model training and simulation verification.
During data application, new data acquisition requirements may arise, for example, acquiring data of a typical scene; at this time, the data service module 121 may obtain the updated data acquisition requirement, and convert the updated data acquisition requirement into an updated acquisition rule; further, the data service module 121 may issue the updated collection rule to the data collection module 111. Finally, the data collection module 111 may complete collection of the typical scene data based on the obtained updated collection rule, and upload the collected typical scene data to the data service module 121. From this, based on vehicle end collection upload, high in the clouds data application, the update download collection demand and accomplish data acquisition, can form the data acquisition closed loop.
Optionally, the data service module 121 is further configured to perform at least one of data cleaning, data diagnosis, data mining, data visualization, data retrieval, data evaluation, and data auditing.
In a specific example, the structure of the data service module 121 may be as shown in fig. 1C, wherein the functions of the data service module 121 may include data storage and core services; core services that may be used to obtain data, evaluate data, and process data; acquiring data, which can comprise data acquisition, data caching, data retrieval and the like; evaluating data, which can comprise data evaluation, data audit and the like; and processing the data, wherein the processing can comprise data cleaning, data labeling, data analysis, data mining, data diagnosis, data visualization, scene extraction and the like.
Optionally, as shown in fig. 1D, the data service module 121 may include a data storage unit 1211; the data storage unit 1211 is configured to store the training data set and the scene data set.
In a specific example, as shown in fig. 1C, the data storage unit 1211 may include different types of databases to store different types of data. For example, the data storage unit 1211 may include a raw database, a simulation database, a data annotation database, a scene library dataset, a static model database, an algorithm model database, and the like. In this embodiment, the vehicle-end data collected by the data collection module 111, the data (e.g., training data set, scene data set, etc.) processed by the data service module 121, various types of algorithm models generated by the model training module 122, and the simulation verification result, verification report, etc. obtained by the simulation verification module 123 may all be stored in the data storage unit 1211, and meanwhile, the data storage unit 1211 may receive the data retrieval request of each module, and may feed back the stored data matching the data retrieval request to the corresponding module.
Optionally, as shown in fig. 1E, the model training module 122 may include a data set extracting unit 1221, an algorithm developing unit 1222, a task training unit 1223, and a model factory unit 1224;
the data set extracting unit 1221 is connected to the task training unit 1223, and is configured to extract a training data set from the data storage unit 1211 and send the training data set to the task training unit 1223; specifically, the data set extraction unit 1221 may retrieve a training data set required for training the current model from the data storage unit 1211.
The algorithm development unit 1222 is connected to the task training unit 1223, and configured to, in response to a configuration instruction for an initial algorithm model, obtain the initial algorithm model, and send the initial algorithm model to the task training unit 1223; in this embodiment, the user may choose to perform code import or code online editing to send configuration instructions for the initial algorithm model to the algorithm development unit 1222. After receiving the configuration instruction for the initial algorithm model, the algorithm development unit 1222 may construct a machine learning model without any function based on the configuration information to serve as the current initial algorithm model, or search a corresponding historical model from the data service module 121 to serve as the current initial algorithm model, and may send the initial algorithm model to the task training unit 1223 for model training.
The task training unit 1223 is connected to the model factory unit 1224, and is configured to generate a model training task in response to a configuration instruction for the model training task, train the initial algorithm model based on the training data set according to the model training task, obtain a trained target model, and send the trained target model to the model factory unit 1224.
In this embodiment, a user may select to configure model training task information, for example, a model training termination condition, a training data set used, and the like, so as to generate a configuration instruction for the model training task and send the configuration instruction to the task training unit 1223, and the task training unit 1223 may generate the model training task according to the model training task information set by the user; alternatively, the task training unit 1223 may automatically generate the model training task according to preset model training task information.
Then, the task training unit 1223 may perform model training on the initial algorithm model by using the corresponding training data set according to the model training task until a trained target model is obtained. In addition, the task training unit 1223 may also allocate training resources and specify training rules for the model training task, and may perform full-process visual monitoring on the model training process.
The model factory unit 1224 is configured to, in response to a version configuration instruction corresponding to a target model, obtain version information corresponding to the target model, generate a mapping relationship between the target model and the version information, and store the mapping relationship between the target model and the version information in a preset database. The version configuration instruction corresponding to the target model may be an instruction of a user to configure version information of the target model.
In this embodiment, the model factory unit 1224 may obtain version information corresponding to the target model according to the version configuration instruction corresponding to the target model, and generate a mapping relationship between the target model and the corresponding version information to be stored in the preset database of the data storage unit 1211. The model plant unit 1224 may also reason about the target model and manage the results of the model reasoning.
In one specific example, the structure of the model training module 122 may be as shown in FIG. 1F; the method comprises a data part, an algorithm/model, a training framework, a training task and training management, wherein the type of the algorithm/model can comprise a perception class, a decision planning class and a control class, the training task can comprise evaluation indexes, task planning and algorithm evaluation, and the training management can comprise functions of training monitoring, process visualization, log generation, report derivation and the like.
In addition, the model training module 122 may preset a typical training framework, such as tensrflow, pytorch, etc., to support the import and online editing of algorithm code, and complete the model training process and the model inference process by configuring a training task. In the model training process, a real-time visual training monitoring result and a training log can be provided, and the training log and a report can be exported after model training is completed.
Optionally, as shown in fig. 1G, the intelligent system upgrade module 124 may include an upgrade control gateway 1241 and a file download gateway 1242;
the upgrade control gateway 1241 is connected to the file download gateway 1242, and is configured to obtain version information corresponding to the target model from a preset database when detecting a software upgrade instruction corresponding to the intelligent vehicle 110; acquiring current version information corresponding to the intelligent vehicle 110, generating an upgrade file distribution instruction corresponding to the intelligent vehicle 110 when detecting that the current version information is different from the version information corresponding to the target model, and sending the upgrade file distribution instruction corresponding to the intelligent vehicle 110 to the file download gateway 1242;
the file download gateway 1242 is configured to, when receiving an upgrade file distribution instruction corresponding to the intelligent vehicle 110, send the target model to the intelligent vehicle 110, so as to control the intelligent vehicle 110 to perform software upgrade.
In this embodiment, the intelligent system upgrade module 124 may periodically collect the software version information of the intelligent vehicle 110 and the cloud recommended software version information, so as to download the file of the intelligent driving software of the intelligent vehicle 110 and update the software version when it is determined that the intelligent vehicle 110 needs to perform software upgrade.
In a specific example, the structure of the intelligent system upgrade module 124 may be as shown in fig. 1H, where the intelligent system upgrade module 124 plays a role of starting and ending in a data-driven closed loop, on one hand, current configuration information of the cloud-end intelligent vehicle 110 needs to be acquired, and on the other hand, an update request and an upgrade package need to be issued for the intelligent vehicle 110 to be upgraded. The intelligent system upgrade module 124 may include an upgrade control gateway 1241 and a file download gateway 1242, where the upgrade control gateway 1241 is configured to undertake communication with a cloud end to complete acquisition of control information including a vehicle type, a component, a version, an upgrade task, and the file download gateway 1242 is configured to be responsible for distributing an upgrade file (target model) to the intelligent vehicle 110.
In a specific embodiment of the present embodiment, the deployment structure of the automated driving data closed loop system 100 for vehicle software upgrade may be as shown in fig. 1I. Wherein, the flow of the automatic driving data closed loop can be as shown in fig. 1J; first, the data collection module 111 uploads the vehicle-side data collected from the smart vehicle 110 to a data center, i.e., a data storage unit 1211, through wireless transmission, hard disk express delivery, web/App import, and the like, where the data center is managed by the data service module 121. The data service module 121 performs grading and classified storage on the acquired vehicle-side data, performs data analysis, unpacking, desensitization and cleaning on the vehicle-side data, further labels and mines the processed data to generate a training data set used for model training and a scene data set used for simulation verification, and stores the training data set and the scene data set to a data center.
Thereafter, the model training module 122 may extract a training data set from the data center to complete algorithm training, so as to obtain a target model, and the model generated in the training process may also be periodically backed up in the data center. Further, the simulation verification module 123 may obtain the trained target model, extract a scene data set for testing from the data center, complete various tests mainly based on the simulation in-loop test, and provide an evaluation result and a test report for the model after the tests. Finally, when it is determined that the target model meets the verification criteria, the target model may be updated into the intelligent vehicle software through the intelligent system upgrade module 124 to complete update iterations of the vehicle-side software, thereby completing vehicle-side data acquisition to a complete data-driven autopilot data closed loop for the data-driven intelligent vehicle 110 to upgrade.
It should be noted that, in the current environment, the smart car software has certain difficulties in personnel, vehicle collection, weather, and scene construction. For example, in terms of personnel, each collection vehicle needs to be provided with a driver and data collection personnel to collect effective scenes; in the aspect of vehicle collection, the collection scene period is long, the same scene is difficult to reproduce, and the scene with lower safety factor cannot be collected in real vehicles; in the aspect of weather, for a low-frequency scene with extreme weather, the acquisition is difficult, and the period is longer; in the aspect of scene processing, the development period of an extreme scene is long, and the vehicle-end cloud verification is difficult. Due to the factors, the acquisition time and the personnel cost are increased sharply, and the whole limit scene is difficult to cover, so that the potential safety hazard brought to the intelligent vehicle 110 is difficult to eliminate, and the intelligent driving technology engineering landing is severely restricted.
In view of the above problems, in this embodiment, an automatic driving data closed-loop system for iterative upgrade of intelligent vehicle software is constructed to efficiently collect and use data, so that a data cycle speed can be increased, and thus, the safety of the intelligent vehicle 110 can be improved, and the intelligent vehicle 110 can meet mass production conditions as soon as possible.
Example two
Fig. 2A is a flowchart of an automatic driving data closed-loop method for vehicle software upgrade according to a second embodiment of the present invention. The method can be applied to the situation of iterative upgrade of intelligent vehicle software based on automatic driving data closed loop, and can be executed by the automatic driving data closed loop system for vehicle software upgrade, which is described in the first embodiment of the invention. As shown in fig. 2A, the method includes:
s210, vehicle end data corresponding to the intelligent vehicle are obtained through a data acquisition module according to a current acquisition rule, data labeling is carried out on the vehicle end data through a data service module to obtain a training data set, and scene extraction is carried out on the vehicle end data to obtain a scene data set.
The data annotation can include at least one of machine vision annotation, video annotation, continuous multi-frame annotation, point cloud annotation and voice annotation. The machine vision labeling can comprise labeling of classification of targets such as lane lines, pedestrians, vehicles and the like; video annotation, which can include frame extraction, content extraction, etc.; the continuous multiframe marking can comprise target tracking; point cloud labeling, which may include object classification, lane line identification, and the like; the voice label can comprise a label of semantics and part of speech, and the like.
S220, obtaining an initial algorithm model through a model training module, training the initial algorithm model based on the training data set, and obtaining a trained target model.
In a specific example, the model training can comprise four steps of data scheduling, algorithm development, task training and model factory; the data scheduling step interacts with the data service module 121 to complete the extraction of the training data set; an algorithm development step, which provides functions of algorithm creation, import, algorithm version configuration and the like; the task training step supports the functions of mirror image configuration, task scheduling, resource monitoring and the like; the model factory steps support functions of model evaluation, model compression, model version configuration and the like. After the model is generated, the above process data may be persistently stored in a corresponding database, and uniformly scheduled and managed by the data service module 121. Through the above steps, the model training module 122 can complete a complete process from algorithm training to model generation.
And S230, performing model test on the target model based on the scene data set through a simulation verification module, and sending the target model to the intelligent vehicle through an intelligent system upgrading module when a model test result is detected to meet a preset model verification standard so as to control the intelligent vehicle to perform software upgrading.
The model test can comprise a simulation test, a virtual-real combination test and a test evaluation; the simulation test may include at least one of a model simulation test, a software simulation test, a hardware-in-the-loop test, a vehicle-in-the-loop test, and a driver-in-the-loop test.
In one specific example, the flow of model testing may be as shown in FIG. 2B. The model test mainly comprises a simulation test, a virtual-real combination test and a test evaluation; the simulation test sequentially corresponds to a model simulation test (MiL), a software simulation test (Sil), a hardware-in-the-loop test (HiL), a vehicle-in-the-loop test (ViL) and a driver-in-the-loop test (DiL) from a model, a code, a controller, a vehicle and a driver. The virtual-real combination test is mainly combined with a vehicle in-loop test and a driver in-loop test to enrich vehicle dynamics and a driver model, so that a model reference is provided for test evaluation. The test evaluation is the operation evaluation of the whole test, if the evaluation is passed, the codes of the whole vehicle, the related algorithms and the models can be upgraded to the intelligent vehicle 110 through the intelligent system upgrading module 124; if the evaluation fails, the algorithm model needs to be updated iteratively.
The preset model verification standard may be preset condition information for determining whether the model can be evaluated through testing. If the model test result of the model meets the preset model verification standard, the model can be determined to successfully pass the test evaluation.
In this embodiment, the simulation verification module 123 may sequentially complete tests of different granularity levels of the intelligent vehicle software from a model, software, hardware, a vehicle, and a driver process, and the test indexes may be configured in test evaluation, and the indexes may include multiple aspects such as laws, regulations, functional correctness, driving experience, economy, and safety. The test evaluation can give corresponding evaluation results according to the test of the process, and only the model passing the test evaluation is possible to be downloaded and updated into the intelligent vehicle software.
In a specific embodiment of this embodiment, the model development and testing process can be as shown in fig. 2C. First, data acquisition, data desensitization and data cleaning are performed through the data service module 121, and data labeling and scene generation are performed on the processed data. Then, in the research and development process, after a training data set generated after data labeling passes through the model training module 122, a series of steps of data set extraction, training task creation, algorithm submission, training execution and model inference evaluation are completed, and finally a model is generated.
Further, in the test flow, after the scene data set generated after the data processing passes through the simulation verification module 123, a series of steps of model extraction, simulation task creation, automated testing, and simulation evaluation are completed, and finally an evaluation report is generated.
In this embodiment, as shown in fig. 2D, under the driving of data, the vehicle end and the cloud end can jointly open a full link closed-loop design from data acquisition, data uploading, data storage, data calculation, model training, simulation verification to intelligent system upgrading.
According to the technical scheme of the embodiment of the invention, vehicle-end data corresponding to the intelligent vehicle is obtained through a data acquisition module according to a current acquisition rule, data labeling is carried out on the vehicle-end data through a data service module so as to obtain a training data set, and scene extraction is carried out on the vehicle-end data so as to obtain a scene data set; further, an initial algorithm model is obtained through a model training module, and the initial algorithm model is trained on the basis of a training data set to obtain a trained target model; finally, performing model test on the target model through the simulation verification module based on the scene data set, and sending the target model to the intelligent vehicle through the intelligent system upgrading module when detecting that a model test result meets a preset model verification standard so as to control the intelligent vehicle to perform software upgrading; by arranging the data service module, the model training module, the simulation verification module and the intelligent system upgrading module on the cloud platform, data format conversion of vehicle-side data in each module of the cloud platform can be automatically completed, the data can be universally circulated in each module, the data circulation speed can be increased, the uploading and downloading times of the data can be reduced, and the network bandwidth and the flow pressure can be reduced.
It should be noted that, in the technical solution of the present embodiment, the acquisition, storage, application, and the like of the personal information of the related user all meet the regulations of the relevant laws and regulations, and do not violate the common customs of the public order.
EXAMPLE III
The third embodiment of the present invention further provides a computer-readable storage medium, where computer instructions are stored, and the computer instructions are used for enabling a processor to implement the automatic driving data closed-loop method for vehicle software upgrading according to the second embodiment of the present invention when the processor executes the computer instructions.
In some embodiments, the automated driving data closed loop method for vehicle software upgrade may be implemented as a computer program tangibly embodied in a computer-readable storage medium, such as a memory unit. In some embodiments, part or all of the computer program may be loaded and/or installed onto the electronic device via the ROM and/or the communication unit. When the computer program is loaded into RAM and executed by the processor, one or more steps of the autopilot data closed loop method for vehicle software upgrade described above may be performed. Alternatively, in other embodiments, the processor may be configured by any other suitable means (e.g., by way of firmware) to perform an autopilot data closed loop method for vehicle software upgrades.
Various implementations of the systems and techniques described here above may be implemented in digital electronic circuitry, integrated circuitry, field Programmable Gate Arrays (FPGAs), application Specific Integrated Circuits (ASICs), application Specific Standard Products (ASSPs), system on a chip (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include: implemented in one or more computer programs that are executable and/or interpretable on a programmable system including at least one programmable processor, which may be special or general purpose, receiving data and instructions from, and transmitting data and instructions to, a storage system, at least one input device, and at least one output device.
A computer program for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus, such that the computer programs, when executed by the processor, cause the functions/acts specified in the flowchart and/or block diagram block or blocks to be performed. A computer program can execute entirely on a machine, partly on the machine, as a stand-alone software package, partly on the machine and partly on a remote machine or entirely on the remote machine or server.
In the context of the present invention, a computer-readable storage medium may be a tangible medium that can contain, or store a computer program for use by or in connection with an instruction execution system, apparatus, or device. A computer readable storage medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. Alternatively, the computer readable storage medium may be a machine readable signal medium. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.
To provide for interaction with a user, the systems and techniques described here can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to a user; and a keyboard and a pointing device (e.g., a mouse or a trackball) by which a user can provide input to the electronic device. Other kinds of devices may also be used to provide for interaction with a user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form, including acoustic, speech, or tactile input.
The systems and techniques described here can be implemented in a computing system that includes a back-end component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front-end component (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such back-end, middleware, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: local Area Networks (LANs), wide Area Networks (WANs), blockchain networks, and the internet.
It should be understood that various forms of the flows shown above, reordering, adding or deleting steps, may be used. For example, the steps described in the present invention may be executed in parallel, sequentially, or in different orders, and are not limited herein as long as the desired results of the technical solution of the present invention can be achieved.
The above-described embodiments should not be construed as limiting the scope of the invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions may be made in accordance with design requirements and other factors. Any modification, equivalent replacement, and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (10)
1. An automatic driving data closed-loop system for vehicle software upgrading is characterized by comprising a data acquisition module, a data service module, a model training module, a simulation verification module and an intelligent system upgrading module, wherein the data acquisition module is arranged on an intelligent vehicle;
the data acquisition module is connected with the data service module and used for acquiring vehicle-end data corresponding to the intelligent vehicle according to a current acquisition rule and sending the vehicle-end data to the data service module;
the data service module is respectively connected with the model training module and the simulation verification module and is used for carrying out data annotation on the vehicle-end data to obtain a training data set and sending the training data set to the model training module; carrying out scene extraction on the vehicle end data to obtain a scene data set, and sending the scene data set to the simulation verification module;
the model training module is connected with the simulation verification module and used for acquiring an initial algorithm model, training the initial algorithm model based on the training data set, acquiring a trained target model and sending the target model to the simulation verification module;
the simulation verification module is connected with the intelligent system upgrading module and used for carrying out model test on the target model based on the scene data set and sending the target model to the intelligent system upgrading module when detecting that a model test result meets a preset model verification standard;
and the intelligent system upgrading module is used for sending the target model to the intelligent vehicle so as to control the intelligent vehicle to carry out software upgrading.
2. The system of claim 1, wherein the data service module is further configured to, when it is detected that there is an update in the current data acquisition requirement, obtain an updated acquisition rule according to the updated data acquisition requirement, and send the updated acquisition rule to the data acquisition module;
the data acquisition module is further configured to obtain updated vehicle-side data corresponding to the intelligent vehicle according to the updated acquisition rule, and send the updated vehicle-side data to the data service module.
3. The system of claim 2, wherein the data service module is further configured to perform at least one of data cleansing, data diagnostics, data mining, data visualization, data retrieval, data evaluation, and data auditing.
4. The system of claim 1, wherein the data service module comprises a data storage unit;
the data storage unit is used for storing the training data set and the scene data set.
5. The system of claim 4, wherein the model training module comprises a data set extraction unit, an algorithm development unit, a task training unit, and a model factory unit;
the data set extraction unit is connected with the task training unit and used for extracting a training data set from the data storage unit and sending the training data set to the task training unit;
the algorithm development unit is connected with the task training unit and used for responding to a configuration instruction of an initial algorithm model, acquiring the initial algorithm model and sending the initial algorithm model to the task training unit;
the task training unit is connected with the model factory unit and used for responding to a configuration instruction of a model training task, generating a model training task, training the initial algorithm model based on the training data set according to the model training task, acquiring a trained target model and sending the target model to the model factory unit;
the model factory unit is used for responding to a version configuration instruction corresponding to a target model, obtaining version information corresponding to the target model, generating a mapping relation between the target model and the version information, and storing the mapping relation between the target model and the version information in a preset database.
6. The system of claim 5, wherein the intelligent system upgrade module comprises an upgrade control gateway and a file download gateway;
the upgrade control gateway is connected with the file download gateway and is used for acquiring version information corresponding to the target model from a preset database when a software upgrade instruction corresponding to the intelligent vehicle is detected; acquiring current version information corresponding to the intelligent vehicle, generating an upgrade file distribution instruction corresponding to the intelligent vehicle when detecting that the current version information is different from the version information corresponding to the target model, and sending the upgrade file distribution instruction corresponding to the intelligent vehicle to the file download gateway;
and the file downloading gateway is used for sending the target model to the intelligent vehicle when receiving an upgrading file distribution instruction corresponding to the intelligent vehicle so as to control the intelligent vehicle to carry out software upgrading.
7. An automatic driving data closed-loop method for vehicle software upgrading, which is applied to the automatic driving data closed-loop system for vehicle software upgrading of any one of claims 1-6, and comprises the following steps:
the method comprises the steps that vehicle end data corresponding to an intelligent vehicle are obtained through a data acquisition module according to a current acquisition rule, data labeling is conducted on the vehicle end data through a data service module to obtain a training data set, and scene extraction is conducted on the vehicle end data to obtain a scene data set;
acquiring an initial algorithm model through a model training module, training the initial algorithm model based on the training data set, and acquiring a trained target model;
and carrying out model test on the target model based on the scene data set through a simulation verification module, and sending the target model to the intelligent vehicle through an intelligent system upgrading module when a model test result is detected to meet a preset model verification standard so as to control the intelligent vehicle to carry out software upgrading.
8. The method of claim 7, wherein the data annotation comprises at least one of a machine vision annotation, a video annotation, a continuous multi-frame annotation, a point cloud annotation, and a voice annotation.
9. The method of claim 7, wherein the model tests include simulation tests, virtual-real combination tests, and test evaluations; the simulation test comprises at least one of a model simulation test, a software simulation test, a hardware-in-loop test, a vehicle-in-loop test and a driver-in-loop test.
10. A computer readable storage medium storing computer instructions for causing a processor to implement the automated driving data closed loop method for vehicle software upgrade of any one of claims 7-9 when executed.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211431339.8A CN115766861A (en) | 2022-11-14 | 2022-11-14 | Automatic driving data closed loop system, method and medium for vehicle software upgrade |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211431339.8A CN115766861A (en) | 2022-11-14 | 2022-11-14 | Automatic driving data closed loop system, method and medium for vehicle software upgrade |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115766861A true CN115766861A (en) | 2023-03-07 |
Family
ID=85371545
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211431339.8A Pending CN115766861A (en) | 2022-11-14 | 2022-11-14 | Automatic driving data closed loop system, method and medium for vehicle software upgrade |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115766861A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116126740A (en) * | 2023-04-18 | 2023-05-16 | 小米汽车科技有限公司 | Model in-loop test method and device, computer equipment and storage medium |
| CN116340192A (en) * | 2023-05-31 | 2023-06-27 | 小米汽车科技有限公司 | Vehicle perception version effect evaluation method and device and vehicle |
| CN117456797A (en) * | 2023-12-26 | 2024-01-26 | 成都运达科技股份有限公司 | Method, system and storage medium for simulating driving training connection |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114379581A (en) * | 2021-11-29 | 2022-04-22 | 江铃汽车股份有限公司 | Algorithm iteration system and method based on automatic driving |
| CN114511827A (en) * | 2021-11-30 | 2022-05-17 | 重庆长安汽车股份有限公司 | Intelligent driving-assisted vehicle cloud perception closed-loop processing method |
| CN114741731A (en) * | 2022-04-27 | 2022-07-12 | 重庆长安汽车股份有限公司 | Intelligent networking automobile data training method based on key information anonymization, electronic equipment and computer readable storage medium |
-
2022
- 2022-11-14 CN CN202211431339.8A patent/CN115766861A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114379581A (en) * | 2021-11-29 | 2022-04-22 | 江铃汽车股份有限公司 | Algorithm iteration system and method based on automatic driving |
| CN114511827A (en) * | 2021-11-30 | 2022-05-17 | 重庆长安汽车股份有限公司 | Intelligent driving-assisted vehicle cloud perception closed-loop processing method |
| CN114741731A (en) * | 2022-04-27 | 2022-07-12 | 重庆长安汽车股份有限公司 | Intelligent networking automobile data training method based on key information anonymization, electronic equipment and computer readable storage medium |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116126740A (en) * | 2023-04-18 | 2023-05-16 | 小米汽车科技有限公司 | Model in-loop test method and device, computer equipment and storage medium |
| CN116340192A (en) * | 2023-05-31 | 2023-06-27 | 小米汽车科技有限公司 | Vehicle perception version effect evaluation method and device and vehicle |
| CN117456797A (en) * | 2023-12-26 | 2024-01-26 | 成都运达科技股份有限公司 | Method, system and storage medium for simulating driving training connection |
| CN117456797B (en) * | 2023-12-26 | 2024-03-15 | 成都运达科技股份有限公司 | Method, system and storage medium for simulating driving training connection |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11733693B2 (en) | Data acquisition method and apparatus | |
| CN115766861A (en) | Automatic driving data closed loop system, method and medium for vehicle software upgrade | |
| CN111338973B (en) | Scene-based automatic driving simulation test evaluation service cloud platform and application method | |
| CN110765661A (en) | Automatic driving simulation scene generation method and device, electronic equipment and storage medium | |
| US10296302B1 (en) | One-click deployment of industrial software | |
| CN116052300B (en) | Digital twinning-based power inspection system and method | |
| CN104123847A (en) | ACP (artificial societies, computational experiments and parallel execution) method-based traffic signal recommendation system and corresponding method | |
| CN114187412A (en) | High-precision map generation method and device, electronic equipment and storage medium | |
| CN114168213A (en) | Software release method and device based on Jenkins and electronic equipment | |
| CN115437339A (en) | Remote diagnosis method and device, electronic equipment and storage medium | |
| CN114997414B (en) | Data processing method, device, electronic equipment and storage medium | |
| CN111373406A (en) | Accelerated simulation setup procedure using a priori knowledge extraction for problem matching | |
| CN115878681A (en) | Method and device for acquiring automatic driving data, storage medium and electronic device | |
| CN117215551A (en) | LLM-based low-code application development method and system | |
| CN116720489A (en) | Page filling method and device, electronic equipment and computer readable storage medium | |
| CN107919998B (en) | Sensor server function testing method and system based on JMeter | |
| CN110716778A (en) | Application compatibility testing method, device and system | |
| CN115935089A (en) | RPA technology-based boarding point recommendation optimization method and system | |
| CN111859370B (en) | Method, apparatus, electronic device and computer readable storage medium for identifying service | |
| KR20220082554A (en) | Customized Autonomous Artificial Intelligence Platform Using Manufacturing Process Data | |
| CN112729319A (en) | Automatic data acquisition and analysis system and method | |
| CN118134119B (en) | Railway construction monitoring information platform based on digital primitive | |
| CN118278822B (en) | Working data acquisition method and system based on image data analysis and electronic equipment | |
| CN116993118A (en) | Service customization method, device, equipment and medium | |
| Bakker‐Reynolds et al. | Implementation and Evaluation of Computer Vision Prototype for Vehicle Detection |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |