Training method and system of automatic driving model, electronic equipment and storage medium

Abstract

The invention relates to the field of automatic driving, in particular to an automatic driving model training method and system, electronic equipment and a storage medium. The training method of the automatic driving model is applied to the computer equipment. The training method comprises the following steps: acquiring a current sample acquisition strategy of the automatic driving model; transmitting the sample collection strategy to the vehicle-mounted equipment; the vehicle-mounted equipment is used for screening target driving data based on a sample acquisition strategy and transmitting the target driving data to the computer equipment; receiving target driving data; adding the target driving data into a training set of an automatic driving model; and training the automatic driving model based on the training set of the automatic driving model. According to the invention, cloud software and hardware burdens can be reduced, and the training efficiency of the automatic driving model is improved.

Description

Technical Field

[0001] The present application relates to the field of autonomous driving, and specifically to a training method, system, electronic device and storage medium for an autonomous driving model. Background Art

[0002] During the training process of the autonomous driving model, the vehicle will collect massive amounts of data and upload it to the cloud. The cloud will store the data uploaded by the vehicle and then use the data to train the autonomous driving model.

[0003] However, since the amount of driving data collected by vehicles is relatively large, uploading it to the cloud requires a lot of bandwidth and software and hardware. Moreover, generally speaking, there is less data that can be used to optimize the autonomous driving model. The massive amount of data uploaded by vehicles will result in low training efficiency of the autonomous driving model. Summary of the invention

[0004] In view of the above, an embodiment of the present application provides a training method for an autonomous driving model, which aims to reduce the burden of cloud software and hardware and improve the training efficiency of the autonomous driving model.

[0005] In a first aspect, an embodiment of the present application provides a training method for an autonomous driving model, which is applied to a computer device, wherein the computer device is communicatively connected to a vehicle-mounted device, and the vehicle-mounted device is configured in a vehicle. The training method for the autonomous driving model includes:

[0006] Obtaining the current sample collection strategy of the autonomous driving model;

[0007] The sample collection strategy is transmitted to the vehicle-mounted device; the vehicle-mounted device is used to filter the target driving data based on the sample collection strategy, and transmit the target driving data to the computer device;

[0008] receiving the target driving data;

[0009] Adding the target driving data to a training set of the autonomous driving model;

[0010] And training the autonomous driving model based on the training set of the autonomous driving model.

[0011] In the process of training the autonomous driving model in the embodiment of the present application, the on-board device uploads the driving data that meets the sample collection strategy to the computer device, which can not only reduce the software and hardware resources occupied by the driving data transmission, save the software and hardware resources used by the computer device to store the driving data, and reduce the workload of data cleaning and screening of the computer device, but also improve the training efficiency of the autonomous driving model.

[0012] In some embodiments, the sample collection strategy includes a driving scenario of the target driving data, and obtaining the current sample collection strategy of the autonomous driving model includes:

[0013] Obtaining a vehicle control effect of the autonomous driving model in a driving scenario to be tested;

[0014] If the vehicle control effect of the automatic driving model in the driving scene to be tested does not meet the preset safe driving standards, the driving scene to be tested is used as the driving scene of the target driving data.

[0015] The embodiments of the present application can formulate a sample collection strategy based on the vehicle control effect of the autonomous driving model in the driving scenario to be tested, thereby improving the vehicle control effect of the autonomous driving model in various driving scenarios and further improving the training efficiency of the autonomous driving model.

[0016] In some embodiments, obtaining the vehicle control effect of the autonomous driving model in the driving scenario to be tested includes:

[0017] Acquire a test sample in a driving scenario to be tested;

[0018] Inputting the test sample into the autonomous driving model to obtain a vehicle control instruction corresponding to the driving scenario to be tested;

[0019] Detect whether the execution of the vehicle control instruction in the driving scenario to be tested meets the preset safe driving standard to obtain the vehicle control effect.

[0020] In some embodiments, after the autonomous driving model is trained based on the training set of the autonomous driving model, the method further includes:

[0021] Continue to enter the step of obtaining the vehicle control effect of the autonomous driving model in the driving scenario to be tested until the vehicle control effect of the autonomous driving model in the driving scenario to be tested meets the preset safe driving standard.

[0022] In a second aspect, an embodiment of the present application further provides a training method for an autonomous driving model, which is applied to an on-board device, wherein the on-board device is configured on a vehicle, and the on-board device is communicatively connected to a computer device, and the training method for the autonomous driving model includes:

[0023] Collecting driving data of the vehicle during driving;

[0024] Receiving a current sample collection strategy of the autonomous driving model from the computer device;

[0025] Based on the sample collection strategy, target driving data is screened out from the collected driving data;

[0026] The target driving data is transmitted to the computer device, and the computer device is used to add the target driving data to the training set of the autonomous driving model; and train the autonomous driving model based on the training set of the autonomous driving model.

[0027] In some embodiments, the sample collection strategy includes a driving scenario of the target driving data, and the step of filtering out the target driving data from the collected driving data based on the sample collection strategy includes:

[0028] Acquire a driving scene of the collected driving data;

[0029] Based on the driving scenario of the collected driving data, driving data that meets the sample collection strategy is screened from the collected driving data to obtain the target driving data.

[0030] In some embodiments, the acquiring the driving scene of the collected driving data includes:

[0031] The collected driving data is input into a pre-trained scene recognition model to obtain a driving scene of the collected driving data.

[0032] In a third aspect, the embodiment of the present application further provides a training system for an autonomous driving model, including:

[0033] A computer device, wherein the computer device is used to execute the training method of the autonomous driving model described in the first aspect;

[0034] At least one vehicle, wherein the vehicle is equipped with an on-board device, wherein the on-board device is communicatively connected to the computer device, and wherein the on-board device is used to execute the training method of the autonomous driving model described in the second aspect.

[0035] In a fourth aspect, an embodiment of the present application further provides an electronic device, comprising a processor and a memory, wherein the memory is used to store instructions, and the processor is used to call the instructions in the memory so that the electronic device executes the training method of the autonomous driving model described in the first aspect, or executes the training method of the autonomous driving model described in the second aspect.

[0036] In a fifth aspect, an embodiment of the present application further provides a computer-readable storage medium, which stores computer instructions. When the computer instructions are executed on an electronic device, the electronic device implements the training method of the autonomous driving model described in the first aspect, or executes the training method of the autonomous driving model described in the second aspect.

[0037] It can be understood that the training system, electronic device and computer-readable storage medium of the autonomous driving model provided above all correspond to the training method of the autonomous driving model mentioned above. Therefore, the beneficial effects that can be achieved can refer to the beneficial effects in the corresponding method provided above, and will not be repeated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 A schematic diagram of a scenario of a training system for an autonomous driving model provided in one embodiment of the present application.

[0039] Figure 2 A flowchart of the steps of a method for training an autonomous driving model applied to a training system provided in one embodiment of the present application.

[0040] Figure 3 A flowchart of the steps of a method for training an autonomous driving model applied to a computer device provided in one embodiment of the present application.

[0041] Figure 4 A flowchart of the steps of a method for training an autonomous driving model applied to a computer device provided in another embodiment of the present application.

[0042] Figure 5 A flowchart of the steps of a method for training an autonomous driving model applied to an on-board device provided in one embodiment of the present application.

[0043] Figure 6 A schematic diagram of the structure of an electronic device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0044] In order to more clearly understand the above-mentioned purposes, features and advantages of the present application, the present application is described in detail below in conjunction with the accompanying drawings and specific implementation methods. It should be noted that the implementation methods of the present application and the features in the implementation methods can be combined with each other without conflict.

[0045] In the following description, many specific details are set forth to facilitate a full understanding of the present application. The described implementations are only part of the implementations of the present application, rather than all the implementations.

[0046] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which this application belongs. The terms used herein in the specification of this application are only for the purpose of describing specific embodiments and are not intended to limit this application.

[0047] It should be further noted that, in this article, the terms "comprises", "includes" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the presence of other identical elements in the process, method, article or device including the element.

[0048] In this application, "at least one" means one or more, and "more than one" means two or more than two. "And / or" describes the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone, where A and B can be singular or plural. The terms "first", "second", "third", "fourth", etc. (if any) in the specification, claims and drawings of this application are used to distinguish similar objects, rather than to describe a specific order or sequence.

[0049] In the embodiments of the present application, words such as "exemplary" or "for example" are used to indicate examples, illustrations or descriptions. Any embodiment or design described as "exemplary" or "for example" in the embodiments of the present application should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of words such as "exemplary" or "for example" is intended to present related concepts in a specific way.

[0050] During the training process of the autonomous driving model, the vehicle will collect massive amounts of data and upload it to the cloud. The cloud will store the data uploaded by the vehicle and then use the data to train the autonomous driving model.

[0051] However, the amount of data that vehicles need to collect is very large. Therefore, the cloud needs to consume a lot of bandwidth, memory and other software and hardware resources to store massive amounts of data. If a complete autonomous driving model is to be trained, the cost of data collection is very high.

[0052] Moreover, the data generated by vehicles during driving are similar, but there is less data that can be used to optimize the autonomous driving model. Therefore, if the autonomous driving model is trained directly based on the uploaded data, the training efficiency of the autonomous driving model will be low.

[0053] After collecting massive amounts of data, the cloud can also clean and filter the data. However, this process also requires a lot of time and software and hardware resources on the cloud.

[0054] In view of the above, the embodiments of the present application provide a training method, system, electronic device and computer-readable storage medium for an autonomous driving model.

[0055] First, refer to Figure 1 As shown, Figure 1 A schematic diagram of the structure of the training system for the autonomous driving model provided in an embodiment of the present application.

[0056] The autonomous driving model may include a computer device 100 and a vehicle-mounted device 200. The vehicle-mounted device 200 is configured in a vehicle 300.

[0057] The computer device 100 may be an independent server, or a server network or server cluster composed of servers. For example, the computer device 100 described in the embodiment of the present application includes but is not limited to a computer, a network host, a single network server, a plurality of network server sets or a cloud server composed of a plurality of servers. The cloud server is composed of a large number of computers or network servers based on cloud computing.

[0058] The vehicle-mounted device 200 is configured in the vehicle 300. The vehicle-mounted device 200 is a device that can automatically perform numerical calculations and / or information processing according to pre-set or stored instructions, and its hardware includes but is not limited to a processor, a microprogrammed control unit (MCU), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), a digital signal processor (DSP), an embedded device, etc. For example, the vehicle-mounted device 200 can be a central gateway in the vehicle, etc.

[0059] The vehicle 300 may also be equipped with other vehicle-mounted devices such as a vehicle speed sensor, a steering angle sensor, a camera, an acceleration sensor (G sensor), a GPS, etc., but is not limited thereto.

[0060] The computer device 100 is used to obtain the current sample collection strategy of the autonomous driving model.

[0061] Among them, the sample collection strategy may include driving scenarios of target driving data to be collected.

[0062] The sample collection strategy may also include the quantity ratio of the target driving data to be collected in each driving scene, which is not limited in the embodiments of the present application.

[0063] The driving scenario is used to describe various types of scenarios that the vehicle 300 may face during driving, such as the road conditions of the driving environment of the vehicle 300, the weather conditions during the driving of the vehicle 300, the brightness of the light in the driving environment of the vehicle 300, etc.

[0064] The road conditions of vehicle 300 may include: uphill, downhill, driving at a fork in the road, driving on a winding mountain road, obstacles on the road surface, and sudden road phenomena, such as a car accident near the driving path, a landslide ahead, acceleration of the vehicle ahead, overtaking of the vehicle behind, etc., but are not limited to these.

[0065] The weather conditions of the vehicle 300 may include driving at night, driving in fog and haze, driving in rainy days, driving in snowy days, etc., but are not limited thereto.

[0066] The driving data is used to describe the driving state of the vehicle 300. For example, the driving scene of the vehicle 300, the speed, direction, acceleration of the vehicle 300, the location of the vehicle 300, etc., but not limited thereto.

[0067] The target driving data is driving data generated by the vehicle 300 in the driving scenario included in the sample collection strategy.

[0068] In some embodiments, the computer device 100 is further used to obtain the vehicle control effect of the autonomous driving model in the driving scene to be tested, for example, to obtain a test sample in the driving scene to be tested, input the test sample into the autonomous driving model, and obtain a vehicle control instruction corresponding to the driving scene to be tested. The vehicle control instruction can be used to control the vehicle 300 to turn, accelerate, etc., but is not limited to this. Then, it is detected whether the execution of the vehicle control instruction in the driving scene to be tested meets the preset safety driving standards to obtain the vehicle control effect; if the vehicle control effect of the autonomous driving model in the driving scene to be tested does not meet the preset safety driving standards, the driving scene to be tested is used as the driving scene of the target driving data.

[0069] The driving scene to be tested may be set according to the environmental requirements for the automatic driving of the vehicle 300. For example, the vehicle 300 needs to drive automatically at night, and the driving scene to be tested may include night, the vehicle 300 needs to pass through an intersection, and the driving scene to be tested may include an intersection, etc., but is not limited thereto.

[0070] The safe driving standard can be set according to actual application requirements. For example, the safe driving standard can be set to the international standard New Car Assessment Program (NCAP).

[0071] The computer device 100 is also used to obtain the current sample collection strategy of the autonomous driving model and transmit the sample collection strategy to the vehicle-mounted device 200.

[0072] The vehicle-mounted device 200 is used to collect driving data of the vehicle 300 during driving, for example, using various sensors such as vehicle speed sensors, steering angle sensors, cameras, G sensors, GPS, etc. to obtain driving data, and receiving the current sample collection strategy of the autonomous driving model from the computer device 100, and then, based on the sample collection strategy, filtering out target driving data from the collected driving data, for example, obtaining the driving scene of the collected driving data; based on the driving scene of the collected driving data, filtering out driving data that meets the sample collection strategy from the collected driving data to obtain the target driving data.

[0073] The vehicle-mounted device 200 may be used to input the collected driving data into a pre-trained scene recognition model to obtain a driving scene of the collected driving data.

[0074] The scene recognition model can be set according to needs. For example, the scene recognition model can use ChatGPT, which can determine the driving scene of a time period based on the data of this time period.

[0075] For example, the driving data is a fragment of an image. The image of the fragment is input into ChatGPT, and ChatGPT outputs text to describe the fragment as "an intersection, the light turns from red to green, and the vehicle 300 in front starts to accelerate", so that the driving scene of the driving data can be accurately determined.

[0076] After screening out the target driving data, the in-vehicle device 200 may transmit the target driving data to the computer device 100 .

[0077] The computer device 100 is also used to add the target driving data to the training set of the autonomous driving model after receiving the target driving data; and train the autonomous driving model based on the training set of the autonomous driving model.

[0078] In some embodiments, after training the autonomous driving model based on the training set of the autonomous driving model, the computer device 100 may continue to obtain the vehicle control effect of the current autonomous driving model in each driving scene to be tested. If the vehicle control effect of the current autonomous driving model in the driving scene to be tested does not meet the preset safe driving standard, the driving scene to be tested may be used as the driving scene of the target driving data to generate a new sample collection strategy. Then, the new sample collection strategy is transmitted to the vehicle-mounted device 200. After receiving the target driving data corresponding to the new sample collection strategy from the vehicle-mounted device 200, training based on the target driving data continues until the vehicle control effect of the autonomous driving model in each driving scene to be tested meets the preset safe driving standard, thereby enabling the autonomous driving model to meet the safe driving standard when driving in all driving scenes to be tested, thereby achieving training optimization of the autonomous driving model.

[0079] The training system of the above-mentioned autonomous driving model can execute the training method of the autonomous driving model, refer to Figure 2 As shown, the training method applied to the training system may include:

[0080] Step 201: The computer device obtains the current sample collection strategy of the autonomous driving model.

[0081] In some embodiments, step 201 may include:

[0082] In step 2011, the computer device detects the vehicle control effect of the automatic driving model in the driving scenario to be tested.

[0083] The vehicle control effect is used to describe whether the vehicle controlled by the autonomous driving model meets the preset safe driving standards when driving in the driving scenario to be tested.

[0084] In some embodiments, step 2011 may include: the computer device 100 may obtain a test sample in the driving scene to be tested; input the test sample into the autonomous driving model to obtain a vehicle control instruction corresponding to the driving scene to be tested; and detect whether the execution of the vehicle control instruction in the driving scene to be tested complies with a preset safe driving standard to obtain the vehicle control effect.

[0085] For example, assuming that the driving scenarios to be tested include: "Vehicle ahead overtaking and cutting in", "Rainy day" and "Night", the computer device 100 can use video clips labeled with "Vehicle ahead overtaking and cutting in", "Rainy day" and "Night" as test samples, input these video clips into the current autonomous driving model, obtain vehicle control instructions, and compare the vehicle control instructions with safe driving standards (such as NCAP) to obtain the vehicle control effect. For example, the vehicle control effect can be that the vehicle control instructions of the autonomous driving model do not meet the safe driving standards in "Rainy day" and "Night".

[0086] If the vehicle control effect of the autonomous driving model in the driving scenario to be tested meets the preset safe driving standards, step 2012 can be executed, that is, the computer device 100 confirms that the current training of the autonomous driving model is completed.

[0087] If the vehicle control effect of the automatic driving model in the driving scenario to be tested does not meet the preset safe driving standards, execute step 2013.

[0088] In step 2013, the computer device uses the driving scenario to be tested as the driving scenario of the target driving data to obtain a sample collection strategy.

[0089] For example, if the vehicle control instructions of the automatic driving model do not meet the safe driving standards on "rainy days" and "nights", "rainy days" and "nights" can be used as driving scenarios for target driving data, that is, the sample collection strategy can be used to describe the driving data of the vehicle 300 on "rainy days" and "nights".

[0090] In some embodiments, the sample collection strategy may further include a ratio of the quantity of target driving data to be collected in each driving scenario.

[0091] For example, the quantity ratio of the target driving data of each driving scene can be determined according to the vehicle control effect of each driving scene to be tested. The worse the vehicle control effect in a driving scene to be tested, the higher the quantity ratio of the target driving data of the driving scene.

[0092] Step 202: The computer device transmits the sample collection strategy to the vehicle-mounted device.

[0093] Step 203: The vehicle-mounted device receives the current sample collection strategy of the autonomous driving model from the computer device.

[0094] Step 204: The vehicle-mounted equipment collects driving data of the vehicle during driving.

[0095] For example, driving data can be acquired using various sensors such as vehicle speed sensor, steering angle sensor, camera, G sensor, GPS, etc.

[0096] In step 205 , the vehicle-mounted device selects target driving data from the collected driving data based on the sample collection strategy.

[0097] In some embodiments, step 205 may include: obtaining the driving scene of the collected driving data, for example, inputting the collected driving data into a pre-trained scene recognition model to obtain the driving scene of the collected driving data, and then, based on the driving scene of the collected driving data, screening the driving data that meets the sample collection strategy from the collected driving data to obtain the target driving data.

[0098] For example, the scene recognition model can identify the driving scene that generates the driving data and label it. For example, if the driving data is generated in the driving scene of "night", the driving data can be labeled "night". The computer device 100 can determine whether the label of each driving data meets the driving scene included in the sample collection strategy. If it meets the driving scene included in the sample collection strategy, the driving data is used as the target driving data; if it does not meet the driving scene included in the sample collection strategy, the driving data can be deleted.

[0099] In step 206 , the vehicle-mounted device transmits the target driving data to the computer device 100 .

[0100] During the target driving data transmission process, the vehicle-mounted device 200 can manage the upload progress of the target driving data to avoid data loss or slow transmission progress, and can also handle problems such as data congestion or deadlock to ensure data integrity.

[0101] In step 207 , the computer device receives the target driving data.

[0102] In step 208 , the computer device adds the target driving data to the training set of the autonomous driving model.

[0103] For example, newly collected target driving data generated at "night" and "rainy days" are added to the training set.

[0104] Step 209: The computer device trains the autonomous driving model based on the training set of the autonomous driving model.

[0105] The computer device 100 may fine-tune the autonomous driving model and then enter step 2011 again to verify the vehicle control effect of the fine-tuned autonomous driving model in each driving scenario to be tested.

[0106] If the vehicle controlled by the autonomous driving model can meet the safe driving standards in each driving scenario to be tested, it can be confirmed that the current training of the autonomous driving model is completed.

[0107] If the vehicle controlled by the autonomous driving model does not meet the safe driving standards in a certain driving scenario to be tested, the sample collection strategy can be re-determined and training can be performed again, that is, executing step 2013 and steps 202 to 209.

[0108] Compared with the vehicle 300 uploading all data to the computer device 100, the vehicle 300 in the embodiment of the present application uploads the driving data that meets the sample collection strategy to the computer device 100, which can not only reduce the software and hardware resources occupied by the driving data transmission, save the software and hardware resources used by the computer device 100 to store driving data, reduce the workload of data cleaning and screening of the computer device 100, but also improve the training efficiency of the autonomous driving model.

[0109] The present application also provides a method for training an autonomous driving model. The method for training an autonomous driving model of the present embodiment can be applied to Figure 1 Computer device 100 is shown.

[0110] refer to Figure 3 As shown, Figure 3 This is a flowchart of the steps of the training method of the autonomous driving model. According to different requirements, the order of the steps in the flowchart can be changed, and some steps can be omitted.

[0111] See also Figure 3 As shown, the training method of the autonomous driving model may include the following steps.

[0112] Step 301, obtaining the current sample collection strategy of the autonomous driving model.

[0113] In some embodiments, reference Figure 4 As shown, step 301 may include:

[0114] Step 3011, detecting the vehicle control effect of the autonomous driving model in the driving scenario to be tested.

[0115] The vehicle control effect is used to describe whether the vehicle controlled by the autonomous driving model meets the preset safe driving standards when driving in the driving scenario to be tested.

[0116] In some embodiments, step 3011 may include: obtaining a test sample in the driving scene to be tested; inputting the test sample into the autonomous driving model to obtain a vehicle control instruction corresponding to the driving scene to be tested; and detecting whether executing the vehicle control instruction in the driving scene to be tested complies with a preset safe driving standard to obtain the vehicle control effect.

[0117] If the vehicle control effect of the autonomous driving model in the driving scenario to be tested meets the preset safe driving standards, execute step 3012, that is, confirm that the current training of the autonomous driving model is completed.

[0118] If the vehicle control effect of the automatic driving model in the driving scenario to be tested does not meet the preset safe driving standards, execute step 3013.

[0119] Step 3013: Use the driving scenario to be tested as the driving scenario of the target driving data to obtain a sample collection strategy.

[0120] Step 302: Transmit the sample collection strategy to the vehicle-mounted device.

[0121] The vehicle-mounted device 200 is used to filter the target driving data based on the sample collection strategy, and transmit the target driving data to the computer device 100 .

[0122] Step 303: Receive target driving data.

[0123] Step 304: Add the target driving data to the training set of the autonomous driving model.

[0124] Step 305: Train the autonomous driving model based on the training set of the autonomous driving model.

[0125] After executing step 305, step 3011 may be entered again until the vehicle control effect of the automatic driving model in each driving scenario to be tested meets the preset safe driving standards.

[0126] The present application also provides a method for training an autonomous driving model. The method for training an autonomous driving model of the present embodiment can be applied to Figure 1 The vehicle-mounted device 200 is shown.

[0127] refer to Figure 5 As shown, Figure 5 This is a flowchart of the steps of the training method of the autonomous driving model. According to different requirements, the order of the steps in the flowchart can be changed, and some steps can be omitted.

[0128] Step 501, collecting driving data of the vehicle during driving;

[0129] Step 502, receiving the current sample collection strategy of the autonomous driving model from the computer device.

[0130] Step 503: based on the sample collection strategy, select the target driving data from the collected driving data.

[0131] In some embodiments, step 503 may include: the vehicle-mounted device 200 obtains the driving scene of the collected driving data; based on the driving scene of the collected driving data, the driving data that meets the sample collection strategy is screened from the collected driving data to obtain the target driving data.

[0132] Wherein, obtaining the driving scene of the collected driving data includes: inputting the collected driving data into a pre-trained scene recognition model to obtain the driving scene of the collected driving data.

[0133] Step 504, transmitting the target driving data to a computer device.

[0134] The computer device 100 is used to add the target driving data to the training set of the autonomous driving model; and to train the autonomous driving model based on the training set of the autonomous driving model.

[0135] The embodiments of the present application can reduce the software and hardware resources occupied by driving data transmission, save the software and hardware resources used by the computer device 100 to store driving data, reduce the workload of data cleaning and screening of the computer device 100, and improve the training efficiency of the autonomous driving model.

[0136] Figure 6 This is a schematic diagram of an embodiment of an electronic device of the present application.

[0137] The electronic device 600 includes a memory 20, a processor 30, and a computer program 40 stored in the memory 20 and executable on the processor 30. When the processor 30 executes the computer program 40, the steps in the training method embodiment of the above-mentioned autonomous driving model are implemented, for example Figure 3 Steps 301 to 305 shown, or Figure 5 Steps 501 to 504 are shown.

[0138] Exemplarily, the computer program 40 may also be divided into one or more modules / units, which are stored in the memory 20 and executed by the processor 30. The one or more modules / units may be a series of computer program instruction segments capable of completing specific functions, and the instruction segments are used to describe the execution process of the computer program 40 in the electronic device 600.

[0139] Those skilled in the art will appreciate that the schematic diagram is merely an example of the electronic device 600 and does not constitute a limitation on the electronic device 600. The electronic device 600 may include more or fewer components than shown in the diagram, or a combination of certain components, or different components. For example, the electronic device 600 may also include input and output devices, network access devices, buses, etc.

[0140] The processor 30 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor, a single-chip microcomputer, or the processor 30 may also be any conventional processor, etc.

[0141] The memory 20 can be used to store the computer program 40 and / or the module / unit. The processor 30 realizes various functions of the electronic device 600 by running or executing the computer program and / or the module / unit stored in the memory 20 and calling the data stored in the memory 20. The memory 20 can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, an application required for at least one function (such as a sound playback function, an image playback function, etc.), etc.; the data storage area can store data (such as audio data) created according to the use of the electronic device 600, etc. In addition, the memory 20 can include a high-speed random access memory, and can also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), at least one disk storage device, a flash memory device, or other non-volatile solid-state storage devices.

[0142] If the module / unit integrated in the electronic device 600 is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present application implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium, and the computer program can implement the steps of the above-mentioned various method embodiments when executed by the processor. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electric carrier signal, telecommunication signal and software distribution medium. It should be noted that the content contained in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electric carrier signals and telecommunication signals.

[0143] In the several embodiments provided in this application, it should be understood that the disclosed electronic device and method can be implemented in other ways. For example, the electronic device embodiments described above are only schematic, for example, the division of the units is only a logical function division, and there may be other division methods in actual implementation.

[0144] In addition, each functional unit in each embodiment of the present application may be integrated into the same processing unit, each unit may exist physically separately, or two or more units may be integrated into the same unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of hardware plus software functional modules.

[0145] It is obvious to those skilled in the art that the present application is not limited to the details of the above exemplary embodiments, and that the present application can be implemented in other specific forms without departing from the spirit or basic features of the present application. Therefore, no matter from which point of view, the embodiments should be regarded as exemplary and non-restrictive. In addition, it is obvious that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. Multiple units or electronic devices stated in the electronic device claim can also be implemented by the same unit or electronic device through software or hardware. The words first, second, etc. are used to indicate names, and do not indicate any particular order.

[0146] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present application and are not intended to limit it. Although the present application has been described in detail with reference to the above embodiments, a person of ordinary skill in the art should understand that the technical solution of the present application may be modified or replaced by equivalents without departing from the spirit and scope of the technical solution of the present application.

Claims

1. A training method for an autonomous driving model, It is characterized in that Applied to a computer device, the computer device is communicatively connected to a vehicle-mounted device, the vehicle-mounted device is configured in a vehicle, and the training method of the autonomous driving model includes: Obtaining the current sample collection strategy of the autonomous driving model; The sample collection strategy is transmitted to the vehicle-mounted device; the vehicle-mounted device is used to filter the target driving data based on the sample collection strategy, and transmit the target driving data to the computer device; receiving the target driving data; Adding the target driving data to a training set of the autonomous driving model; And training the autonomous driving model based on the training set of the autonomous driving model.

2. The method for training an autonomous driving model according to claim 1, It is characterized in that The sample collection strategy includes the driving scenario of the target driving data, and the obtaining of the current sample collection strategy of the autonomous driving model includes: Obtaining a vehicle control effect of the autonomous driving model in a driving scenario to be tested; If the vehicle control effect of the automatic driving model in the driving scene to be tested does not meet the preset safe driving standards, the driving scene to be tested is used as the driving scene of the target driving data.

3. The method for training an autonomous driving model according to claim 2, It is characterized in that The obtaining of the vehicle control effect of the automatic driving model in the driving scenario to be tested includes: Acquire a test sample in a driving scenario to be tested; Inputting the test sample into the autonomous driving model to obtain a vehicle control instruction corresponding to the driving scenario to be tested; Detect whether the execution of the vehicle control instruction in the driving scenario to be tested meets the preset safe driving standard to obtain the vehicle control effect.

4. The method for training an autonomous driving model according to claim 2, It is characterized in that After the autonomous driving model is trained based on the training set of the autonomous driving model, the method further includes: Continue to enter the step of obtaining the vehicle control effect of the autonomous driving model in the driving scenario to be tested until the vehicle control effect of the autonomous driving model in the driving scenario to be tested meets the preset safe driving standard.

5. A training method for an autonomous driving model, It is characterized in that Applied to a vehicle-mounted device, the vehicle-mounted device is configured in a vehicle, the vehicle-mounted device is communicatively connected with a computer device, and the training method of the autonomous driving model includes: Collecting driving data of the vehicle during driving; Receiving a current sample collection strategy of the autonomous driving model from the computer device; Based on the sample collection strategy, target driving data is screened out from the collected driving data; The target driving data is transmitted to the computer device, and the computer device is used to add the target driving data to the training set of the autonomous driving model; and train the autonomous driving model based on the training set of the autonomous driving model.

6. The method for training an autonomous driving model according to claim 5, It is characterized in that The sample collection strategy includes the driving scenario of the target driving data, and the target driving data is screened out from the collected driving data based on the sample collection strategy, including: Acquire a driving scene of the collected driving data; Based on the driving scenario of the collected driving data, driving data satisfying the sample collection strategy is screened from the collected driving data to obtain the target driving data.

7. The training method of the autonomous driving model according to claim 6, It is characterized in that The acquiring of the driving scenario of the collected driving data includes: The collected driving data is input into a pre-trained scene recognition model to obtain a driving scene of the collected driving data.

8. A training system for an autonomous driving model, It is characterized in that include: A computer device, wherein the computer device is used to execute the training method of the autonomous driving model according to any one of claims 1 to 4; At least one vehicle, wherein the vehicle is equipped with an on-board device, wherein the on-board device is communicatively connected to the computer device, and wherein the on-board device is used to execute the training method of the autonomous driving model as described in any one of claims 5 to 7.

9. An electronic device, comprising a processor and a memory, It is characterized in that The memory is used to store instructions, and the processor is used to call the instructions in the memory, so that the electronic device executes the training method of the autonomous driving model as described in any one of claims 1 to 4, or executes the training method of the autonomous driving model as described in any one of claims 5 to 7.

10. A computer-readable storage medium, It is characterized in that The computer-readable storage medium stores computer instructions, and when the computer instructions are executed on an electronic device, the electronic device executes the training method for the autonomous driving model as described in any one of claims 1 to 4, or executes the training method for the autonomous driving model as described in any one of claims 5 to 7.

Images