CN110478911A - The unmanned method of intelligent game vehicle and intelligent vehicle, equipment based on machine learning - Google Patents
The unmanned method of intelligent game vehicle and intelligent vehicle, equipment based on machine learning Download PDFInfo
- Publication number
- CN110478911A CN110478911A CN201910744970.5A CN201910744970A CN110478911A CN 110478911 A CN110478911 A CN 110478911A CN 201910744970 A CN201910744970 A CN 201910744970A CN 110478911 A CN110478911 A CN 110478911A
- Authority
- CN
- China
- Prior art keywords
- unmanned
- training
- model
- image data
- scene image
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 65
- 238000010801 machine learning Methods 0.000 title claims abstract description 45
- 238000012549 training Methods 0.000 claims abstract description 93
- 238000005457 optimization Methods 0.000 claims description 12
- 238000001514 detection method Methods 0.000 claims description 10
- 230000003252 repetitive effect Effects 0.000 claims description 7
- 238000012795 verification Methods 0.000 claims description 7
- 230000002159 abnormal effect Effects 0.000 claims description 6
- 238000000605 extraction Methods 0.000 claims description 6
- 230000000007 visual effect Effects 0.000 abstract description 4
- 238000010586 diagram Methods 0.000 description 8
- 238000004891 communication Methods 0.000 description 7
- 230000005540 biological transmission Effects 0.000 description 5
- 238000013135 deep learning Methods 0.000 description 4
- 238000012360 testing method Methods 0.000 description 4
- 238000007781 pre-processing Methods 0.000 description 3
- 238000013473 artificial intelligence Methods 0.000 description 2
- 230000015572 biosynthetic process Effects 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 230000006399 behavior Effects 0.000 description 1
- 238000004590 computer program Methods 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 230000013016 learning Effects 0.000 description 1
- 238000003032 molecular docking Methods 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 230000036632 reaction speed Effects 0.000 description 1
Classifications
-
- A—HUMAN NECESSITIES
- A63—SPORTS; GAMES; AMUSEMENTS
- A63F—CARD, BOARD, OR ROULETTE GAMES; INDOOR GAMES USING SMALL MOVING PLAYING BODIES; VIDEO GAMES; GAMES NOT OTHERWISE PROVIDED FOR
- A63F13/00—Video games, i.e. games using an electronically generated display having two or more dimensions
- A63F13/25—Output arrangements for video game devices
-
- A—HUMAN NECESSITIES
- A63—SPORTS; GAMES; AMUSEMENTS
- A63F—CARD, BOARD, OR ROULETTE GAMES; INDOOR GAMES USING SMALL MOVING PLAYING BODIES; VIDEO GAMES; GAMES NOT OTHERWISE PROVIDED FOR
- A63F13/00—Video games, i.e. games using an electronically generated display having two or more dimensions
- A63F13/80—Special adaptations for executing a specific game genre or game mode
- A63F13/803—Driving vehicles or craft, e.g. cars, airplanes, ships, robots or tanks
-
- A—HUMAN NECESSITIES
- A63—SPORTS; GAMES; AMUSEMENTS
- A63F—CARD, BOARD, OR ROULETTE GAMES; INDOOR GAMES USING SMALL MOVING PLAYING BODIES; VIDEO GAMES; GAMES NOT OTHERWISE PROVIDED FOR
- A63F13/00—Video games, i.e. games using an electronically generated display having two or more dimensions
- A63F13/90—Constructional details or arrangements of video game devices not provided for in groups A63F13/20 or A63F13/25, e.g. housing, wiring, connections or cabinets
- A63F13/98—Accessories, i.e. detachable arrangements optional for the use of the video game device, e.g. grip supports of game controllers
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- G—PHYSICS
- G09—EDUCATION; CRYPTOGRAPHY; DISPLAY; ADVERTISING; SEALS
- G09B—EDUCATIONAL OR DEMONSTRATION APPLIANCES; APPLIANCES FOR TEACHING, OR COMMUNICATING WITH, THE BLIND, DEAF OR MUTE; MODELS; PLANETARIA; GLOBES; MAPS; DIAGRAMS
- G09B9/00—Simulators for teaching or training purposes
- G09B9/02—Simulators for teaching or training purposes for teaching control of vehicles or other craft
- G09B9/04—Simulators for teaching or training purposes for teaching control of vehicles or other craft for teaching control of land vehicles
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Software Systems (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Medical Informatics (AREA)
- Evolutionary Computation (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Mathematical Physics (AREA)
- Artificial Intelligence (AREA)
- Aviation & Aerospace Engineering (AREA)
- Business, Economics & Management (AREA)
- Educational Administration (AREA)
- Educational Technology (AREA)
- Image Analysis (AREA)
Abstract
The embodiment of the present invention discloses a kind of unmanned method of intelligent game vehicle and intelligent vehicle, equipment based on machine learning, and wherein method includes the following steps: the scene image data that current driving road is acquired based on FPV；Training is iterated to scene image data using machine learning method, obtains unmanned Controlling model；It is carried out based on unmanned Controlling model unmanned.Using the present invention, pass through the sample data of FPV collection model training, the training sample that can guarantee acquisition is the data at normal driving visual angle, then by the unmanned model of machine learning training, the model trained while improving model training efficiency can preferably meet the unmanned of intelligent game vehicle.
Description
Technical field
The present invention relates to the unmanned technical field of intelligent vehicle more particularly to a kind of intelligent game vehicles based on machine learning
Unmanned method and intelligent vehicle, equipment.
Background technique
With the fast development of artificial intelligence, it is especially unmanned that artificial intelligence has in depth been applied to intelligent driving
Technical field, intelligent vehicle can carry out the training of unmanned model by modes such as supervised learnings at present, and model training
Training data be collected by driver's manual drive intelligent vehicle, and need to acquire a large amount of training data and to data into
Rower note could preferably train unmanned model, be greatly lowered the efficiency of model training in this way, and according to certain
The non-genuine intelligent vehicle of game or the teaching of scale smaller causes to instruct since true man can not be used to drive acquisition training data
The model practised is not well positioned to meet the unmanned of game car.
Summary of the invention
The embodiment of the present invention provides a kind of unmanned method of intelligent game vehicle based on machine learning and intelligent vehicle, sets
It is standby, pass through the sample data of FPV collection model training, it is ensured that the training sample of acquisition is the number at normal driving visual angle
According to then by the unmanned model of machine learning training, the model trained while improving model training efficiency can be more
Good meets the unmanned of intelligent game vehicle.
First aspect of the embodiment of the present invention provides a kind of unmanned method of intelligent game vehicle based on machine learning, can
Include:
Scene image data based on FPV acquisition current driving road；
Training is iterated to scene image data using machine learning method, obtains unmanned Controlling model；
It is carried out based on unmanned Controlling model unmanned.
Further, the above-mentioned scene image data based on FPV acquisition current driving road, comprising:
The Driving control operation that user is carried out based on FPV terminal is obtained, corresponding control instruction is operated based on Driving control
Carry out controlled traveling；
During controlled traveling, the scene image data in driving process is acquired according to default frequency acquisition.
Further, the above method further include:
Whether normal detect unpiloted driving process；
When the result of detection is abnormal, it is transferred to and executes scene image of the step based on FPV acquisition current driving road
Data carry out model optimization training.
Further, above-mentioned that training is iterated to scene image data using machine learning method, it obtains unmanned
Controlling model, comprising:
Pre-processed to obtain the corresponding steering angle of each frame image in scene and speed to scene image data, it is pre- to locate
Reason includes image study and feature extraction；
Training is iterated to pretreated scene image data using machine learning method, obtains unmanned control
Model.
Further, the above method further include:
During repetitive exercise, calculates training sample and verify the loss late between sample；
When loss late is less than loss-rate threshold, determine that the unmanned model trained is valid model；
Wherein, training sample is the training set that scene image data is formed, and verifying sample is based on scene image data
The empirical verification set of formation.
Second aspect of the embodiment of the present invention provides a kind of intelligent game vehicle based on machine learning, it may include:
Data acquisition module, for the scene image data based on FPV acquisition current driving road；
Model training module obtains nobody for being iterated training to scene image data using machine learning method
Driving control model；
Model uses module, unmanned for being carried out based on unmanned Controlling model.
Further, above-mentioned data acquisition module includes:
Operation control unit, the Driving control operation carried out for obtaining user based on FPV terminal, is grasped based on Driving control
Make corresponding control instruction and carries out controlled traveling；
Data acquisition unit, for being acquired in driving process according to default frequency acquisition during controlled traveling
Scene image data.
Further, above-mentioned intelligent vehicle further include:
Detection module is driven, it is whether normal for detecting unpiloted driving process；
Model optimization module executes step based on FPV acquisition current line for being transferred to when the result of detection is abnormal
The scene image data of road is sailed, model optimization training is carried out.
Further, above-mentioned model training module includes:
Data pre-processing unit, it is corresponding for being pre-processed to obtain each frame image in scene to scene image data
Steering angle and speed, pretreatment include image study and feature extraction；
Model training unit, for being iterated instruction to pretreated scene image data using machine learning method
Practice, obtains unmanned Controlling model.
Further, above-mentioned intelligent game vehicle further include:
Loss late computing module, for during repetitive exercise, calculating training sample and verifying the damage between sample
Mistake rate；
Valid model determining module, for determining the unmanned mould trained when loss late is less than loss-rate threshold
Type is valid model；
Wherein, training sample is the training set that scene image data is formed, and verifying sample is based on scene image data
The empirical verification set of formation.
The third aspect of the embodiment of the present invention provides a kind of electronic equipment, which includes processor and memory, described
At least one instruction, at least a Duan Chengxu, code set or instruction set are stored in memory, described at least one instructs, is described
An at least Duan Chengxu, the code set or instruction set loaded as the processor and executed with realize described in above-mentioned aspect based on
The unmanned method of intelligent game vehicle of machine learning.
Fourth aspect of the embodiment of the present invention provides a kind of computer storage medium, is stored in the computer storage medium
At least one instruction, at least a Duan Chengxu, code set or instruction set, at least one instruction, an at least Duan Chengxu, institute
Code set or instruction set is stated to be loaded as processor and executed to realize the intelligent game described in above-mentioned aspect based on machine learning
The unmanned method of vehicle.
In embodiments of the present invention, it by the FPV sample data that training acquisition is trained manually, realizes and is regarded based on machine
Feel and the training of the unmanned model of machine learning also makes to train resulting nothing under the premise of guaranteeing model training efficiency
People's driving model has preferably been used on intelligent game vehicle.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
Some embodiments of invention for those of ordinary skill in the art without creative efforts, can be with
It obtains other drawings based on these drawings.
Fig. 1 is a kind of process of unmanned method of intelligent game vehicle based on machine learning provided in an embodiment of the present invention
Schematic diagram；
Fig. 2 is the stream of another unmanned method of intelligent game vehicle based on machine learning provided in an embodiment of the present invention
Journey schematic diagram；
Fig. 3 is a kind of structural schematic diagram of intelligent vehicle provided in an embodiment of the present invention；
Fig. 4 is the structural schematic diagram of data acquisition module provided in an embodiment of the present invention；
Fig. 5 is the structural schematic diagram of model training module provided in an embodiment of the present invention；
Fig. 6 is the structural schematic diagram of a kind of electronic equipment provided in an embodiment of the present invention.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts every other
Embodiment shall fall within the protection scope of the present invention.
Term " includes " in description and claims of this specification and above-mentioned attached drawing and " having " and they appoint
What is deformed, it is intended that covering non-exclusive includes that term " first " and " second " be not merely to difference name, represents number
The size or sequence of word.Such as it contains the process, method, system, product or equipment of a series of steps or units and does not limit
Due to listed step or unit, but optionally further comprising the step of not listing or unit, or optionally further comprising it is right
In other intrinsic step or units of these process, methods, product or equipment.
It should be noted that the intelligent game vehicle unmanned method provided by the present application based on machine learning can be applied
In unpiloted intelligent game vehicle, the intelligent carriage etc. of teaching application scenarios in.
In the embodiment of the present invention, the unmanned method of intelligent game vehicle based on machine learning can be applied to electronic equipment
In, the computer equipment can be embedded microcomputer intelligent game vehicle or it is other have calculating, processing capacity and
Robot with similar vehicles driving functions is also possible to the unmanned control device carried on true intelligent automobile.
It should be noted that being equipped with the software or device for carrying out data interaction with the FPV helmet in the application, in electronic equipment, and should
Electronic equipment has the function of wireless image transmission.
As shown in Figure 1, the unmanned method of intelligent game vehicle based on machine learning at least may include following step
It is rapid:
S101, the scene image data based on FPV acquisition current driving road.
It is understood that the figure at the first visual angle for the scene can be acquired for existing road scene intelligent vehicle
As data, specific implementation, which may is that, obtains the Driving control operation that user is carried out based on FPV terminal, is based on the operation pair
The control instruction answered carries out controlled traveling, during controlled traveling, is acquired in driving process according to default frequency acquisition
Scene image data.For example, driver can see the current driving road of intelligent vehicle by wearing the FPV helmet with the first visual angle
Scene, then according to current scene make drive reaction carry out Driving control operation (process be similar to present Dummy
Feel game, user has impression on the spot in person as long as wearing relevant device when playing game), further, driven controlled
During, intelligent vehicle carries out the frequency of 20 frame per second or so by the FPV docking facilities installed on vehicle or built-in software
Acquire image.
In an alternative embodiment, FPV implementation is as follows: passing through WIFI or other wireless transmission sides on game intelligence vehicle
Formula, transmits going game intelligent vehicle camera acquired image data, and the receiving end of wireless image transmission includes tablet computer or general
The above-mentioned image data of the reception such as talking computer, the FPV realtime graphic that then user receives according to tablet computer or common computer etc., into
For row to the real-time control of game intelligence vehicle, the mode of control includes passing through WIFI or other wireless transmission methods, will be current: preceding
Into, retreat, turn left (including steering angle), turn right (including steering angle) and brake and the command adapted thereto accelerated are issued
Game intelligence vehicle, game intelligence vehicle receive instruction in real time and are manipulated immediately by instruction.
It is understood that can achieve by the time interval that FPV carries out game intelligence vehicle acquisition image data per second
10~100 frames or so certainly here will the wireless image transmission data capability according to game intelligence vehicle and the behaviour between FPV terminal
Control reaction speed is adjusted, and theoretically the time interval of normal speed is that 20 frame per second belongs to average value ranges.
In an alternative embodiment, can stop when intelligent vehicle has travelled 10 times or cycleoperation reaches 10 times or more
Manipulation stops data acquisition, is transferred to the process of data prediction and model training.Specific data collection cycle can regard mould
Type optimization degree adjust at any time, for example, for the first time model training be based on the training data for acquiring 10 times, but finally training
Modelling effect is not good enough, then during the circuit training of next time, can acquire the scene image data of 15 times driving processes.
S102 is iterated training to scene image data using machine learning method, obtains unmanned Controlling model.
In the specific implementation, machine learning method can be deep learning algorithm or improved deep learning algorithm, right
Model is iterated before training, can be pre-processed to obtain corresponding turn of each frame image in scene to scene image data
To angle and speed, training is then iterated for steering angle and speed using deep learning algorithm, is obtained unmanned
Controlling model.It is understood that above-mentioned steering angle and speed can be obtained by image study and feature extraction, in iteration
During, model is continuously available optimization, number when specific the number of iterations is optimal for model, such as can be 10 times
It is arrived between hundreds of times above.
It should be noted that can calculate training sample during repetitive exercise and verify the loss between sample
Rate determines that the unmanned model trained is valid model, i.e. currently trained mould when loss late is less than loss-rate threshold
The result that type meets deep learning is also that can satisfy unpiloted model.It is understood that wherein, training sample can be with
It is the training set that scene image data is formed, verifying sample is the empirical verification set formed based on scene image data.
It should be noted that in general loss-rate threshold is the smaller the better, but in the training process, sometimes when Loss rate
(loss-rate threshold) at 20%~40%, obtained unmanned Controlling model is also possible to can be used normally, this mistake
Journey need test and test, in different racing track scenes in the following, loss-rate threshold with scene, vision camera, illumination, interference because
The objective factors such as element, picture noise are varied.
S103 is carried out unmanned based on unmanned Controlling model.
It, can be with it should be noted that when carrying out unmanned based on the unmanned model cootrol intelligent vehicle trained
Detect whether entire driving procedure is normal, think that trained model is not good enough if there is exception, it is possible to due to acquisition
Model training less effective caused by training data is insufficient or the number of iterations reasons such as not enough, need to resurvey data into
The above-mentioned training process of row, until model is optimal.
In embodiments of the present invention, it by the FPV sample data that training acquisition is trained manually, realizes and is regarded based on machine
Feel and the training of the unmanned model of machine learning also makes to train resulting nothing under the premise of guaranteeing model training efficiency
People's driving model has preferably been used on intelligent game vehicle.
Fig. 2 is the stream of another unmanned method of intelligent game vehicle based on machine learning provided in an embodiment of the present invention
Journey schematic diagram, comprising:
S201 obtains the Driving control operation that user is carried out based on FPV terminal, operates corresponding control based on Driving control
Instruction carries out controlled traveling.
S202 acquires the scene image data in driving process according to default frequency acquisition.
S203 is pre-processed to obtain the corresponding steering angle of each frame image in scene and speed to scene image data
Degree.
S204 is iterated training to pretreated scene image data using machine learning method, obtains nobody and drive
Sail Controlling model.
It is understood that can calculate training sample during repetitive exercise and verify the loss between sample
Rate, and when loss late is less than loss-rate threshold, determine that the unmanned model trained is valid model.Wherein, training sample
This is the training set that scene image data is formed, and verifying sample is the empirical verification set formed based on scene image data.
Whether normal S205 detects unpiloted driving process.
It is understood that being transferred to execution step S206 terminates entire control process, instead when above-mentioned testing result, which is, is
It, is transferred to and executes step S201, repeats above-mentioned link, until testing result is normal.
S206 terminates.
It should be noted that the detailed implementation of the present embodiment may refer to the specific descriptions of embodiment one kind, herein
It repeats no more.
In embodiments of the present invention, it by the FPV sample data that training acquisition is trained manually, realizes and is regarded based on machine
Feel and the training of the unmanned model of machine learning also makes to train resulting nothing under the premise of guaranteeing model training efficiency
People's driving model has preferably been used on intelligent game vehicle.
Below in conjunction with attached drawing 3- attached drawing 5, to the intelligent game vehicle provided in an embodiment of the present invention based on machine learning into
Row is discussed in detail.It should be noted that the attached intelligent vehicle shown in fig. 5 of attached drawing 3-, real shown in Fig. 1 and Fig. 2 of the present invention for executing
The method for applying example, for ease of description, only parts related to embodiments of the present invention are shown, and particular technique details does not disclose
, please refer to Fig. 1 of the present invention and embodiment shown in Fig. 2.
Fig. 3 is referred to, for the embodiment of the invention provides a kind of structural schematic diagrams of intelligent vehicle.As shown in figure 3, of the invention
The intelligent vehicle 10 of embodiment may include: data acquisition module 101, model training module 102, model using module 103, drive
Detection module 104, model optimization module 105, loss late computing module 106 and valid model determining module 107.Wherein, data
Acquisition module 101 is as shown in figure 4, include operation control unit 1011 and data acquisition unit 1012, model training module 102 is such as
Shown in Fig. 5, including data pre-processing unit 1021 and model training unit 1022.
Data acquisition module 101, for the scene image data based on FPV acquisition current driving road.
Model training module 102 is obtained for being iterated training to the scene image data using machine learning method
To unmanned Controlling model.
In an alternative embodiment, model training module 102 includes:
Data pre-processing unit 1021, for being pre-processed to obtain each frame figure in scene to the scene image data
As corresponding steering angle and speed, the pretreatment includes image study and feature extraction.
Model training unit 1022, for being iterated using machine learning method to pretreated scene image data
Training, obtains unmanned Controlling model.
Model uses module 103, unmanned for being carried out based on the unmanned Controlling model.
In an alternative embodiment, data acquisition module 101 includes:
Operation control unit 1011, the Driving control operation carried out for obtaining user based on FPV terminal, is driven based on described
It sails the corresponding control instruction of control operation and carries out controlled traveling.
Data acquisition unit 1012, for acquiring driving process according to default frequency acquisition during controlled traveling
In scene image data.
Detection module 104 is driven, it is whether normal for detecting the unpiloted driving process；
Model optimization module 105 is adopted for when the result of the detection is abnormal, being transferred to execution step based on FPV
Collect the scene image data of current driving road, carries out model optimization training.
Loss late computing module 106, for during the repetitive exercise, calculate training sample and verifying sample it
Between loss late；
Valid model determining module 107, for determining nobody trained when the loss late is less than loss-rate threshold
Driving model is valid model；
Wherein, the training sample be the scene image data formed training set, the verifying sample for based on
The empirical verification set that the scene image data is formed.
It should be noted that the implementation procedure of each unit module may refer to above method embodiment in the embodiment of the present invention
In specific descriptions, details are not described herein again.
In embodiments of the present invention, it by the FPV sample data that training acquisition is trained manually, realizes and is regarded based on machine
Feel and the training of the unmanned model of machine learning also makes to train resulting nothing under the premise of guaranteeing model training efficiency
People's driving model has preferably been used on intelligent game vehicle.
The embodiment of the invention also provides a kind of computer storage medium, the computer storage medium can store more
Item instruction, described instruction are suitable for being loaded by processor and executing the method and step such as above-mentioned Fig. 1 and embodiment illustrated in fig. 2, specifically
Implementation procedure may refer to illustrating for Fig. 1 and embodiment illustrated in fig. 2, herein without repeating.
The embodiment of the present application also provides a kind of electronic equipment.As shown in fig. 6, electronic equipment 20 may include: at least one
A processor 201, such as CPU, at least one radio network interface 204, memory 205, at least one communication bus 202, until
A few vision camera 203 can also include optionally display screen 206.Wherein, communication bus 202 is for realizing these groups
Connection communication between part.Wherein, network interface 204 optionally may include standard wireline interface and wireless interface (such as WI-
FI interface), it can be established and be communicated to connect with server by network interface 204.Memory 205 can be high speed RAM memory,
It is also possible to non-labile memory (non-volatile memory), for example, at least a magnetic disk storage, memory
205 include the flash in the embodiment of the present invention.Memory 205 optionally can also be that at least one is located remotely from aforementioned processing
The storage system of device 201.As shown in fig. 6, as may include operation system in a kind of memory 205 of computer storage medium
System, network communication module, Subscriber Interface Module SIM and program instruction.
It should be noted that network interface 204 can connect receiver, transmitter or other communication modules, other communications
Module can include but is not limited to WiFi module, bluetooth module etc., it will be understood that electronic equipment can also be in the embodiment of the present invention
Including receiver, transmitter and other communication modules etc..
Processor 201 can be used for calling the program instruction stored in memory 205, and it is following to execute electronic equipment 20
Operation:
Scene image data based on FPV acquisition current driving road；
Training is iterated to the scene image data using machine learning method, obtains unmanned Controlling model；
It is carried out based on the unmanned Controlling model unmanned.
In some embodiments, equipment 20 is specific to use when based on the scene image data of FPV acquisition current driving road
In:
The Driving control operation that user is carried out based on FPV terminal is obtained, corresponding control is operated based on the Driving control
Instruction carries out controlled traveling；
During controlled traveling, the scene image data in driving process is acquired according to default frequency acquisition.
In some embodiments, equipment 20 is also used to:
Whether normal detect the unpiloted driving process；
When the result of the detection is abnormal, it is transferred to and executes scene of the step based on FPV acquisition current driving road
Image data carries out model optimization training.
In some embodiments, equipment 20 is being iterated instruction to the scene image data using machine learning method
Practice, when obtaining unmanned Controlling model, be specifically used for:
Pre-processed to obtain the corresponding steering angle of each frame image in scene and speed to the scene image data,
The pretreatment includes image study and feature extraction；
Training is iterated to pretreated scene image data using machine learning method, obtains unmanned control
Model.
In some embodiments, equipment 20 is also used to:
During the repetitive exercise, calculates training sample and verify the loss late between sample；
When the loss late is less than loss-rate threshold, determine that the unmanned model trained is valid model；
Wherein, the training sample be the scene image data formed training set, the verifying sample for based on
The empirical verification set that the scene image data is formed.
In embodiments of the present invention, it by the FPV sample data that training acquisition is trained manually, realizes and is regarded based on machine
Feel and the training of the unmanned model of machine learning also makes to train resulting nothing under the premise of guaranteeing model training efficiency
People's driving model has preferably been used on intelligent game vehicle.
Those of ordinary skill in the art will appreciate that realizing all or part of the process in above-described embodiment method, being can be with
Relevant hardware is instructed to complete by computer program, the program can be stored in computer-readable storage medium
In, the program is when being executed, it may include such as the process of the embodiment of above-mentioned each method.Wherein, the storage medium can be magnetic
Dish, CD, read-only memory (Read-Only Memory, ROM) or random access memory (Random Access
Memory, RAM) etc..
The above disclosure is only the preferred embodiments of the present invention, cannot limit the right model of the present invention with this certainly
It encloses, therefore equivalent changes made in accordance with the claims of the present invention, is still within the scope of the present invention.
Claims (10)
1. a kind of unmanned method of intelligent game vehicle based on machine learning characterized by comprising
Scene image data based on FPV acquisition current driving road；
Training is iterated to the scene image data using machine learning method, obtains unmanned Controlling model；
It is carried out based on the unmanned Controlling model unmanned.
2. the method according to claim 1, wherein the scene figure based on FPV acquisition current driving road
As data, comprising:
The Driving control operation that user is carried out based on FPV terminal is obtained, corresponding control instruction is operated based on the Driving control
Carry out controlled traveling；
During controlled traveling, the scene image data in driving process is acquired according to default frequency acquisition.
3. the method according to claim 1, wherein the method also includes:
Whether normal detect the unpiloted driving process；
When the result of the detection is abnormal, it is transferred to and executes scene image of the step based on FPV acquisition current driving road
Data carry out model optimization training.
4. the method according to claim 1, wherein described use machine learning method to the scene image number
According to training is iterated, unmanned Controlling model is obtained, comprising:
Pre-processed to obtain the corresponding steering angle of each frame image in scene and speed to the scene image data, it is described
Pretreatment includes image study and feature extraction；
Training is iterated to pretreated scene image data using machine learning method, obtains unmanned control mould
Type.
5. according to the method described in claim 4, it is characterized in that, the method also includes:
During the repetitive exercise, calculates training sample and verify the loss late between sample；
When the loss late is less than loss-rate threshold, determine that the unmanned model trained is valid model；
Wherein, the training sample is the training set that the scene image data is formed, and the verifying sample is based on described
The empirical verification set that scene image data is formed.
6. a kind of intelligent game vehicle based on machine learning characterized by comprising
Data acquisition module, for the scene image data based on FPV acquisition current driving road；
Model training module obtains nobody for being iterated training to the scene image data using machine learning method
Driving control model；
Model uses module, unmanned for being carried out based on the unmanned Controlling model.
7. intelligent game vehicle according to claim 6, which is characterized in that the data acquisition module includes:
Operation control unit, the Driving control operation carried out for obtaining user based on FPV terminal, is grasped based on the Driving control
Make corresponding control instruction and carries out controlled traveling；
Data acquisition unit, for acquiring the scene in driving process according to default frequency acquisition during controlled traveling
Image data.
8. intelligent game vehicle according to claim 7, which is characterized in that the intelligent vehicle further include:
Detection module is driven, it is whether normal for detecting the unpiloted driving process；
Model optimization module executes step based on FPV acquisition current line for being transferred to when the result of the detection is abnormal
The scene image data of road is sailed, model optimization training is carried out.
9. a kind of electronic equipment, which is characterized in that the electronic equipment includes processor and memory, is stored in the memory
Have at least one instruction, at least a Duan Chengxu, code set or instruction set, at least one instruction, an at least Duan Chengxu,
The code set or instruction set are loaded by the processor and are executed to realize and be based on as described in any one of claim 1 to 5
The unmanned method of intelligent game vehicle of machine learning.
10. a kind of computer readable storage medium, which is characterized in that be stored at least one instruction, extremely in the storage medium
A few Duan Chengxu, code set or instruction set, at least one instruction, an at least Duan Chengxu, the code set or instruction
Collection is loaded by processor and is executed to realize such as the intelligent game vehicle described in any one of claim 1 to 5 based on machine learning
Unmanned method.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910744970.5A CN110478911A (en) | 2019-08-13 | 2019-08-13 | The unmanned method of intelligent game vehicle and intelligent vehicle, equipment based on machine learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910744970.5A CN110478911A (en) | 2019-08-13 | 2019-08-13 | The unmanned method of intelligent game vehicle and intelligent vehicle, equipment based on machine learning |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110478911A true CN110478911A (en) | 2019-11-22 |
Family
ID=68550999
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910744970.5A Pending CN110478911A (en) | 2019-08-13 | 2019-08-13 | The unmanned method of intelligent game vehicle and intelligent vehicle, equipment based on machine learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110478911A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111126336A (en) * | 2019-12-31 | 2020-05-08 | 潍柴动力股份有限公司 | Sample collection method, device and equipment |
| CN112382165A (en) * | 2020-11-19 | 2021-02-19 | 北京罗克维尔斯科技有限公司 | Driving strategy generation method, device, medium, equipment and simulation system |
| WO2023178510A1 (en) * | 2022-03-22 | 2023-09-28 | 深圳市大疆创新科技有限公司 | Image processing method, device, and system and movable platform |
Citations (22)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103871079A (en) * | 2014-03-18 | 2014-06-18 | 南京金智视讯技术有限公司 | Vehicle tracking method based on machine learning and optical flow |
| CN105488534A (en) * | 2015-12-04 | 2016-04-13 | 中国科学院深圳先进技术研究院 | Method, device and system for deeply analyzing traffic scene |
| CN105718879A (en) * | 2016-01-19 | 2016-06-29 | 华南理工大学 | Free-scene egocentric-vision finger key point detection method based on depth convolution nerve network |
| CN105760894A (en) * | 2016-03-16 | 2016-07-13 | 哈尔滨伟方智能科技开发有限责任公司 | Robot navigation method based on machine vision and machine learning |
| CN106080590A (en) * | 2016-06-12 | 2016-11-09 | 百度在线网络技术（北京）有限公司 | Control method for vehicle and device and the acquisition methods of decision model and device |
| CN106682649A (en) * | 2017-01-24 | 2017-05-17 | 成都容豪电子信息科技有限公司 | Vehicle type recognition method based on deep learning |
| CN107169567A (en) * | 2017-03-30 | 2017-09-15 | 深圳先进技术研究院 | The generation method and device of a kind of decision networks model for Vehicular automatic driving |
| CN107491073A (en) * | 2017-09-05 | 2017-12-19 | 百度在线网络技术（北京）有限公司 | The data training method and device of automatic driving vehicle |
| US20180081356A1 (en) * | 2016-09-19 | 2018-03-22 | The Geek Group | Multirotor game system |
| CN108139804A (en) * | 2015-10-15 | 2018-06-08 | 索尼公司 | Information processing unit and information processing method |
| CN108229489A (en) * | 2016-12-30 | 2018-06-29 | 北京市商汤科技开发有限公司 | Crucial point prediction, network training, image processing method, device and electronic equipment |
| CN108319289A (en) * | 2017-01-16 | 2018-07-24 | 翔升（上海）电子技术有限公司 | Head-wearing display device, unmanned plane, flight system and unmanned aerial vehicle (UAV) control method |
| CN108388873A (en) * | 2018-03-01 | 2018-08-10 | 路志宏 | A kind of supervisory systems of machine vision, method and client computer, storage medium |
| CN108416783A (en) * | 2018-02-01 | 2018-08-17 | 湖北工业大学 | Road scene dividing method based on full convolutional Neural network |
| CN108520220A (en) * | 2018-03-30 | 2018-09-11 | 百度在线网络技术（北京）有限公司 | model generating method and device |
| CN108540451A (en) * | 2018-03-13 | 2018-09-14 | 北京理工大学 | A method of classification and Detection being carried out to attack with machine learning techniques |
| CN108875595A (en) * | 2018-05-29 | 2018-11-23 | 重庆大学 | A kind of Driving Scene object detection method merged based on deep learning and multilayer feature |
| CN109040555A (en) * | 2018-08-22 | 2018-12-18 | 信利光电股份有限公司 | A kind of display system and method for FPV image |
| CN109448499A (en) * | 2019-01-07 | 2019-03-08 | 北京智能佳科技有限公司 | Miniature multilevel traffic Simulation Experimental Platform |
| CN109658688A (en) * | 2017-10-11 | 2019-04-19 | 深圳市哈工大交通电子技术有限公司 | The detection method and device of access connection traffic flow based on deep learning |
| CN109685066A (en) * | 2018-12-24 | 2019-04-26 | 中国矿业大学(北京) | A kind of mine object detection and recognition method based on depth convolutional neural networks |
| KR20190063883A (en) * | 2017-11-30 | 2019-06-10 | 로보웍스(주) | System for providing augmented reality interactive game contents using a drones |
-
2019
- 2019-08-13 CN CN201910744970.5A patent/CN110478911A/en active Pending
Patent Citations (22)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103871079A (en) * | 2014-03-18 | 2014-06-18 | 南京金智视讯技术有限公司 | Vehicle tracking method based on machine learning and optical flow |
| CN108139804A (en) * | 2015-10-15 | 2018-06-08 | 索尼公司 | Information processing unit and information processing method |
| CN105488534A (en) * | 2015-12-04 | 2016-04-13 | 中国科学院深圳先进技术研究院 | Method, device and system for deeply analyzing traffic scene |
| CN105718879A (en) * | 2016-01-19 | 2016-06-29 | 华南理工大学 | Free-scene egocentric-vision finger key point detection method based on depth convolution nerve network |
| CN105760894A (en) * | 2016-03-16 | 2016-07-13 | 哈尔滨伟方智能科技开发有限责任公司 | Robot navigation method based on machine vision and machine learning |
| CN106080590A (en) * | 2016-06-12 | 2016-11-09 | 百度在线网络技术（北京）有限公司 | Control method for vehicle and device and the acquisition methods of decision model and device |
| US20180081356A1 (en) * | 2016-09-19 | 2018-03-22 | The Geek Group | Multirotor game system |
| CN108229489A (en) * | 2016-12-30 | 2018-06-29 | 北京市商汤科技开发有限公司 | Crucial point prediction, network training, image processing method, device and electronic equipment |
| CN108319289A (en) * | 2017-01-16 | 2018-07-24 | 翔升（上海）电子技术有限公司 | Head-wearing display device, unmanned plane, flight system and unmanned aerial vehicle (UAV) control method |
| CN106682649A (en) * | 2017-01-24 | 2017-05-17 | 成都容豪电子信息科技有限公司 | Vehicle type recognition method based on deep learning |
| CN107169567A (en) * | 2017-03-30 | 2017-09-15 | 深圳先进技术研究院 | The generation method and device of a kind of decision networks model for Vehicular automatic driving |
| CN107491073A (en) * | 2017-09-05 | 2017-12-19 | 百度在线网络技术（北京）有限公司 | The data training method and device of automatic driving vehicle |
| CN109658688A (en) * | 2017-10-11 | 2019-04-19 | 深圳市哈工大交通电子技术有限公司 | The detection method and device of access connection traffic flow based on deep learning |
| KR20190063883A (en) * | 2017-11-30 | 2019-06-10 | 로보웍스(주) | System for providing augmented reality interactive game contents using a drones |
| CN108416783A (en) * | 2018-02-01 | 2018-08-17 | 湖北工业大学 | Road scene dividing method based on full convolutional Neural network |
| CN108388873A (en) * | 2018-03-01 | 2018-08-10 | 路志宏 | A kind of supervisory systems of machine vision, method and client computer, storage medium |
| CN108540451A (en) * | 2018-03-13 | 2018-09-14 | 北京理工大学 | A method of classification and Detection being carried out to attack with machine learning techniques |
| CN108520220A (en) * | 2018-03-30 | 2018-09-11 | 百度在线网络技术（北京）有限公司 | model generating method and device |
| CN108875595A (en) * | 2018-05-29 | 2018-11-23 | 重庆大学 | A kind of Driving Scene object detection method merged based on deep learning and multilayer feature |
| CN109040555A (en) * | 2018-08-22 | 2018-12-18 | 信利光电股份有限公司 | A kind of display system and method for FPV image |
| CN109685066A (en) * | 2018-12-24 | 2019-04-26 | 中国矿业大学(北京) | A kind of mine object detection and recognition method based on depth convolutional neural networks |
| CN109448499A (en) * | 2019-01-07 | 2019-03-08 | 北京智能佳科技有限公司 | Miniature multilevel traffic Simulation Experimental Platform |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111126336A (en) * | 2019-12-31 | 2020-05-08 | 潍柴动力股份有限公司 | Sample collection method, device and equipment |
| CN111126336B (en) * | 2019-12-31 | 2023-07-21 | 潍柴动力股份有限公司 | Sample collection method, device and equipment |
| CN112382165A (en) * | 2020-11-19 | 2021-02-19 | 北京罗克维尔斯科技有限公司 | Driving strategy generation method, device, medium, equipment and simulation system |
| WO2023178510A1 (en) * | 2022-03-22 | 2023-09-28 | 深圳市大疆创新科技有限公司 | Image processing method, device, and system and movable platform |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110146100B (en) | Trajectory prediction method, apparatus and storage medium | |
| CN113762252B (en) | Unmanned aerial vehicle intelligent following target determining method, unmanned aerial vehicle and remote controller | |
| CN110478911A (en) | The unmanned method of intelligent game vehicle and intelligent vehicle, equipment based on machine learning | |
| EP3796112B1 (en) | Virtual vehicle control method, model training method, control device and storage medium | |
| CN107609502A (en) | Method and apparatus for controlling automatic driving vehicle | |
| CN109785368A (en) | A kind of method for tracking target and device | |
| CN107481292A (en) | The attitude error method of estimation and device of vehicle-mounted camera | |
| CN109584299B (en) | Positioning method, positioning device, terminal and storage medium | |
| CN108734303A (en) | Vehicle drive data predication method, equipment and computer readable storage medium | |
| JP7273635B2 (en) | Vehicle image processing method, vehicle image processing apparatus and computer readable storage medium | |
| CN112068549A (en) | Unmanned system cluster control method based on deep reinforcement learning | |
| CN109657077A (en) | Model training method, lane line generation method, equipment and storage medium | |
| CN111079533A (en) | Unmanned vehicle driving decision method, unmanned vehicle driving decision device and unmanned vehicle | |
| US20230367934A1 (en) | Method and apparatus for constructing vehicle dynamics model and method and apparatus for predicting vehicle state information | |
| CN109196437A (en) | Intelligent driving method, apparatus and storage medium | |
| CN109493611A (en) | A kind of licence plate recognition method and device | |
| CN114972911A (en) | Method and equipment for collecting and processing output data of automatic driving perception algorithm model | |
| CN110723152B (en) | Artificial intelligence detection method, device, equipment or storage medium for guaranteeing automatic driving safety | |
| CN109919107B (en) | Traffic police gesture recognition method based on deep learning and unmanned vehicle | |
| CN115507867A (en) | Target trajectory prediction method, target trajectory prediction device, electronic device, and storage medium | |
| CN112671487A (en) | Vehicle testing method, server and testing vehicle | |
| CN108600691B (en) | Image acquisition method, device and system | |
| CN109683491A (en) | Vehicle-mounted camera analogue system | |
| CN109977884B (en) | Target following method and device | |
| CN109360436B (en) | Video generation method, terminal and storage 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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20191122 |