CN112052959A - Automatic driving training method, equipment and medium based on federal learning - Google Patents
Automatic driving training method, equipment and medium based on federal learning Download PDFInfo
- Publication number
- CN112052959A CN112052959A CN202010931767.1A CN202010931767A CN112052959A CN 112052959 A CN112052959 A CN 112052959A CN 202010931767 A CN202010931767 A CN 202010931767A CN 112052959 A CN112052959 A CN 112052959A
- Authority
- CN
- China
- Prior art keywords
- preset
- model
- target
- training
- trained
- 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.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Software Systems (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Medical Informatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Artificial Intelligence (AREA)
- Traffic Control Systems (AREA)
Abstract
The application discloses an automatic driving training method, equipment and a medium based on federal learning, which are applied to a cloud server, wherein the method comprises the following steps: when a training instruction is detected, acquiring a preset model to be trained, and sending the preset model to be trained to each target vehicle within a preset driving range; acquiring first target model parameters obtained after each target vehicle trains the preset model to be trained on the basis of respective real-time local data sets; executing a preset federal flow based on each first target model parameter, and performing iterative training on the preset model to be trained to obtain a second target model parameter; and sending the second target model parameters to the target vehicles so that each target vehicle can obtain a respective target model based on the second target model parameters to realize a respective automatic driving function. The method and the device solve the technical problem that an effective automatic driving model is difficult to rapidly and accurately obtain in the prior art.
Description
Technical Field
The application relates to the technical field of artificial intelligence of internet science and technology, in particular to an automatic driving training method, equipment and medium based on federal learning.
Background
With the continuous development of financial science and technology, especially internet science and technology, more and more technologies (such as distributed, Blockchain, artificial intelligence and the like) are applied to the internet science and technology field, but the internet science and technology field also puts forward higher requirements on the technologies, and for example, the internet science and technology field also has higher requirements on federal learning-based automatic driving training.
Recently, the automatic driving technology is developed very rapidly, and the application of automatic driving has the advantages of reducing driving working intensity, relieving driving fatigue, improving driving safety, reducing accident rate and the like, wherein environment perception is one of core technologies of automatic driving, and automatic driving carries out analysis and decision through the result of environment perception as the basis of path planning, thereby realizing automatic driving, and the environment perception comprises lane detection, pedestrian detection, traffic sign recognition, obstacle detection and the like.
At present, environment perception is achieved through a model of an identification target obtained through deep learning training, but in the prior art, the model of the identification target is obtained through training in an offline mode through deep learning based on offline data, and then practical application of each environment scene is performed, the model obtained through offline training often has locality, namely the problem that the model obtained through offline training is not matched with an actual environment, and offline training data are often huge, so that data transmission load in a training process is too much, the model training efficiency is low, and the technical problem that an effective automatic driving model is difficult to obtain quickly and accurately exists in the prior art.
Disclosure of Invention
The application mainly aims to provide an automatic driving training method, equipment and medium based on federal learning, and aims to solve the technical problem that an effective automatic driving model is difficult to obtain quickly and accurately in the prior art.
In order to achieve the above object, the present application provides an automatic driving training method based on federal learning, which is applied to a cloud server, and the automatic driving training method based on federal learning includes:
when a training instruction is detected, acquiring a preset model to be trained, and sending the preset model to be trained to each target vehicle within a preset driving range;
acquiring first target model parameters obtained after each target vehicle trains the preset model to be trained on the basis of respective real-time local data sets;
executing a preset federal flow based on each first target model parameter, and performing iterative training on the preset model to be trained to obtain a second target model parameter;
and sending the second target model parameters to the target vehicles so that each target vehicle can obtain a respective target model based on the second target model parameters to realize a respective automatic driving function.
Optionally, the step of obtaining a preset model to be trained when the training instruction is detected, and sending the preset model to be trained to each target vehicle within a preset driving range includes:
when a training instruction is detected, selecting the preset model to be trained from a preset model set according to the region attribute carried in the training instruction;
and sending the preset model to be trained to each target vehicle within a preset driving range.
Optionally, when the training instruction is detected, the step of selecting the preset model to be trained from a preset model set according to the region attribute carried in the training instruction includes:
when a training instruction is detected, selecting a preset model subset from a preset model set according to the region attribute carried in the training instruction;
and selecting the preset model to be trained from a preset model subset according to the time information and the vehicle type information carried in the training instruction.
Optionally, the step of sending the second target model parameter to the target vehicles, so that each target vehicle obtains a respective target model based on the second target model parameter to implement a respective automatic driving function includes:
encrypting the second target model parameter to obtain an encrypted model parameter;
and sending the encrypted model parameters to the target vehicles so that each target vehicle can obtain a respective target model based on the second target model parameters to realize a respective automatic driving function.
Optionally, the step of obtaining a preset model to be trained and sending the preset model to be trained to each target vehicle within a preset driving range when the training instruction is detected further includes:
when a training instruction is detected, acquiring a preset model to be trained;
and communicating with each base station in a preset driving range, and distributing the preset model to be trained to each base station so that each base station can send the preset model to be trained to each target vehicle in the preset driving range based on the distance information between each base station and each target vehicle.
Optionally, before the step of acquiring a preset model to be trained and sending the preset model to be trained to each target vehicle within a preset driving range when the training instruction is detected, the method includes:
acquiring a preset basic model;
performing iterative training on the preset basic model based on offline training data with preset identification labels to train and update model parameters in the preset basic model until the preset basic model reaches a preset training completion condition;
and setting the preset basic model after reaching the preset training completion condition as a preset model to be trained.
Optionally, the offline training data with the preset identification tag includes offline laser radar scanning data, offline camera data, and offline millimeter wave radar data with the preset identification tag.
The application also provides an automatic driving training method based on federal learning, which is applied to vehicles and comprises the following steps:
receiving a preset model to be trained sent by a cloud server;
training the preset model to be trained based on a real-time local data set to obtain a first target model parameter;
encrypting and sending the first target model parameter to the cloud server so that the cloud server can execute a preset federal flow, and performing iterative training on the preset model to be trained to obtain a second target model parameter;
and receiving the second target model parameter sent by the cloud server in an encrypted manner, and obtaining a target model based on the second target model parameter so as to realize an automatic driving function.
The application still provides an autopilot trainer based on federal study, is applied to high in the clouds server, autopilot trainer based on federal study includes:
the system comprises a first acquisition module, a second acquisition module and a control module, wherein the first acquisition module is used for acquiring a preset model to be trained when a training instruction is detected, and sending the preset model to be trained to each target vehicle within a preset driving range;
the second acquisition module is used for acquiring each first target model parameter obtained after each target vehicle trains the preset model to be trained based on each real-time local data set;
the first training module is used for executing a preset federal flow based on each first target model parameter and carrying out iterative training on the preset model to be trained to obtain a second target model parameter;
and the sending module is used for sending the second target model parameters to the target vehicles so that each target vehicle can obtain a respective target model based on the second target model parameters to realize a respective automatic driving function.
Optionally, the first obtaining module includes:
the selection unit is used for selecting the preset model to be trained from a preset model set according to the area attribute carried in the training instruction when the training instruction is detected;
and the sending unit is used for sending the preset model to be trained to each target vehicle in a preset driving range.
Optionally, the selecting unit includes:
the first selection subunit is used for selecting a preset model subset from a preset model set according to the area attribute carried in the training instruction when the training instruction is detected;
and the second selection subunit is used for selecting the preset model to be trained from a preset model subset according to the time information and the vehicle type information carried in the training instruction.
Optionally, the sending module includes:
the encryption unit is used for carrying out encryption processing on the second target model parameter to obtain an encryption model parameter;
and the sending unit is used for sending the encrypted model parameters to the target vehicles so that each target vehicle can obtain a respective target model based on the second target model parameters to realize a respective automatic driving function.
Optionally, the first obtaining module further includes:
the detection unit is used for acquiring a preset model to be trained when a training instruction is detected;
and the distribution unit is used for communicating with each base station in a preset driving range and distributing the preset model to be trained to each base station so that each base station can send the preset model to be trained to each target vehicle in the preset driving range based on the distance information between each base station and each target vehicle.
Optionally, the federal learning based automatic driving training device further comprises:
the third acquisition module is used for acquiring a preset basic model;
the second training module is used for carrying out iterative training on the preset basic model based on off-line training data with preset identification labels so as to train and update model parameters in the preset basic model until the preset basic model reaches a preset training completion condition;
and the setting module is used for setting the preset basic model after the preset training completion condition is reached as a preset model to be trained.
Optionally, the offline training data with the preset identification tag includes offline laser radar scanning data, offline camera data, and offline millimeter wave radar data with the preset identification tag.
The application also provides an autopilot training device based on federal learning, is applied to the vehicle, autopilot training device based on federal learning includes:
the first receiving module is used for receiving a preset model to be trained sent by the cloud server;
the fourth acquisition module is used for training the preset model to be trained based on the real-time local data set to obtain a first target model parameter;
the fifth obtaining module is used for encrypting and sending the first target model parameter to the cloud server so that the cloud server can execute a preset federal flow, and performing iterative training on the preset model to be trained to obtain a second target model parameter;
and the second receiving module is used for receiving the second target model parameters sent by the cloud server in an encrypted manner, and obtaining a target model based on the second target model parameters so as to realize an automatic driving function.
The application also provides an autopilot training device based on federal learning, autopilot training device based on federal learning is entity equipment, autopilot training device based on federal learning includes: a memory, a processor, and a program of the federal learning based autopilot training method stored in the memory and operable on the processor, the program of the federal learning based autopilot training method being executable by the processor to perform the steps of the federal learning based autopilot training method as described above.
The present application also provides a medium having a program stored thereon for implementing the above-described federal-learning-based automatic driving training method, wherein the program for implementing the above-described federal-learning-based automatic driving training method implements the steps of the above-described federal-learning-based automatic driving training method when executed by a processor.
The application provides an automatic driving training method, equipment and medium based on federal learning, compared with the prior art that a model for identifying a target is obtained by training local off-line data and then the model is put into different practical environments for use, the method comprises the steps of obtaining a preset model to be trained when a training instruction is detected, sending the preset model to be trained to each target vehicle in a preset driving range, and obtaining first target model parameters obtained after each target vehicle trains the preset model to be trained on the basis of respective real-time local data sets; executing a preset federal flow based on each first target model parameter, and performing iterative training on the preset model to be trained to obtain a second target model parameter; and sending the second target model parameters to the target vehicles so that each target vehicle can obtain a respective target model based on the second target model parameters to realize a respective automatic driving function. In this application, after obtaining and predetermineeing the model of waiting to train, based on each target vehicle online real-time training federal model under the actual environment scene, avoid the problem that federal model and actual environment do not match or the federal model has local limitation, and because in this application, only carry out the interaction of model parameter, therefore, avoid the data transmission load too much in the federal model training process, lead to the problem that model training efficiency is low to realize obtaining effective autopilot model fast and accurately.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and together with the description, serve to explain the principles of the application.
In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings needed to be used in the description of the embodiments or the prior art will be briefly described below, and it is obvious for those skilled in the art to obtain other drawings without inventive exercise.
FIG. 1 is a schematic flow chart of a first embodiment of an automatic driving training method based on federal learning according to the present application;
fig. 2 is a detailed flowchart illustrating steps of acquiring a preset model to be trained and sending the preset model to be trained to each target vehicle within a preset driving range when a training instruction is detected in the first embodiment of the federal learning-based automatic driving training method;
fig. 3 is a schematic device structure diagram of a hardware operating environment according to an embodiment of the present application.
The objectives, features, and advantages of the present application will be further described with reference to the accompanying drawings.
Detailed Description
It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.
The embodiment of the application provides an automatic driving training method based on federal learning, and in a first embodiment of the automatic driving training method based on federal learning, referring to fig. 1, the automatic driving training method based on federal learning is applied to a cloud server, and comprises the following steps:
step S10, when a training instruction is detected, acquiring a preset model to be trained, and sending the preset model to be trained to each target vehicle within a preset driving range;
step S20, acquiring first target model parameters obtained after each target vehicle trains the preset model to be trained based on the respective real-time local data set;
step S30, executing a preset federal flow based on each first target model parameter, and performing iterative training on the preset model to be trained to obtain a second target model parameter;
step S40, sending the second target model parameter to the target vehicles, so that each target vehicle obtains its own target model based on the second target model parameter, so as to implement its own automatic driving function.
The method comprises the following specific steps:
step S10, when a training instruction is detected, acquiring a preset model to be trained, and sending the preset model to be trained to each target vehicle within a preset driving range;
in this embodiment, the automatic driving training method based on federal learning is applied to a cloud server, the cloud server is in communication connection with each participant through a base station, it should be noted that, in this embodiment, each participant is a target vehicle, wherein the target vehicle and the cloud server together form an automatic driving training system based on federal learning, the automatic driving training system based on federal learning belongs to an automatic driving training device based on federal learning, it should be noted that, in this embodiment, the automatic driving training system based on federal learning can be divided into automatic driving training subsystems based on federal learning in different areas, specifically, for example, the automatic driving training subsystem based on federal learning in area a and the automatic driving training subsystem based on federal learning in area B, wherein a training instruction can be triggered by a user through the automatic driving training system (or subsystem) based on federal learning, after an automatic driving training system based on federal learning triggers a training instruction, a cloud server can detect the training instruction, the cloud server obtains a preset model to be trained after detecting the training instruction, the preset model to be trained is sent to each target vehicle within a preset driving range, and it needs to be explained that a plurality of models exist in the cloud server, so that the preset model to be trained needs to be selected from the plurality of models, and the preset model to be trained is sent to each target vehicle within the preset driving range, wherein the mode of selecting the preset model to be trained from the plurality of models can be that target model attribute information is extracted from the training instruction, so that the preset model to be trained is selected from the plurality of models, or the preset model to be trained is selected from the plurality of models based on model information associated with the training instruction.
When a training instruction is detected, acquiring a preset model to be trained, and sending the preset model to be trained to each target vehicle within a preset driving range on line in real time, wherein the preset driving range can be determined by the preset model to be trained, namely the preset model to be trained, and the preset driving range is determined, and the sending of the preset model to be trained to each target vehicle within the preset driving range can refer to: the model to be trained is sent to all vehicles within a preset driving range, or the model to be trained is sent to a certain type of vehicles within the preset driving range, for example, the model to be trained is sent to a truck type vehicle within the preset driving range, the model to be trained is sent to a car type vehicle within the preset driving range, or the model to be trained is sent to a passenger type vehicle within the preset driving range. In this embodiment, it should be noted that the purpose of sending the preset model to be trained to a certain type of vehicle within the preset driving range is to: the accuracy of training is promoted because, the vehicle is different, and the environmental perception of vehicle camera or radar is different, therefore trains alone according to different vehicle types, and data pertinence is strong, therefore, can promote the training rate of accuracy.
It should be noted that, in this embodiment, when the preset model to be trained is sent to a certain type of vehicle within a preset driving range, the vehicle type needs to be identified when the vehicle enters the preset driving range, specifically, the type of the vehicle may be identified according to a roadside unit of a road system, and after the identification, the cloud server receives the identified vehicle type to finally determine the target vehicle.
Referring to fig. 2, the step of acquiring a preset model to be trained and sending the preset model to be trained to each target vehicle within a preset driving range when a training instruction is detected includes:
step S11, when a training instruction is detected, selecting the preset model to be trained from a preset model set according to the area attribute carried in the training instruction;
in this embodiment, when a training instruction is detected, the preset model to be trained is selected from a preset model set according to the area attribute carried in the training instruction, for example, there are 100 models in the preset model set, and the 100 models correspond to different cities, and then the preset model to be trained is selected from the preset model set according to the city information carried in the training instruction.
And step S12, sending the preset model to be trained to each target vehicle in a preset driving range.
After the preset model to be trained is obtained, the preset model to be trained is sent to each target vehicle within the preset driving range in real time on line, namely, as long as the vehicle enters the preset driving range, the preset model to be trained is sent to each target vehicle within the preset driving range, so that the target vehicles can carry out local training.
When a training instruction is detected, the step of selecting the preset model to be trained from a preset model set according to the region attribute carried in the training instruction comprises the following steps:
step S111, when a training instruction is detected, selecting a preset model subset from a preset model set according to the area attribute carried in the training instruction;
and step S112, selecting the preset model to be trained from a preset model subset according to the time information and the vehicle type information carried in the training instruction.
In this embodiment, after selecting the preset model subset from the preset model set, the preset model to be trained is selected from the preset model subset according to the time information and the vehicle type information carried in the training instruction, that is, in this embodiment, the selection of the preset model to be trained is simultaneously associated with the vehicle type and the time, because the light change rule is different and the shooting of the vehicle-mounted camera on the vehicle is changed in different time periods due to different sunshine and different sunshine, so that the preset model to be trained is selected according to the vehicle type and the time, specifically, assuming that there are 100 models in the preset model subset and there are 10 models corresponding to the summer and the passenger cars, the preset model to be trained is selected from the 10 models, in this embodiment, according to the time information and the vehicle type information carried in the training instruction, specifically, the preset model to be trained is selected from the preset model subset, so that the model training speed and accuracy are improved, and it is to be noted that the preset model to be trained is a preset model to be trained in spring, and in autumn, the corresponding model in spring cannot be updated to the environment in autumn when being applied.
Before the step of acquiring a preset model to be trained and sending the preset model to be trained to each target vehicle within a preset driving range when a training instruction is detected, the method comprises the following steps:
step S01, acquiring a preset basic model;
in this embodiment, it should be noted that although there is a certain possibility of a deviation of the model generated by the offline training, that is, the correlation between the sample of the offline data and the acquired place, environment, and the like is relatively large, and there is a case that a local sample is often present, for example, the data acquisition of the offline training is limited to a certain area and limited to a certain specific environment (illumination, weather, and the like), so as to affect the training result, in order to ensure the security of the initial training stage, in this embodiment, the target model obtained after the offline training of the offline data is used as the preset model to be trained in the initial training stage, that is, the preset model to be trained is obtained by offline training through a preset basic model.
Step S02, performing iterative training on the preset basic model based on the off-line training data with the preset identification label to perform training and updating on the model parameters in the preset basic model until the preset basic model reaches a preset training completion condition;
and performing iterative training on the preset basic model based on the offline training data with the preset identification tag to train and update the model parameters in the preset basic model until the preset basic model reaches a preset training completion condition, wherein the offline training data with the preset identification tag comprises offline laser radar scanning data, offline shooting data, offline millimeter wave radar data and the like with the preset identification tag. The method comprises the steps of carrying out iterative training on a preset basic model based on offline training data with preset identification labels, specifically, offline training data with preset identification labels in different areas (or different time and different vehicle types) to obtain different types of preset models to be trained, specifically, carrying out iterative federal training on the preset basic model based on the offline training data with the preset identification labels to train and update model parameters in the preset basic model until the preset basic model reaches preset training completion conditions, wherein the preset training completion conditions comprise that the training times reach a second preset number or a first preset loss function is converged.
And step S03, setting the preset basic model after reaching the preset training completion condition as a preset model to be trained.
And setting the preset basic model after reaching the preset training completion condition as a preset model to be trained so as to provide, wherein the number of the preset models to be trained is multiple.
Step S20, acquiring first target model parameters obtained after each target vehicle trains the preset model to be trained based on the respective real-time local data set;
in this embodiment, after determining the target vehicles, first target model parameters obtained after the preset model to be trained is trained by the target vehicles based on the respective real-time local data sets are obtained, specifically, first target model parameters obtained after the preset model to be trained is trained by the target vehicles based on the respective real-time local data sets with corresponding labels are obtained, where the target vehicles may obtain corresponding first target model parameters after being trained for a second preset number of times.
Step S30, executing a preset federal flow based on each first target model parameter, and performing iterative training on the preset model to be trained to obtain a second target model parameter;
executing a preset federal flow based on each first target model parameter, performing iterative training on the preset model to be trained to obtain a second target model parameter, specifically, each target vehicle sends the first target model parameter to a cloud server, the cloud server receives each first target model parameter, performing aggregation means such as averaging and the like on each first target model parameter to obtain a first aggregation parameter, after obtaining the first aggregation parameter, sending the first aggregation parameter to each target vehicle for each target vehicle to perform next round of local iterative training to obtain a target model parameter corresponding to a next round, after obtaining the target model parameter corresponding to the next round, each target vehicle sends the target model parameter of the next round to the cloud server for aggregation by the cloud server, and continuously performing iterative training based on the first aggregation parameter, and setting the parameters of the converged model as second target model parameters until the second preset loss function of the system converges.
Step S40, sending the second target model parameter to the target vehicles, so that each target vehicle obtains its own target model based on the second target model parameter, so as to implement its own automatic driving function.
In this embodiment, after obtaining the second target model parameter, the second target model parameter is sent to the target vehicles, and the target vehicles perfect the local target models based on the second target model parameter, that is, each target vehicle obtains its own target model based on the second target model parameter, and after obtaining the target models, the automatic driving function of each target vehicle is realized.
Compared with the prior method that different sample data are converted into an embedded vector by a participant and interactive federated modeling is carried out on the embedded vector and a server, the method obtains the sample data, wherein the sample data comprises user data and article data; determining, locally at the first party, a user-embedded vector of the user data, and receiving an item-embedded vector sent by a server, the item-embedded vector being generated locally at the server based on the item data; and obtaining a preset prediction model of the first participant through federal learning training based on the user embedding vector and the article embedding vector. In the application, after sample data is obtained, a user embedded vector of the user data is determined locally at the first participant, an article embedded vector determined based on the article data is received at a server side, and then a preset prediction model of the first participant is obtained through federal learning training based on the user embedded vector and the article embedded vector.
In another embodiment of the federal learning-based automatic driving training method, the step of sending the second target model parameter to the target vehicles so that each target vehicle obtains a respective target model based on the second target model parameter to realize a respective automatic driving function includes:
step A1, carrying out encryption processing on the second target model parameter to obtain an encryption model parameter;
in this embodiment, in order to ensure security, when the target vehicle interacts with the cloud server, the target vehicle interacts with the cloud server by encryption, and if the first target model parameter is encrypted and sent to the cloud server, the cloud server encrypts the second target model parameter to obtain the encrypted model parameter.
Step a2, sending the encrypted model parameters to the target vehicles, so that each target vehicle obtains its own target model based on the second target model parameters, so as to implement its own automatic driving function.
In this embodiment, the encrypted model parameters may be sent to the target vehicles through a block chain, so that each target vehicle obtains a respective target model based on the second target model parameters, so as to achieve a respective automatic driving function.
In this embodiment, the second target model parameter is encrypted to obtain an encrypted model parameter; and sending the encrypted model parameters to the target vehicles so that each target vehicle can obtain a respective target model based on the second target model parameters to realize a respective automatic driving function. In this embodiment, a secure target model is achieved.
In another embodiment of the federal learning-based automatic driving training method, when a training instruction is detected, a preset model to be trained is obtained, and the preset model to be trained is sent to each target vehicle within a preset driving range, where the method further includes:
step B1, when a training instruction is detected, acquiring a preset model to be trained;
and step B2, communicating with each base station in a preset driving range, and distributing the preset model to be trained to each base station so that each base station can send the preset model to be trained to each target vehicle in the preset driving range based on the distance information between each base station and each target vehicle.
In this embodiment, when the training instruction is detected, the cloud server obtains a preset model to be trained, communicates with each base station within a preset driving range, distributes the preset model to be trained to each base station, so that each base station can send the preset model to be trained to each target vehicle in the preset driving range based on the distance information between each base station and each target vehicle, namely, in the embodiment, in order to reduce the communication time length for each target vehicle to rapidly obtain the preset model to be trained, and finally, quickly obtaining the target model, and sending the preset model to be trained to each target vehicle within a preset driving range by each base station based on the distance information between each base station and each target vehicle, specifically, enabling each target vehicle to obtain the preset model to be trained within the shortest communication distance.
In the embodiment, when a training instruction is detected, a preset model to be trained is obtained; and communicating with each base station in a preset driving range, and distributing the preset model to be trained to each base station so that each base station can send the preset model to be trained to each target vehicle in the preset driving range based on the distance information between each base station and each target vehicle. In this embodiment, each target vehicle obtains a preset model to be trained based on the closest distance, and the model training efficiency is improved.
The embodiment of the application provides an automatic driving training method based on federal learning, and in another embodiment of the automatic driving training method based on federal learning, the automatic driving training method based on federal learning is applied to a vehicle, and comprises the following steps:
receiving a preset model to be trained sent by a cloud server;
training the preset model to be trained based on a real-time local data set to obtain a first target model parameter;
encrypting and sending the first target model parameter to the cloud server so that the cloud server can execute a preset federal flow, and performing iterative training on the preset model to be trained to obtain a second target model parameter;
and receiving the second target model parameter sent by the cloud server in an encrypted manner, and obtaining a target model based on the second target model parameter so as to realize an automatic driving function.
In this embodiment, for a vehicle, a preset model to be trained sent by a cloud server is received; training the preset model to be trained based on a real-time local data set to obtain a first target model parameter; encrypting and sending the first target model parameter to the cloud server so that the cloud server can execute a preset federal flow, and performing iterative training on the preset model to be trained to obtain a second target model parameter; and receiving the second target model parameter sent by the cloud server in an encrypted manner, and obtaining a target model based on the second target model parameter so as to realize an automatic driving function. In the embodiment, the target vehicle trains the target model on line in real time through the federal so as to realize automatic driving.
Referring to fig. 3, fig. 3 is a schematic device structure diagram of a hardware operating environment according to an embodiment of the present application.
As shown in fig. 3, the federal learning based automatic drive training device may include: a processor 1001, such as a CPU, a memory 1005, and a communication bus 1002. The communication bus 1002 is used for realizing connection communication between the processor 1001 and the memory 1005. The memory 1005 may be a high-speed RAM memory or a non-volatile memory (e.g., a magnetic disk memory). The memory 1005 may alternatively be a memory device separate from the processor 1001 described above.
Optionally, the federal learning based autopilot training device may further include a rectangular user interface, a network interface, a camera, RF (Radio Frequency) circuitry, sensors, audio circuitry, a WiFi module, and the like. The rectangular user interface may comprise a Display screen (Display), an input sub-module such as a Keyboard (Keyboard), and the optional rectangular user interface may also comprise a standard wired interface, a wireless interface. The network interface may optionally include a standard wired interface, a wireless interface (e.g., WI-FI interface).
Those skilled in the art will appreciate that the federal learning based autopilot training device architecture illustrated in fig. 3 does not constitute a limitation of the federal learning based autopilot training device and may include more or fewer components than shown, or some components in combination, or a different arrangement of components.
As shown in fig. 3, a memory 1005, which is one type of computer medium, may include an operating system, a network communication module, and a federal learning based automated driving training program. The operating system is a program that manages and controls the hardware and software resources of the federal learning based autopilot training facility, supporting the operation of the federal learning based autopilot training program as well as other software and/or programs. The network communication module is used to enable communication between components within the memory 1005, as well as with other hardware and software in the federal learning based automatic drive training system.
In the federal learning based automatic driving training device shown in fig. 3, the processor 1001 is configured to execute the federal learning based automatic driving training program stored in the memory 1005, and implement any of the steps of the federal learning based automatic driving training method described above.
The specific implementation of the automatic driving training equipment based on federal learning in the application is basically the same as that of the above automatic driving training method based on federal learning, and is not described herein again.
The application also provides an autopilot trainer based on federal study, and the application provides an autopilot trainer based on federal study, is applied to high in the clouds server, autopilot trainer based on federal study includes:
the system comprises a first acquisition module, a second acquisition module and a control module, wherein the first acquisition module is used for acquiring a preset model to be trained when a training instruction is detected, and sending the preset model to be trained to each target vehicle within a preset driving range;
the second acquisition module is used for acquiring each first target model parameter obtained after each target vehicle trains the preset model to be trained based on each real-time local data set;
the first training module is used for executing a preset federal flow based on each first target model parameter and carrying out iterative training on the preset model to be trained to obtain a second target model parameter;
and the sending module is used for sending the second target model parameters to the target vehicles so that each target vehicle can obtain a respective target model based on the second target model parameters to realize a respective automatic driving function.
Optionally, the first obtaining module includes:
the selection unit is used for selecting the preset model to be trained from a preset model set according to the area attribute carried in the training instruction when the training instruction is detected;
and the sending unit is used for sending the preset model to be trained to each target vehicle in a preset driving range.
Optionally, the selecting unit includes:
the first selection subunit is used for selecting a preset model subset from a preset model set according to the area attribute carried in the training instruction when the training instruction is detected;
and the second selection subunit is used for selecting the preset model to be trained from a preset model subset according to the time information and the vehicle type information carried in the training instruction.
Optionally, the sending module includes:
the encryption unit is used for carrying out encryption processing on the second target model parameter to obtain an encryption model parameter;
and the sending unit is used for sending the encrypted model parameters to the target vehicles so that each target vehicle can obtain a respective target model based on the second target model parameters to realize a respective automatic driving function.
Optionally, the first obtaining module further includes:
the detection unit is used for acquiring a preset model to be trained when a training instruction is detected;
and the distribution unit is used for communicating with each base station in a preset driving range and distributing the preset model to be trained to each base station so that each base station can send the preset model to be trained to each target vehicle in the preset driving range based on the distance information between each base station and each target vehicle.
Optionally, the federal learning based automatic driving training device further comprises:
the third acquisition module is used for acquiring a preset basic model;
the second training module is used for carrying out iterative training on the preset basic model based on off-line training data with preset identification labels so as to train and update model parameters in the preset basic model until the preset basic model reaches a preset training completion condition;
and the setting module is used for setting the preset basic model after the preset training completion condition is reached as a preset model to be trained.
Optionally, the offline training data with the preset identification tag includes offline laser radar scanning data, offline camera data, and offline millimeter wave radar data with the preset identification tag.
The application also provides an autopilot training device based on federal learning, is applied to the vehicle, autopilot training device based on federal learning includes:
the first receiving module is used for receiving a preset model to be trained sent by the cloud server;
the fourth acquisition module is used for training the preset model to be trained based on the real-time local data set to obtain a first target model parameter;
the fifth obtaining module is used for encrypting and sending the first target model parameter to the cloud server so that the cloud server can execute a preset federal flow, and performing iterative training on the preset model to be trained to obtain a second target model parameter;
and the second receiving module is used for receiving the second target model parameters sent by the cloud server in an encrypted manner, and obtaining a target model based on the second target model parameters so as to realize an automatic driving function.
The specific implementation of the automatic driving training device based on federal learning in the application is basically the same as that of the above automatic driving training method based on federal learning, and is not repeated herein.
The present embodiments provide a medium storing one or more programs, which may be further executed by one or more processors for implementing the steps of any of the above-mentioned federal learning based automatic driving training methods.
The specific implementation of the medium of the present application is substantially the same as that of each embodiment of the above automatic driving training method based on federal learning, and is not described herein again.
The above description is only a preferred embodiment of the present application, and not intended to limit the scope of the present application, and all modifications of equivalent structures and equivalent processes, which are made by the contents of the specification and the drawings, or which are directly or indirectly applied to other related technical fields, are included in the scope of the present application.
Claims (10)
1. The automatic driving training method based on the federal learning is characterized by being applied to a cloud server and comprises the following steps:
when a training instruction is detected, acquiring a preset model to be trained, and sending the preset model to be trained to each target vehicle within a preset driving range;
acquiring first target model parameters obtained after each target vehicle trains the preset model to be trained on the basis of respective real-time local data sets;
executing a preset federal flow based on each first target model parameter, and performing iterative training on the preset model to be trained to obtain a second target model parameter;
and sending the second target model parameters to the target vehicles so that each target vehicle can obtain a respective target model based on the second target model parameters to realize a respective automatic driving function.
2. The federal learning-based automatic driving training method as claimed in claim 1, wherein the step of obtaining a preset model to be trained and transmitting the preset model to be trained to each target vehicle within a preset driving range when a training instruction is detected comprises:
when a training instruction is detected, selecting the preset model to be trained from a preset model set according to the region attribute carried in the training instruction;
and sending the preset model to be trained to each target vehicle within a preset driving range.
3. The federal learning-based automatic driving training method as claimed in claim 2, wherein the step of selecting the preset model to be trained from a preset model set according to the region attribute carried in the training instruction when the training instruction is detected comprises:
when a training instruction is detected, selecting a preset model subset from a preset model set according to the region attribute carried in the training instruction;
and selecting the preset model to be trained from a preset model subset according to the time information and the vehicle type information carried in the training instruction.
4. The federal learning-based automated driving training method as claimed in claim 1, wherein said step of sending said second target model parameters to said target vehicles for each of said target vehicles to obtain a respective target model based on said second target model parameters to implement a respective automated driving function comprises:
encrypting the second target model parameter to obtain an encrypted model parameter;
and sending the encrypted model parameters to the target vehicles so that each target vehicle can obtain a respective target model based on the second target model parameters to realize a respective automatic driving function.
5. The federal learning-based automatic driving training method as claimed in claim 1, wherein the step of obtaining a preset model to be trained and sending the preset model to be trained to each target vehicle within a preset driving range when a training instruction is detected further comprises:
when a training instruction is detected, acquiring a preset model to be trained;
and communicating with each base station in a preset driving range, and distributing the preset model to be trained to each base station so that each base station can send the preset model to be trained to each target vehicle in the preset driving range based on the distance information between each base station and each target vehicle.
6. The federal learning-based automatic drive training method as claimed in claim 1, wherein before the step of obtaining a preset model to be trained and sending the preset model to be trained to each target vehicle within a preset driving range when a training instruction is detected, the method comprises:
acquiring a preset basic model;
performing iterative training on the preset basic model based on offline training data with preset identification labels to train and update model parameters in the preset basic model until the preset basic model reaches a preset training completion condition;
and setting the preset basic model after reaching the preset training completion condition as a preset model to be trained.
7. The federal learning-based automatic drive training method as claimed in claim 6, wherein the offline training data with preset identification tags includes offline lidar scanning data, offline camera data and offline millimeter wave radar data with preset identification tags.
8. An automatic driving training method based on federal learning, which is characterized by being applied to a vehicle, and comprises the following steps:
receiving a preset model to be trained sent by a cloud server;
training the preset model to be trained based on a real-time local data set to obtain a first target model parameter;
encrypting and sending the first target model parameter to the cloud server so that the cloud server can execute a preset federal flow, and performing iterative training on the preset model to be trained to obtain a second target model parameter;
and receiving the second target model parameter sent by the cloud server in an encrypted manner, and obtaining a target model based on the second target model parameter so as to realize an automatic driving function.
9. An federal learning-based autopilot training device, comprising: a memory, a processor, and a program stored on the memory for implementing the federal learning based automatic drive training method,
the memory is used for storing a program for realizing the automatic driving training method based on the federal learning;
the processor is configured to execute a program for implementing the federal learning based automated driving training method to implement the steps of the federal learning based automated driving training method as claimed in any one of claims 1 to 8.
10. A medium having a program for implementing the federal learning based autopilot training method stored thereon for execution by a processor to perform the steps of the federal learning based autopilot training method as claimed in any one of claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010931767.1A CN112052959B (en) | 2020-09-04 | 2020-09-04 | Automatic driving training method, equipment and medium based on federal learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010931767.1A CN112052959B (en) | 2020-09-04 | 2020-09-04 | Automatic driving training method, equipment and medium based on federal learning |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112052959A true CN112052959A (en) | 2020-12-08 |
| CN112052959B CN112052959B (en) | 2023-08-25 |
Family
ID=73609868
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010931767.1A Active CN112052959B (en) | 2020-09-04 | 2020-09-04 | Automatic driving training method, equipment and medium based on federal learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112052959B (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112486180A (en) * | 2020-12-10 | 2021-03-12 | 深圳前海微众银行股份有限公司 | Vehicle control method, device, equipment, storage medium and program product |
| CN112766138A (en) * | 2021-01-14 | 2021-05-07 | 深圳前海微众银行股份有限公司 | Positioning method, device and equipment based on image recognition and storage medium |
| CN113313264A (en) * | 2021-06-02 | 2021-08-27 | 河南大学 | Efficient federal learning method in Internet of vehicles scene |
| CN113920780A (en) * | 2021-09-01 | 2022-01-11 | 同济大学 | Cloud and mist collaborative personalized forward collision risk early warning method based on federal learning |
| CN114548608A (en) * | 2022-04-26 | 2022-05-27 | 腾讯科技（深圳）有限公司 | Model processing method and device, target traffic equipment and storage medium |
| CN114740970A (en) * | 2022-02-23 | 2022-07-12 | 广东工业大学 | Millimeter wave gesture recognition method and system based on federal learning |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106446754A (en) * | 2015-08-11 | 2017-02-22 | 阿里巴巴集团控股有限公司 | Image identification method, metric learning method, image source identification method and devices |
| CN109747659A (en) * | 2018-11-26 | 2019-05-14 | 北京汽车集团有限公司 | The control method and device of vehicle drive |
| CN110196593A (en) * | 2019-05-16 | 2019-09-03 | 济南浪潮高新科技投资发展有限公司 | A kind of more scene environments detections of automatic Pilot and decision system and method |
| CN110867096A (en) * | 2019-10-08 | 2020-03-06 | 江苏大学 | Mountain road safety control system and method suitable for vehicles with multiple automatic driving grades |
| US20200189590A1 (en) * | 2018-12-18 | 2020-06-18 | Beijing DIDI Infinity Technology and Development Co., Ltd | Systems and methods for determining driving action in autonomous driving |
| CN111428881A (en) * | 2020-03-20 | 2020-07-17 | 深圳前海微众银行股份有限公司 | Recognition model training method, device, equipment and readable storage medium |
-
2020
- 2020-09-04 CN CN202010931767.1A patent/CN112052959B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106446754A (en) * | 2015-08-11 | 2017-02-22 | 阿里巴巴集团控股有限公司 | Image identification method, metric learning method, image source identification method and devices |
| CN109747659A (en) * | 2018-11-26 | 2019-05-14 | 北京汽车集团有限公司 | The control method and device of vehicle drive |
| US20200189590A1 (en) * | 2018-12-18 | 2020-06-18 | Beijing DIDI Infinity Technology and Development Co., Ltd | Systems and methods for determining driving action in autonomous driving |
| CN110196593A (en) * | 2019-05-16 | 2019-09-03 | 济南浪潮高新科技投资发展有限公司 | A kind of more scene environments detections of automatic Pilot and decision system and method |
| CN110867096A (en) * | 2019-10-08 | 2020-03-06 | 江苏大学 | Mountain road safety control system and method suitable for vehicles with multiple automatic driving grades |
| CN111428881A (en) * | 2020-03-20 | 2020-07-17 | 深圳前海微众银行股份有限公司 | Recognition model training method, device, equipment and readable storage medium |
Non-Patent Citations (2)
| Title |
|---|
| 潘如晟等: "联邦学习可视化:挑战与框架", 《计算机辅助设计与图形学学报》 * |
| 潘如晟等: "联邦学习可视化:挑战与框架", 《计算机辅助设计与图形学学报》, no. 04, 30 April 2020 (2020-04-30), pages 513 - 519 * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112486180A (en) * | 2020-12-10 | 2021-03-12 | 深圳前海微众银行股份有限公司 | Vehicle control method, device, equipment, storage medium and program product |
| CN112766138A (en) * | 2021-01-14 | 2021-05-07 | 深圳前海微众银行股份有限公司 | Positioning method, device and equipment based on image recognition and storage medium |
| CN113313264A (en) * | 2021-06-02 | 2021-08-27 | 河南大学 | Efficient federal learning method in Internet of vehicles scene |
| CN113313264B (en) * | 2021-06-02 | 2022-08-12 | 河南大学 | Efficient federal learning method in Internet of vehicles scene |
| CN113920780A (en) * | 2021-09-01 | 2022-01-11 | 同济大学 | Cloud and mist collaborative personalized forward collision risk early warning method based on federal learning |
| CN114740970A (en) * | 2022-02-23 | 2022-07-12 | 广东工业大学 | Millimeter wave gesture recognition method and system based on federal learning |
| CN114740970B (en) * | 2022-02-23 | 2024-05-24 | 广东工业大学 | Millimeter wave gesture recognition method and system based on federal learning |
| CN114548608A (en) * | 2022-04-26 | 2022-05-27 | 腾讯科技（深圳）有限公司 | Model processing method and device, target traffic equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112052959B (en) | 2023-08-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112052959B (en) | Automatic driving training method, equipment and medium based on federal learning | |
| Xu et al. | V2x-vit: Vehicle-to-everything cooperative perception with vision transformer | |
| CN113366497B (en) | Agent Prioritization for Autonomous Vehicles | |
| Lin et al. | Intelligent transportation system (ITS): Concept, challenge and opportunity | |
| US20200209874A1 (en) | Combined virtual and real environment for autonomous vehicle planning and control testing | |
| US20210302585A1 (en) | Smart navigation method and system based on topological map | |
| US20220215667A1 (en) | Method and apparatus for monitoring vehicle, cloud control platform and system for vehicle-road collaboration | |
| US11210770B2 (en) | AI-based inspection in transportation | |
| JP2019215849A (en) | Method, device, apparatus, and medium for classifying driving scene data | |
| WO2021135948A1 (en) | Data transmission method and device | |
| US11967103B2 (en) | Multi-modal 3-D pose estimation | |
| CN111859597A (en) | Evaluation method and system of automatic driving algorithm | |
| US20230078241A1 (en) | Driving assistance processing method and apparatus, computer-readable medium, and electronic device | |
| CN114228743B (en) | Unmanned logistics vehicle control method, device and system and readable storage medium | |
| US20230080076A1 (en) | Platooning processing method and apparatus, computer-readable medium, and electronic device | |
| CN113326826A (en) | Network model training method and device, electronic equipment and storage medium | |
| CN113189989B (en) | Vehicle intention prediction method, device, equipment and storage medium | |
| US11745762B2 (en) | System and methods of adaptive trajectory prediction for autonomous driving | |
| CN114821537A (en) | Activity intention prediction method and device and unmanned vehicle | |
| US20220084228A1 (en) | Estimating ground truth object keypoint labels for sensor readings | |
| CN115061386A (en) | Intelligent driving automatic simulation test system and related equipment | |
| AlKishri et al. | Object recognition for organizing the movement of self-driving car | |
| CN112150811A (en) | Urban road intersection traffic information acquisition system based on image recognition technology | |
| CN112766138A (en) | Positioning method, device and equipment based on image recognition and storage medium | |
| CN112861701A (en) | Illegal parking identification method and device, electronic equipment and computer readable medium |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |