CN109697875A - Plan the method and device of driving trace - Google Patents
Plan the method and device of driving trace Download PDFInfo
- Publication number
- CN109697875A CN109697875A CN201710993734.8A CN201710993734A CN109697875A CN 109697875 A CN109697875 A CN 109697875A CN 201710993734 A CN201710993734 A CN 201710993734A CN 109697875 A CN109697875 A CN 109697875A
- Authority
- CN
- China
- Prior art keywords
- vehicle
- network model
- driving
- driving trace
- section
- 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
Classifications
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/09—Arrangements for giving variable traffic instructions
- G08G1/0962—Arrangements for giving variable traffic instructions having an indicator mounted inside the vehicle, e.g. giving voice messages
- G08G1/0967—Systems involving transmission of highway information, e.g. weather, speed limits
- G08G1/096708—Systems involving transmission of highway information, e.g. weather, speed limits where the received information might be used to generate an automatic action on the vehicle control
- G08G1/096725—Systems involving transmission of highway information, e.g. weather, speed limits where the received information might be used to generate an automatic action on the vehicle control where the received information generates an automatic action on the vehicle control
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
- G06Q10/047—Optimisation of routes or paths, e.g. travelling salesman problem
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/16—Anti-collision systems
- G08G1/166—Anti-collision systems for active traffic, e.g. moving vehicles, pedestrians, bikes
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Theoretical Computer Science (AREA)
- Strategic Management (AREA)
- Life Sciences & Earth Sciences (AREA)
- Economics (AREA)
- Data Mining & Analysis (AREA)
- Biomedical Technology (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Computational Linguistics (AREA)
- Atmospheric Sciences (AREA)
- Biophysics (AREA)
- General Health & Medical Sciences (AREA)
- Development Economics (AREA)
- Artificial Intelligence (AREA)
- Game Theory and Decision Science (AREA)
- Health & Medical Sciences (AREA)
- Entrepreneurship & Innovation (AREA)
- Marketing (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Traffic Control Systems (AREA)
Abstract
This application discloses a kind of method and devices for planning driving trace, belong to technical field of data processing.The described method includes: server determines target nerve network model corresponding with the section where the target position that the first vehicle is presently in, and the target nerve network model is sent to the first vehicle, so that the first vehicle determines driving trace by the target nerve network model.It is obtained since the target nerve network model is server according to the training of the driving trace of multiple second vehicles, and multiple second vehicle refers to the vehicle for passing through the section where the target position that the first vehicle is presently in front of current time.Namely, when the first vehicle is when target location determines the driving trace of itself by target nerve network model, the driving trace for passing through multiple second vehicles in the section where the target position before with reference to current time, improves the feasibility of the driving trace determined.
Description
Technical field
This application involves technical field of data processing, in particular to a kind of method and device for planning driving trace.
Background technique
Automatic Pilot, that is, planning driving trace by vehicle itself, and according to planning during vehicle driving
Driving trace traveling.However in practical applications, if the driving trace of vehicle planning is improper, be easy to other vehicles and
The safety of itself impacts, therefore how to plan that driving trace is just particularly important.
In the related technology, when needing to plan driving trace, vehicle determines the traveling task needed to be implemented and in current road
Locating lane position in road, and the lane position planning traveling rail locating according to determining traveling task and in present road
Mark.Wherein, traveling task includes keeping straight on, turn left, turn right and turning around, and locating lane position includes straight traffic in present road
Road, left turn lane and right-turn lane.For example, the traveling task determined is to turn right, locating lane position is in present road
Through Lane, the driving trace planned at this time can be with are as follows: from the center line for being presently in lane, along being presently in lane
Center line and right-hand lane center line between connecting line traveling, until reach right-hand lane center line.
However, the driving trace determined according to the method described above when vehicle is when driving, it is easy to accident occurs, that is to say, vehicle
Accident rate is higher when driving for determining driving trace according to the method described above, therefore, the traveling planned according to the method described above
The feasibility of track is lower.
Summary of the invention
In order to solve the problems, such as that the driving trace feasibility planned in the related technology is lower, this application provides a kind of planning
The method and device of driving trace.The technical solution is as follows:
In a first aspect, providing a kind of method for planning driving trace, it is applied to the first vehicle, this method comprises:
The target nerve network model that server is sent is received, the target nerve network model refers to and first vehicle
The corresponding neural network model in section where the target position being presently in, and the target nerve network model is described
Server is obtained according to the training of the driving trace of multiple second vehicles, and the multiple second vehicle refers to pass through before current time
Cross the vehicle in the section where the target position;
According to the traveling task of first vehicle and driving information and the driving information of the first obstacle vehicle, pass through
The target nerve network model determines the driving trace of first vehicle;
Wherein, the traveling task includes keeping straight on, turn left, turn right and turning around, and the driving information includes being presently in
Location information, driving direction and the travel speed of position, the first obstacle vehicle are the distance between described first vehicle
Less than the vehicle of pre-determined distance threshold value.
In this application, since the target nerve network model is server according to the driving trace of multiple second vehicles instruction
It gets, and multiple second vehicle refers to that current time passes through the road where the target position that the first vehicle is presently in front of
The vehicle of section.That is, when the first vehicle is when target location determines the driving trace of itself by target nerve network model,
The driving information for not only allowing for first vehicle and the first obstacle vehicle reference is also made to pass through the target position before current time
The driving trace of multiple second vehicles in the section where setting, to reduce the first vehicle according to the driving trace determined when driving
Accident rate, that is, improving the feasibility of the driving trace determined.
Optionally, the section where the target position is crossing;
It is described according to the traveling task and driving information of first vehicle and the driving information of the first obstacle vehicle,
The driving trace of first vehicle is determined by the target nerve network model, comprising:
By the traveling task of first vehicle and driving information, the driving information of the first obstacle vehicle and described
Input of the signal lamp state corresponding with the target position as the target nerve network model, passes through the mesh at crossing
Mark neural network model determines the driving trace of first vehicle.
Specifically, when the section where the target position is crossing, the is determined by target nerve network model at this time
When the driving trace of one vehicle, need to consider signal lamp state corresponding with the target position at the crossing, also to further increase
The feasibility for the driving trace determined.
Optionally, the target nerve network model refer to where the target position section and first vehicle
The corresponding neural network model of traveling task, and the multiple second vehicle refers to and passes through the target before current time
Section and traveling task vehicle identical with the traveling task of first vehicle where position.
Further, at some section, server can train different neural networks for different traveling tasks
Model, at this point, the first vehicle can be according to corresponding with the traveling task in section and the first vehicle where the target position
Target nerve network model determines driving trace.
Second aspect provides the method for another planning driving trace, is applied to server, this method comprises:
It is corresponding with the section where the target position that the first vehicle is presently in from determination in the neural network model of storage
Target nerve network model, the target nerve network model be according to the driving trace of multiple second vehicles training obtain
, the multiple second vehicle refers to the vehicle for passing through the section where the target position before current time;
The target nerve network model is sent to first vehicle, so that first vehicle passes through the target mind
The driving trace of first vehicle is determined through network model.
In this application, when the first vehicle driving to target position, server can be directly determined and the target position
The corresponding target nerve network model in the section at place, and the target nerve network model is sent to the first vehicle, so that first
Vehicle determines driving trace by the target nerve network model.Since the target nerve network model is server according to multiple
The driving trace training of second vehicle obtains, and multiple second vehicle refers to that current time is passed through before where the target position
Section vehicle.That is, when the first vehicle determines the traveling rail of itself in target location by target nerve network model
When mark, pass through the driving trace of multiple second vehicles in the section where the target position, before with reference to current time with drop
Accident rate of low first vehicle according to the driving trace determined when driving, that is, improving the traveling rail determined
The feasibility of mark.
Optionally, in the neural network model from storage where the determining target position being presently in the first vehicle
The corresponding target nerve network model in section before, further includes:
Determine all second vehicles in preset time period by the section where the target position driving trace and
The scoring of the driving trace of each second vehicle;
Selection scoring is greater than N number of driving trace of default scoring from all driving traces got, and the N is greater than 1
And it is less than or equal to the total quantity of the driving trace got;
The neural network model of initialization is trained by N number of driving trace, obtains the target nerve net
Network model.
Since the first vehicle is that the target nerve network model sent according to server determines driving trace,
In the application, server also needs to predefine the target nerve network model.Further, in order to pass through the target nerve network
The feasibility for the driving trace that model is determined, server can be according to the scorings of each driving trace from multiple driving traces
Select outstanding driving trace, and the outstanding driving trace training objective neural network model by selecting.
Optionally, described that the neural network model of initialization is trained by N number of driving trace, it obtains described
Target nerve network model, comprising:
Determine the traveling task of N number of second vehicle and the driving information of driving information and N number of second obstacle vehicle;
Wherein, N number of second vehicle is the corresponding vehicle of the N number of driving trace, N number of second obstacle vehicle with
N number of second vehicle corresponds, and the second obstacle vehicle be the distance between corresponding second vehicle be less than it is default away from
Vehicle from threshold value, the traveling task include keeping straight on, turn left, turn right and turning around, and the driving information includes being presently in
Position, driving direction and travel speed;
According to the driving information of the traveling task of N number of second vehicle and driving information, N number of second obstacle vehicle
And the driving trace of N number of second vehicle, the neural network model of initialization is trained, the target nerve is obtained
Network model.
Wherein, the process being trained to target nerve network model passes through determination that is, determining sample data later
Sample data out is trained the neural network model of initialization, to obtain the target nerve network model.
Optionally, the section where the target position is crossing;
The traveling task of N number of second vehicle of determination and the driving information of driving information and N number of second obstacle vehicle
Later, further includes:
Determine N number of signal lamp state, N number of signal lamp state and N number of second vehicle correspond, each signal
Lamp state refers to corresponding second vehicle corresponding signal lamp state at crossing when by the crossing;
Correspondingly, the traveling task and driving information according to N number of second vehicle, N number of second obstacle vehicle
Driving information and N number of second vehicle driving trace, the neural network model of initialization is trained, is obtained
The target nerve network model, comprising:
By the driving information of the traveling task of N number of second vehicle and driving information, N number of second obstacle vehicle with
And the input of neural network model of the N number of signal lamp state as the initialization, by the traveling of N number of second vehicle
Output of the track as the neural network model of the initialization, is trained the neural network model of the initialization, obtains
To the target nerve network model.
Further, when the section where target position is crossing, server determines the training target nerve net at this time
The sample data of network model, the sample data can also include that each second vehicle is corresponding at the crossing when by the crossing
Signal lamp state.
Optionally, the scoring of the driving trace of each second vehicle of the determination, comprising:
For any second vehicle in all second vehicles, described is determined according to the driving trace of second vehicle
The driving condition of two vehicles in the process of moving, whether the driving condition includes the number to collide, observe traffic rules and regulations,
Lane change number, traveling duration and whether be smooth ride;
According to the driving condition of second vehicle in the process of moving, commenting for the driving trace of second vehicle is determined
Point.
Wherein, server can by the number that collides of second vehicle when by the crossing, whether abide by
Traffic rules, lane change number, traveling duration and whether be driving trace that the factors such as smooth ride determine second vehicle
Scoring.
Optionally, the driving trace according to second vehicle determines the row of second vehicle in the process of moving
Sail situation, comprising:
Third obstacle vehicle is determined in the driving trace by the crossing, the third obstacle vehicle is described second
It is less than the vehicle of pre-determined distance threshold value when vehicle passes through the crossing at a distance from second vehicle;
According to the driving trace of the driving trace of second vehicle and the third obstacle vehicle, second vehicle is determined
The number to collide between the third obstacle vehicle;
Determine second vehicle by the corresponding signal lamp state in crossing described during the crossing;
According to the driving trace of second vehicle and second vehicle by the crossing during the crossing
The signal lamp state at place, determines whether second vehicle observes traffic rules and regulations;
According to the driving trace of second vehicle, determines the lane change number of second vehicle, travels duration and be
No is smooth ride.
Specifically, server can determine the second vehicle colliding when by the crossing by the above method
Whether number observes traffic rules and regulations, lane change number, traveling duration and whether be smooth ride.
Optionally, the driving condition according to second vehicle in the process of moving, determines second vehicle
The scoring of driving trace, comprising:
If the number that second vehicle collides in the process of moving is greater than or equal to default collision frequency, really
Fixed collision scoring is scored for first, otherwise, it determines collision scoring is the second scoring, wherein collision scores and collides
Number negative correlation;
If second vehicle is observed traffic rules and regulations, it is determined that traffic rules scoring is that third scores, otherwise, it determines institute
Traffic rules scoring is stated as the 4th scoring;
If the lane change number of second vehicle in the process of moving is greater than or equal to described in second vehicle process
Minimum lane change number needed for crossing, it is determined that lane change scoring is the 5th scoring, otherwise, it determines lane change scoring is commented for the 6th
Point, wherein lane change scoring and lane change number negative correlation;
If second vehicle is greater than or equal to preset travel duration by the long when driving of the crossing, it is determined that when
Long scoring is the 7th scoring, otherwise, it determines duration scoring is the 8th scoring, wherein duration scoring is with traveling duration in negative
Correlativity;
If second vehicle is smooth ride, it is determined that drive scoring for the 9th scoring, otherwise, it determines the driving
Scoring is the tenth scoring;
Collision scoring, traffic rules scoring, lane change scoring, duration scoring and the driving are commented
/ the scoring with the driving trace for being determined as second vehicle.
Further, server is determining the number that collides of second vehicle when by the crossing, is being
It is no observe traffic rules and regulations, lane change number, traveling duration and when whether being smooth ride, can determine collision scoring, friendship respectively
Drift then scores, lane change scoring, duration scores and driving scoring, according to collision scoring, the traffic rules scoring, change determined
Road scoring, duration scoring and driving score and determine the scoring of the driving trace of second vehicle.
Optionally, in the neural network model from storage where the determining target position being presently in the first vehicle
The corresponding target nerve network model in section, comprising:
According to the traveling task of the location information of the target position and first vehicle, from the neural network mould of storage
It is determined and the corresponding target nerve network of traveling task in the section and first vehicle where the target position in type
Model;
Correspondingly, the multiple second vehicle refers to the section where passing through the target position before current time and row
Sail task vehicle identical with the traveling task of first vehicle.
Further, at some section, server can train different neural networks for different traveling tasks
Model, at this point, server can send the traveling task with section and the first vehicle where the target position to the first vehicle
Corresponding target nerve network model.
Optionally, all second vehicles in the section where the target position are passed through in the determining preset time period
Driving trace, comprising:
Determine the second vehicle at the first moment by the section where the target position, first moment refers to institute
State any moment in preset time period;
For determining obtained any second vehicle, determine second vehicle by the section where the target position
Driving process in multiple second moment;
The driving trace of second vehicle is determined based on the driving information of the second vehicle described at each second moment.
Wherein, server determines the driving trace of the second vehicle, that is, determining that the second vehicle is passing through the target position institute
Section during each second moment when driving information.
The third aspect provides a kind of device for planning driving trace, is applied to the first vehicle, the planning driving trace
Device has the function of realizing the method that driving trace is planned in above-mentioned first aspect.It is described planning driving trace device include
At least one unit, at least one unit is for realizing the method for planning driving trace provided by above-mentioned first aspect.
Fourth aspect provides the device of another planning driving trace, is applied to server, the planning driving trace
Device has the function of realizing the method that driving trace is planned in above-mentioned second aspect.It is described planning driving trace device include
At least one unit, at least one unit is for realizing the method for planning driving trace provided by above-mentioned second aspect.
5th aspect, provide it is a kind of plan driving trace device, it is described planning driving trace device structure in
Including processor and memory, the memory supports beam size enlargement apparatus to execute provided by above-mentioned first aspect for storing
The program of the method for driving trace, and storage are planned for realizing the side of driving trace is planned provided by above-mentioned first aspect
Data involved in method.The processor is configured to for executing the program stored in the memory.The storage equipment
Operating device can also include communication bus, the communication bus is for establishing connection between the processor and memory.
6th aspect provides the device of another planning driving trace, the structure of the device of the planning driving trace
In include processor and memory, the memory supports beam size enlargement apparatus to execute above-mentioned second aspect and provided for storing
Planning driving trace method program, and storage is for realizing planning driving trace provided by above-mentioned second aspect
Data involved in method.The processor is configured to for executing the program stored in the memory.The storage is set
Standby operating device can also include communication bus, and the communication bus is for establishing connection between the processor and memory.
7th aspect, provides a kind of computer readable storage medium, is stored in the computer readable storage medium
Instruction, when run on a computer, so that computer executes the method for planning driving trace described in above-mentioned first aspect.
Eighth aspect provides another computer readable storage medium, stores in the computer readable storage medium
There is instruction, when run on a computer, so that computer executes the side for planning driving trace described in above-mentioned second aspect
Method.
9th aspect, provides a kind of computer program product comprising instruction, when run on a computer, so that
Computer executes the method that driving trace is planned described in above-mentioned first aspect.
Tenth aspect, provide another computer program product comprising instruction makes when run on a computer
It obtains computer and executes the method for planning driving trace described in above-mentioned second aspect.
In the above-mentioned third aspect, the 5th aspect, the 7th aspect and the 9th aspect technical effect obtained and first aspect
The technical effect that corresponding technological means obtains is approximate, repeats no more herein.Above-mentioned fourth aspect, the 6th aspect, eighth
Face and the tenth aspect technical effect obtained are approximate with the technical effect that technological means corresponding in second aspect obtains, at this
In equally repeat no more.
Technical solution provided by the present application has the benefit that
In this application, when the first vehicle is currently at target position, server can be directly determined and the target position
The corresponding target nerve network model in section where setting, and the target nerve network model is sent to the first vehicle, so that the
One vehicle determines driving trace by the target nerve network model.Since the target nerve network model is server according to more
The driving trace training of a second vehicle obtains, and multiple second vehicle refers to that pass through the first vehicle before current time current
The vehicle in the section where locating target position.That is, when the first vehicle passes through target nerve network mould in target location
When type determines the driving trace of itself, the driving information of first vehicle and the first obstacle vehicle is not only allowed for, reference is also made to
Pass through the driving trace of multiple second vehicles in the section where the target position before current time, is pressed with reducing the first vehicle
According to the accident rate of the driving trace determined when driving, that is, improving the feasibility of the driving trace determined.
Detailed description of the invention
Fig. 1 is a kind of system schematic for planning driving trace provided in an embodiment of the present invention;
Fig. 2 is a kind of server schematic diagram provided in an embodiment of the present invention;
Fig. 3 is a kind of structural schematic diagram of server provided in an embodiment of the present invention;
Fig. 4 is a kind of vehicle schematic diagram provided in an embodiment of the present invention;
Fig. 5 is a kind of structural schematic diagram of vehicle provided in an embodiment of the present invention;
Fig. 6 A is a kind of method flow diagram for planning driving trace provided in an embodiment of the present invention;
Fig. 6 B is a kind of crossing schematic diagram provided in an embodiment of the present invention;
Fig. 7 is the method flow diagram of another planning driving trace provided in an embodiment of the present invention;
Fig. 8 is a kind of device block diagram for planning driving trace provided in an embodiment of the present invention;
Fig. 9 A is a kind of device block diagram for planning driving trace provided in an embodiment of the present invention;
Fig. 9 B is the device block diagram of another planning driving trace provided in an embodiment of the present invention.
Specific embodiment
To keep the purposes, technical schemes and advantages of the application clearer, below in conjunction with attached drawing to the application embodiment party
Formula is described in further detail.
In order to make it easy to understand, doing simple introduction to application scenarios involved in the embodiment of the present invention first.
For the vehicle of automatic Pilot, vehicle is to travel according to the driving trace planned in advance, rather than picture has driver
The vehicle of driving is travelled according to the instruction that driver issues in real time like that, causes the vehicle of automatic Pilot cannot be as there is driver driving
Vehicle is like that in real time adaptively adjusted driving direction or travel speed according to the environment that vehicle is presently in.For example, vehicle
Current traveling task is to turn right, and locating lane position is Through Lane, the driving trace planned at this time in present road
Are as follows: from the center line for being presently in lane, along between the center line for being presently in lane and right-hand lane center line
Connecting line traveling, until reaching the center line of right-hand lane.And it is practical during vehicle is turned right and travelled, vehicle periphery may
There are other vehicles lane change behavior, if continuing to travel according to the driving trace of above-mentioned planning at this time, it is easy to other vehicles
And the safety of itself impacts.Therefore, how to plan that driving trace is just particularly important.And the embodiment of the present invention is just answered
In the scene of driving trace for how to plan vehicle.
After the application scenarios to the embodiment of the present invention are introduced, go below to planning provided in an embodiment of the present invention
The system for sailing track is simply introduced.
As shown in Figure 1, the embodiment of the invention provides a kind of system for planning driving trace, which is
System 100 includes server 101 and at least one vehicle 102, and between server 101 and each vehicle 102 wirelessly
Connection is to be communicated.
Wherein, server 101 is for counting history driving trace, to pass through history driving trace training nerve net
Network model.Vehicle 102 passes through the preparatory training for obtaining the neural network model trained in advance from server 101
Neural network plan driving trace.
Optionally, server 101 is also used to push the mind trained in advance at least one vehicle 102 by broadcast mode
Through network model, vehicle 102 from server 101 without actively obtaining the neural network model trained in advance at this time.
Fig. 2 is a kind of schematic diagram of server 200 provided in an embodiment of the present invention.Referring to fig. 2, which includes
Data acquisition module 201, data memory module 202, task analysis module 203, time locus analysis module 204, signal lamp point
Analyse module 205, track grading module 206, model training module 207 and model pushing module 208.
Wherein, data acquisition module 201 is for acquiring data, and the data of the acquisition can be to be installed on default section
The data that the camera of multiple fixed installations reports are also possible to preset the number that the vehicle in section is reported to server by this
According to.If the data of the acquisition are the data that multiple cameras report, the data of the acquisition should be included at least: the view of acquisition
Frequently, the information such as direction of the video corresponding time, the geographical location of camera, the height of camera and camera.If
The data of the acquisition are the data presetting the vehicle in section by this and reporting, then the data acquired should include at least: the vehicle institute
Position, driving direction, speed and the corresponding temporal information at place.Optionally, the data of the acquisition can also include being mounted on
The video of the vehicle-mounted camera shooting of the vehicle, or the vehicle-mounted software by being mounted on the vehicle handle the video of the shooting
Information later.It wherein, may include signal information and periphery dynamic barrier in the video of vehicle-mounted camera shooting
The information such as position, driving direction and speed.
Data memory module 202 is for storing the data that data acquisition module 201 acquires.
Task analysis module 203 is used to read the data of storage from data memory module 202, and by pre- setting video
Reason technology determines the driving trace of each vehicle, according to the driving trace of each vehicle, analyzes the vehicle at this and presets section
Traveling task.For example, turning left, keeping straight on, turn right or turning around.
Time locus analysis module 204 reads the data of storage from data memory module 202, and by pre- setting video
Reason technology determines that the driving trace of each vehicle, the driving trace of the vehicle are included under certain time sequence, locating for the vehicle
The information such as position, driving direction, speed, that is, the driving trace of the vehicle is to preset section at this by the vehicle driving
The information such as the location of the vehicle, driving direction and speed form when different moments.
Signal lamp analysis module 205 reads the data of storage from data memory module 202, similarly, passes through default view
Frequency processing technique analyzes the information that the signal lamp in the default section changes over time.For example, analyze in time period t 0 to t1,
The straight trip signal lamp in western direction eastwards is green.
Track grading module 206 is for reading above-mentioned task analysis module 203, time locus analysis module 204 and letter
Treated the data of signal lamp analysis module 205, these data are interrelated according to the time, and by pre-determined distance threshold value not
It is considered as obstacle vehicle each other with vehicle, to score all driving traces, the standard of the scoring may include: whether abide by
Traffic rules preset collision frequency between section time, lane change number and obstacle vehicle by this etc..
Model training module 207 is analyzed according to above-mentioned task analysis module 203, time locus analysis module 204, signal lamp
The score data of module 205 and track grading module carries out history driving trace using the neural network model of initialization
Training obtains neural network model, and store so that neural network model study has the feature of the driving trace of higher scoring
Neural network model after training.
Model pushing module 208 is used for when receiving the model request of vehicle, and the neural network model of storage is pushed
To vehicle, which is also used to push the neural network mould to a certain range of vehicle in a broadcast manner
Type.
Fig. 3 is the structural schematic diagram of another server 300 provided in an embodiment of the present invention.The server 300 is for real
The function for the modules that existing server 200 shown in Fig. 2 includes.Specifically, as shown in figure 3, the server 300 includes at least
One processor 301, communication bus 302, memory 303 and at least one communication interface 304.
Processor 301 can be a general central processor (Central Processing Unit, CPU), micro process
Device, application-specific integrated circuit (application-specific integrated circuit, ASIC) or one or more
A integrated circuit executed for controlling application scheme program.
Communication bus 302 may include an access, and information is transmitted between said modules.
Memory 303 can be read-only memory (read-only memory, ROM) or can store static information and instruction
Other types of static storage device, random access memory (random access memory, RAM)) or can store
The other types of dynamic memory of information and instruction, is also possible to Electrically Erasable Programmable Read-Only Memory
(Electrically Erasable Programmable Read-Only Memory, EEPROM), CD-ROM (Compact
Disc Read-Only Memory, CD-ROM) or other optical disc storages, optical disc storage (including compression optical disc, laser disc, light
Dish, Digital Versatile Disc, Blu-ray Disc etc.), magnetic disk storage medium or other magnetic storage apparatus or can be used in carry or
Store have instruction or data structure form desired program code and can by any other medium of computer access, but
It is without being limited thereto.Memory 303, which can be, to be individually present, and is connected by communication bus 302 with processor 301.Memory 303
It can be integrated with processor 301.
Communication interface 304, using the device of any transceiver one kind, for other equipment or communication, such as
Ethernet, wireless access network (RAN), WLAN (Wireless Local Area Networks, WLAN) etc..
In the concrete realization, as one embodiment, processor 301 may include one or more CPU.
In the concrete realization, as one embodiment, server may include multiple processors.It is every in these processors
One can be monokaryon (single-CPU) processor, be also possible to multicore (multi-CPU) processor.Here
Processor can refer to one or more equipment, circuit, and/or the processing for handling data (such as computer program instructions)
Core.For example, for above-mentioned task analysis module 203 shown in Fig. 2, time locus analysis module 204, signal lamp analysis module
205, each module in track grading module 206 and model training module 207, the module can by a processor Lai
Realize the function of the module.
In the concrete realization, as one embodiment, server 300 can also include output equipment and input equipment.It is defeated
Equipment and processor 301 communicate out, can show information in many ways.For example, output equipment can be liquid crystal display
(liquid crystal display, LCD), Light-Emitting Diode (light emitting diode, LED) show equipment, yin
Extreme ray pipe (cathode ray tube, CRT) shows equipment or projector (projector) etc..Input equipment and processor
301 communications, can receive the input of user in many ways.For example, input equipment can be mouse, keyboard, touch panel device
Or sensing equipment etc..
In the concrete realization, server can be desktop computer, portable computer, network server, palm PC
(Personal Digital Assistant, PDA), cell phone, tablet computer, wireless terminal device, communication equipment or
Embedded device.The embodiment of the present invention does not limit the type of server.
Wherein, memory 303 is used to store the program code for executing application scheme, and is held by processor 301 to control
Row, to realize according to history driving trace training neural network model.Processor 301 is used to execute to store in memory 303
Program code.It may include one or more software modules in program code.
Fig. 4 be a kind of schematic diagram of vehicle 400 provided in an embodiment of the present invention, the vehicle 400 include locating module 401,
Model request module 402, vehicle-carrying communication module 403, mission planning module 404, sensing module 405, trajectory computation module 406 with
And vehicle control module 407.
Wherein, the location information that acquisition vehicle is presently in real time of locating module 401, and will be collected with certain frequency
Location information be sent to model request module 402.
Model request module 402 is used to request neural network model to server by vehicle-carrying communication module 403, that is,
Model request module 402 is used to pass through vehicle-carrying communication module 403 to server transmission pattern acquisition request, the model acquisition request
The target position being presently in including vehicle.It is returned when model request module 402 receives server by vehicle-carrying communication module 403
When the neural network model returned, the neural network model is stored.
Optionally, when the model pushing module 208 in above-mentioned server shown in Fig. 2 by way of broadcast to certain model
When vehicle in enclosing pushes neural network model, the model request module 402 by vehicle-carrying communication module 403 for directly being received
The neural network model of server push, and store the neural network model.
In addition, the model request module 402 is also used to root when corresponding to different neural network models due to different sections
Check locally whether there is the corresponding neural network model of current driving road segment according to collected location information, when discovery is local not
The neural network model is obtained there are when the neural network model, then by server.
Mission planning module 404 is used to determine the traveling task of vehicle.Sensing module 405 is used to determine the traveling letter of vehicle
The information such as breath, the driving information of obstacle vehicle and signal lamp state.Wherein, driving information includes the position being presently in, row
Sail the information such as direction and travel speed.
Trajectory computation module 406 is used for the data determined according to mission planning module 404 and sensing module 405, passes through mould
The neural network model that type request module 402 obtains, determines the driving trace of vehicle.Vehicle control module 407 is for controlling vehicle
According to trajectory computation module 406 determine driving trace travel.
Fig. 5 is the structural schematic diagram of another vehicle 500 provided in an embodiment of the present invention, and the vehicle 500 is for realizing upper
State the function for the modules that vehicle 400 shown in Fig. 4 includes.Specifically, as shown in figure 5, the vehicle 500 includes at least one
Processor 501, communication bus 502, memory 503 and at least one communication interface 504.
Wherein, the structure and function of processor 501 and processor shown in Fig. 3 301 are essentially identical, communication bus 502
It is essentially identical with the structure and function of communication bus 502 shown in Fig. 3, memory 503 and memory shown in Fig. 3 503
Structure and function are essentially identical, and the structure and function of communication interface 504 and communication interface shown in Fig. 3 504 are essentially identical,
It no longer elaborates herein.,
The difference is that memory 503 is used to store the program code for executing application scheme, and by processor 501
It is executed to control, driving trace is planned as vehicle according to trained neural network model to realize.
The embodiment of the present invention is described in further detail below in conjunction with attached drawing.
Fig. 6 A is a kind of method flow diagram for planning driving trace provided in an embodiment of the present invention, and this method is applied to Fig. 1
Shown in system, in order to subsequent convenient for explanation, the vehicle for currently needing to plan driving trace is known as the first vehicle.Such as Fig. 6 A
Shown, this method comprises the following steps:
Step 601: server determining target position being presently in the first vehicle from the neural network model of storage
The corresponding target nerve network model in the section at place, the target nerve network model are servers according to multiple second vehicles
Driving trace training obtains, and multiple second vehicle refers to the vehicle for passing through the section where the target position before current time
?.
Since in practical application, the geography of different sections of highway has a long way to go, such as the path formation degree of different sections of highway
And the fixed obstacle on road is likely to difference, therefore, in embodiments of the present invention, server can be in advance for not
The different neural network model of same section training.Since different sections corresponds to different neural network models, when
When one vehicle is currently at target position, server can determine target nerve net corresponding with the section where the target position
Network model.
Particularly, since the geography at crossing is complex, the first vehicle by the accident rate at crossing compared with
Height, and the geography in the section at other non-crossings is then more relatively easy, therefore, in embodiments of the present invention, server can
With the neural network model different only for different crossing training.That is, when the section where target position is crossing,
The target nerve network model is neural network model corresponding with the crossing at this time.
It should be noted that in embodiments of the present invention, which can be by haing point at two different intersections
Branch crossing, or the either branch crossing at this two different intersection Chu Suoyou branch crossings.For example, such as Fig. 6 B institute
Show, in one crossroad of infall formation, which wraps for four respectively in all directions towards upper for road 1 and road 2
Include four different branch crossings.
Wherein, when the crossing in the embodiment of the present invention is the crossroad, server is according to by the cross at this time
The history driving trace at all branch crossings that crossing includes determines neural network model corresponding with the crossing.When of the invention real
Crossing in example is applied when can be for the branch crossing of any direction in the crossroad, at this time server be according to by this ten
The history driving trace at branch crossing determines neural network model corresponding with the branch crossing, that is, on the crossroad,
It can be for each towards upper branch crossing, 4 different neural network models of training.
In addition, for the same section, gap of the different traveling tasks at the section between corresponding driving trace
It is larger.And during training neural network, it is to need to be divided for a kind of training sample with some unified rule
Therefore analysis in embodiments of the present invention, the rule of data in learning training sample is capable of for the ease of neural network model,
For the same section, different neural network models can be respectively trained according to traveling task.Namely at the section, for
One neural network model of each traveling task training.
At this point, server is it is confirmed that with the first vehicle driving task and mesh when the first vehicle need to determine driving trace
The corresponding target nerve network model in section where cursor position.Correspondingly, server is in the row according to multiple second vehicles
When sailing track training target nerve network model, multiple second vehicle, which refers to, passes through the target position institute before current time
Section and identical with the traveling task of first vehicle vehicle of traveling task.
Wherein, server obtains the realization side of target nerve network model according to the training of the driving trace of multiple second vehicles
Formula will be described in detail in the following embodiments, not illustrate first herein.
Step 602: server sends the target nerve network model to the first vehicle.
Server sends the target to the first vehicle after determining the target nerve network model by step 601
Neural network model, in order to which the first vehicle is determined by following step 603 and step 604 according to the target nerve network model
The driving trace of first vehicle.
Step 603: the first vehicle receives the target nerve network model that server is sent.
It, can be true by following step 604 when the first vehicle receives the target nerve network model of server transmission
The driving trace of fixed first vehicle.
Step 604: the first vehicle is according to the traveling task and driving information of the first vehicle and the row of the first obstacle vehicle
Information is sailed, the driving trace of the first vehicle is determined by the target nerve network model, wherein traveling task includes straight trip, a left side
Turn, turn right and turn around, driving information includes the position being presently in, driving direction and travel speed, the first obstacle vehicle be with
The distance between first vehicle is less than the vehicle of pre-determined distance threshold value.
In embodiments of the present invention, server can be according within a preset period of time by the section where the target position
All second vehicles driving trace, training initialization neural network model, obtain and the section where the target position
Corresponding target nerve network model, in order to which the first vehicle determines driving trace according to the target nerve network model.That is,
When the first vehicle is when target location determines the driving trace of itself by target nerve network model, not only allow for this
The driving information of one vehicle and the first obstacle vehicle reference is also made to pass through the section where the target position before current time
The driving trace of multiple second vehicles, to reduce accident rate of first vehicle according to the driving trace determined when driving,
That is, improving the feasibility of the driving trace determined.
Fig. 7 is the method flow diagram of another planning driving trace provided in an embodiment of the present invention, and this method is applied to Fig. 1
Shown in system.Wherein, embodiment shown in Fig. 7 is the further description to embodiment shown in Fig. 6 A, such as Fig. 7 institute
Show, method includes the following steps:
Step 701: server determines in preset time period by all second vehicles in the section where the target position
Driving trace, target position is the position that is presently in of the first vehicle.
In embodiments of the present invention, since target nerve network model is traveling rail of the server according to multiple second vehicles
Mark training obtains, therefore server is before send the target nerve network model to the first vehicle, also needs to multiple the
The driving trace of two vehicles is trained, to determine the target nerve network model.Specifically, step 701 to step can be passed through
Rapid 705 determine the target nerve network model.
In addition, capableing of the rule of data in learning training sample for the ease of neural network model, for the same road
Section, can be respectively trained different neural network models according to traveling task.Due to the area of the different neural network model of training
It is not the difference of training sample, it is therefore, in embodiments of the present invention, corresponding with the section where the target position with training
It is illustrated for neural network model, the training process of other kinds of neural network model is no longer described in detail.
It is worth noting that, the key of training neural network is to determine training sample, when needs are trained and the target position
It, need to be according to passing through the section where the target position before current time when the corresponding neural network model in section where setting
Driving trace determines training sample.
In addition, the possible quantity of the driving trace in the section as where passing through the target position before current time is more,
And the geography in the section before current time where the target position is it can also happen that variation, namely apart from current time
The data of the driving trace occurred when longer may be no longer valid, therefore, only need to travel rail according to the history in preset time period
Mark trains neural network model.
For example, the preset time period can be 1 month before current time, at this time only need to be according to current time before
Neural network model is determined by the driving trace of second vehicle in the section where the target position in 1 month.
Specifically, it is determined that second vehicle in the section where passing through the target position at the first moment, first moment are
Refer to any moment in the preset time period, for determining obtained any second vehicle, determines second vehicle by the mesh
Multiple second moment in the driving process in the section where cursor position, the row based on second vehicle at each second moment
Sail the driving trace that information determines second vehicle.
In one possible implementation, pass through the road where the target position in the preset time period in order to determining
All driving traces of section, can will mark off multiple first moment in the preset time period, determine each first moment warp
Second vehicle in the section where the target position is crossed, all first moment pass through second vehicle in the section where the target position
It is within a preset period of time by all vehicles in the section where the target position.
Wherein, the driving trace of each second vehicle is the row by second vehicle by each second moment at the crossing
Sail information composition, the driving information include second vehicle be presently at second moment position, driving direction and
The information such as travel speed.
Specifically, by data memory module shown in Fig. 2 determine data acquisition module within a preset period of time at this
The video acquired at section where target position, and each second vehicle is determined by time locus shown in Fig. 2 analysis
Driving trace.
In embodiments of the present invention, feasible driving trace is obtained by neural network model in order to realize, Ying Caiyong is gone through
More outstanding driving trace training neural network model in history driving trace.That is, determining in preset time period by being somebody's turn to do
After the driving trace of all second vehicles in the section where target position, also need to determine wherein outstanding driving trace.
Wherein, when the section where the target position is crossing, service implement body can pass through following step 702 and step
Rapid 703 determine the scoring of each driving trace, in order to being determined by following step 704 according to the scoring of each driving trace
Wherein outstanding driving trace.
Optionally, when the section where the target position is other kinds of section, following step can equally be referred to
702 determine wherein outstanding driving trace to step 704, no longer elaborate herein.
Step 702: server is for any second vehicle in all second vehicles, according to the traveling rail of second vehicle
Mark determines that the driving condition of the second vehicle in the process of moving, the driving condition include the number to collide, whether abide by
Traffic rules, lane change number, traveling duration and whether be smooth ride.
Wherein it is determined that the number that collides is specifically as follows: determining traveling of third obstacle vehicle when by the crossing
Track, the third obstacle vehicle are the vehicle for being less than pre-determined distance threshold value at a distance from second vehicle.According to second vehicle
Driving trace and the third obstacle vehicle driving trace, determine and touched between second vehicle and the third obstacle vehicle
The number hit.
Since the driving trace of the second vehicle is to be believed by second vehicle by the traveling at each second moment at the crossing
Breath composition, therefore, the driving trace of third obstacle vehicle are also to be believed by the traveling at each second moment of the third obstacle vehicle
Breath composition, that is, being corresponded at the time of in the driving trace of the driving trace of the second vehicle and third obstacle vehicle.At this point,
Determine that the implementation of the number to collide between second vehicle and the third obstacle vehicle can be with are as follows: for each second
Moment determines the position that the second vehicle is presently in and the position that third obstacle vehicle is presently in, if the current institute of the second vehicle
The distance between position that the position at place and third obstacle vehicle are presently in is less than default collision distance, it is determined that this second
Moment collides between the second vehicle and the third obstacle vehicle.
Wherein, presetting collision distance is pre-set distance, the default collision distance can for 0.1m, 0.15m or
0.2m etc..
It is specifically as follows secondly, determining whether the second vehicle observes traffic rules and regulations: determines second vehicle by the crossing
During the corresponding signal lamp state in the crossing, which is passed through according to the driving trace of second vehicle and second vehicle
During corresponding signal lamp state at the crossing, determine whether second vehicle observes traffic rules and regulations.
That is, being determined at each second moment crossing in the second vehicle during travelling at each second moment
Corresponding signal lamp state, for each second moment, at second moment, if with corresponding signal lamp state at the crossing
For red light, then show that second vehicle is not observed traffic rules and regulations at this time, if being green with corresponding signal lamp state at the crossing
Lamp then shows that second vehicle is observed traffic rules and regulations at this time.
In addition, determining the lane change number of second vehicle according to the driving trace of second vehicle, travelling duration and be
No is smooth ride.
Wherein, in order to make the traveling duration that can pass through the duration at the crossing with accurate description second vehicle, traveling is determined
Duration is specifically as follows: determining that the second vehicle passes through the time at the crossing, according to corresponding signal lamp state in the time, determines
Duration that no through traffic when signal lamp is red light, subtracts this in the time from the second vehicle by the crossing and forbids communication time,
The obtained time is the traveling duration.
Determine whether the second vehicle is that smooth ride is specifically as follows: determining the traveling of each second moment, second vehicle
Speed carries out variance calculating to the travel speed of each second moment, second vehicle, obtains velocity variance.Since variance can be with
The dispersion degree of one group of data is described, therefore, if the velocity variance is greater than default variance, shows that the speed of the second vehicle is unstable
Fixed namely the second vehicle is not smooth ride, if the velocity variance is less than default variance, shows the velocity-stabilization of the second vehicle,
Namely second vehicle be smooth ride.
Step 703: server determines the row of second vehicle according to the driving condition of the second vehicle in the process of moving
Sail the scoring of track.
Outstanding driving trace in driving trace in order to determine all second vehicles, can be to the row of each second vehicle
Track is sailed to score.Wherein, the standard of scoring is the driving condition of the second vehicle in the process of moving.
Specifically, if the number that second vehicle collides in the process of moving is greater than or equal to default collision time
Number, it is determined that collision scoring is scored for first, otherwise, it determines collision scoring is the second scoring, wherein collision is scored and occurred
The number negative correlation of collision.
If second vehicle is observed traffic rules and regulations, it is determined that traffic rules scoring is that third scores, otherwise, it determines the friendship
Drift then scores as the 4th scoring.
If the lane change number of second vehicle in the process of moving is greater than or equal to second vehicle and passes through the crossing institute
The minimum lane change number needed, it is determined that lane change scoring is the 5th scoring, otherwise, it determines lane change scoring is the 6th scoring, wherein
Lane change scoring and lane change number negative correlation.
If second vehicle is greater than or equal to preset travel duration by the long when driving of the crossing, it is determined that duration is commented
It is divided into the 7th scoring, otherwise, it determines duration scoring is the 8th scoring, wherein duration scoring and the traveling negatively correlated pass of duration
System.
If second vehicle is smooth ride, it is determined that drive scoring for the 9th scoring, otherwise, it determines the driving is scored
For the tenth scoring.
The sum of collision scoring, traffic rules scoring, lane change scoring, duration scoring and driving scoring are determined
For the scoring of the driving trace of second vehicle.
Wherein, first scoring, second scoring, third scoring, the 4th scoring, the 5th scoring, the 6th scoring, the 7th scoring,
8th scoring, the 9th scoring and the tenth scoring are the score of default setting, and the score of the default setting can be any score, only
Above-mentioned condition need to be met.
For example, the second scoring, third scoring, the 6th scoring, the 8th scoring and the 9th scoring are set as+5 points, first is commented
Point, the 4th scoring, the 5th scoring and the tenth scoring be set as -5 points.If second vehicle collides in the process of moving
Number be greater than or equal to default collision frequency, it is determined that collision scoring is+5 points, that is, the traveling rail of second vehicle at this time
The scoring of mark will add 5 points.If second vehicle is not observed traffic rules and regulations, it is determined that traffic rules scoring is -5 points, that is,
The scoring of the driving trace of second vehicle at this time will subtract 5 point.
That is, in embodiments of the present invention, can determine in preset time period and pass through with through the above steps 701 to step 703
Cross the scoring of the driving trace and each driving trace of all second vehicles in the section where the target position.
For example, will be marked off in preset time period multiple first moment labeled as t0, t1, t2 ..., tm, from first the
One moment t0 starts, and determines that the first moment t0 passes through second vehicle at the crossing and the driving trace of the second vehicle, and lead to
It crosses above-mentioned steps 802 and step 803 determines the scoring of the driving trace.Continue next moment t1, repeat the above process, determines
The first moment t1 is scored by the driving trace of second vehicle at the crossing ..., and so on, until determining the last one
The driving trace scoring that first moment tm passes through second vehicle at the crossing.
Step 704: server selects scoring to be greater than the default N number of traveling rail to score from all driving traces got
Mark, N are greater than 1 and are less than or equal to the total quantity of the driving trace got.
Since the scoring of driving trace is determined according to the driving condition of the second vehicle in the process of moving, traveling
The scoring of track can be used for describing the outstanding degree of the driving trace and show the row that is, the scoring of the driving trace is higher
It is more outstanding to sail track.Therefore, pass through driving trace more outstanding in the available history driving trace of step 704.
Wherein, presetting scoring is pre-set scoring, which can be 80 points, 90 points or 95 points etc..
Step 705: server is trained by neural network model of N number of driving trace to initialization, obtains target
Neural network model.
For the training sample of the neural network model in the embodiment of the present invention, the training sample include multiple independents variable and
With multiple independents variable multiple dependent variables correspondingly, for ease of description, by multiple independents variable labeled as x1, x2 ..., xn,
Will with multiple independents variable correspondingly multiple dependent variables labeled as y1, y2 ..., yn.Training neural network model, that is, making
The neural network model of initialization learns multiple independent variable and between multiple independent variable correspondingly multiple dependent variables
Mapping relations obtain y=f (x), which is neural network model after training.
Therefore, above-mentioned steps 705 are specifically as follows, and determine the traveling task and driving information of N number of second vehicle, and
The driving information of N number of third obstacle vehicle, N number of second vehicle are the corresponding vehicle of the N number of driving trace, N number of second barrier
Vehicle and N number of second vehicle is hindered to correspond, and the second obstacle vehicle is that the distance between corresponding second vehicle is less than
The vehicle of pre-determined distance threshold value.
The driving information of the traveling task of N number of second vehicle and driving information and N number of second obstacle vehicle is made
For the input of the neural network model of the initialization, using the driving trace of N number of second vehicle as the nerve net of the initialization
The output of network model is trained the neural network model of the initialization, obtains corresponding with the section where the target position
Neural network model.
Further, when the section where the target position is crossing, above-mentioned steps 705 are specifically as follows: determining the N
The traveling task and driving information of a second vehicle, at the crossing with N number of second vehicle N number of signal lamp shape correspondingly
The driving information of state and N number of second obstacle vehicle
By the traveling task of N number of second vehicle and driving information, N number of signal lamp state and N number of second obstacle
Input of the driving information of vehicle as the neural network model of the initialization, using the driving trace of N number of second vehicle as
The output of the neural network model of the initialization is trained the neural network model of the initialization, obtains and the crossing pair
The neural network model answered.
That is, by the traveling task of N number of third vehicle and driving information, the crossing with N number of second vehicle one by one
The driving information of corresponding signal lamp state and N number of second obstacle vehicle is as the independent variable in training sample, by the N
Dependent variable of the driving trace of a second vehicle as training sample, to determine the mapping relations y between independent variable and dependent variable
=f (x), that is, determining neural network model.
Since the data of training sample are the traveling rails from all second vehicles for passing through the crossing within a preset period of time
It is determined in mark, therefore obtained neural network model is neural network model corresponding with the crossing.
It should be noted that server is being determined as neural network model corresponding with the section where the target position
Later, it can store the neural network model, that is, being stored with where neural network model and the target position in the server
Section between corresponding relationship, that is, the corresponding neural network model in section.
Later, server can be sent and first vehicle according to the model acquisition request that the first vehicle is sent to the first vehicle
The corresponding target nerve network model in the target position being presently in.Alternatively, server is in a broadcast manner to by the road
The vehicle of section pushes the corresponding neural network model in the section.
As above-mentioned steps 701 to step 705 it is found that in order to obtain with the target position where the corresponding nerve net in section
Network model can determine training sample by the driving trace in the section where the target position according in the preset time period.
It therefore,, can be according to by the road if expecting neural network model corresponding with some traveling task for some section
Section and traveling task are that the driving trace of the traveling task determines training sample, after obtaining training sample, pass through training sample
This trains the process of corresponding neural network model then essentially identical with the process of above-mentioned trained neural network model, and the present invention is real
Example is applied no longer to be described in detail herein.
That is, server can be different for different section training in advance through the above steps 701 to step 705
Neural network model passes through the corresponding target nerve network in section where the target position being presently in order to the first vehicle
Model determines driving trace.Specifically, the driving trace of the first vehicle can be determined by following step 706 to step 707.
Step 706: server determining target position being presently in the first vehicle from the neural network model of storage
The corresponding target nerve network model in the section at place.
In one possible implementation, when server receives the model acquisition request of the first vehicle transmission, root
According to the target position that the first vehicle carried in the model acquisition request is presently in, from the corresponding nerve of the different sections of highway of storage
In network model determine the target position where the corresponding target nerve network model in section, and by following step 707 to
First vehicle sends the target nerve network model.
In alternatively possible implementation, for the section where the target position, difference of the server from storage
The corresponding target nerve network model in section where the target position is determined in the corresponding neural network model in section, and is passed through
Vehicle of the following step 707 in a broadcast manner in the range of where the section pushes the target nerve network model.
Correspondingly, the first vehicle for being currently at the target position can directly receive the target nerve network model.
Step 707: server sends the target nerve network model to the first vehicle, so that the first vehicle passes through the target
Neural network model determines the driving trace of first vehicle.
Wherein, the first vehicle specifically can be true according to the target nerve network model by following step 708 and step 709
Determine driving trace.
Step 708: the first vehicle receives the target nerve network model that server is sent.
By step 707 it is found that in embodiments of the present invention, the first vehicle can be by two different implementations from clothes
The target nerve network model is received at business device.It no longer elaborates herein.
Step 709: the first vehicle is according to the traveling task and driving information of the first vehicle and the row of the first obstacle vehicle
Information is sailed, the driving trace of the first vehicle is determined by the target nerve network model, wherein traveling task includes straight trip, a left side
Turn, turn right and turn around, driving information includes the position being presently in, driving direction and travel speed, the first obstacle vehicle be with
The distance between first vehicle is less than the vehicle of pre-determined distance threshold value.
Specifically, step 709 can be realized by following two step:
(1) the traveling letter of the traveling task of the first vehicle, the driving information of the first vehicle and the first obstacle vehicle is obtained
Breath.
Specifically, the first vehicle can determine itself current traveling task by mission planning module shown in Fig. 4, obtain
To the traveling task of the first vehicle.The position itself being presently in is determined by locating module, and is determined certainly by sensing module
The letter such as position, driving direction and travel speed that the driving direction and travel speed of body and the first obstacle vehicle are presently in
Breath, obtains the driving information of the first vehicle and the first obstacle vehicle.
Wherein, mission planning module determines current driving according to predetermined guidance path and the position being presently in
Task is straight trip, turns left, turns right and still turn around.Locating module can by GPS (Global Positioning System, entirely
Ball positioning system) technology determines the position that the first vehicle is presently in.Sensing module can be by being installed on taking the photograph for the first vehicle
As video that head acquires determines the driving information of the driving direction of the first vehicle, travel speed and the first obstacle vehicle.
In addition, pre-determined distance threshold value is pre-set distance, when the distance between two vehicles are less than the pre-determined distance
When threshold value, the travel situations of one of vehicle may impact the travel situations of another vehicle, at this time the two
Vehicle obstacle vehicle each other.The pre-determined distance threshold value can be 1 meter, 0.75 meter or 0.5 meter etc..
Optionally, when the section where the target position that the first vehicle is presently in is crossing, in order to reduce by the first vehicle
Occur the probability of traffic accident when driving at the crossing, when determining the driving trace of the first vehicle, also needs to consider the road
Signal lamp state corresponding with the target position at mouthful.
Therefore, the first vehicle is obtaining the traveling task of the first vehicle, the driving information and the first obstacle of the first vehicle
After the driving information of vehicle, also need to obtain signal lamp state corresponding with the target position at the crossing.
(2) according to the traveling task and driving information of the first vehicle and the driving information of the first obstacle vehicle, by this
Target nerve network model determines the driving trace of the first vehicle.
Specifically, the first vehicle can be by the traveling task of the first vehicle, the driving information of the first vehicle, the first obstacle vehicle
Input of the driving information as the neural network model, the first vehicle is determined by neural network model corresponding with the section
Driving trace.
Since the target nerve network model is according to the more of the section where passing through the target position before current time
What the driving trace training of a second vehicle obtained, therefore, which has learnt to history driving trace
Feature, so, when by the driving information of the driving information of the traveling task of the first vehicle, the first vehicle and the first obstacle vehicle
When input as the target nerve network model, which can be travelled by the history learnt
The feature of track determines the driving trace of the first vehicle.
Further, when the section where the target position that the first vehicle is presently in is crossing, at this point, the first vehicle
It can be by the traveling task of the first vehicle, the driving information of the first vehicle, the driving information of the first obstacle vehicle and the crossing
Locate input of the signal lamp state corresponding with the target position as the neural network model, passes through nerve corresponding with the section
Network model determines the driving trace of the first vehicle.
Wherein, neural network model corresponding with the section where the target position is server previously according to by the road
The history driving trace training of the history vehicle of section obtains.Server training nerve corresponding with the section where the target position
Network model will be discussed in detail in the following embodiments, not illustrate first herein.
In embodiments of the present invention, server can be according within a preset period of time by the section where the target position
All second vehicles driving trace, training initialization neural network model, obtain and the section where the target position
Corresponding target nerve network model, in order to which the first vehicle determines driving trace according to the target nerve network model.That is,
When the first vehicle is when target location determines the driving trace of itself by target nerve network model, not only allow for this
The driving information of one vehicle and the first obstacle vehicle reference is also made to pass through the section where the target position before current time
The driving trace of multiple second vehicles, to reduce accident rate of first vehicle according to the driving trace determined when driving,
That is, improving the feasibility of the driving trace determined.
In addition to providing the method for planning driving trace described in above-described embodiment, the embodiment of the invention also provides planning to go
The device of track is sailed, following embodiments will introduce this expansion.
Fig. 8 is a kind of device 800 for planning driving trace provided in an embodiment of the present invention, is applied to vehicle shown in FIG. 1
In, as shown in figure 8, the device 800 includes receiving unit 801 and the first determination unit 802:
Receiving unit 801, for executing in step 603 or embodiment shown in Fig. 7 in embodiment shown in Fig. 6 A
Step 708;
First determination unit 802, for executing step 604 or embodiment shown in Fig. 7 in embodiment shown in Fig. 6 A
In step 709;
Wherein, traveling task includes keeping straight on, turn left, turn right and turning around, and driving information includes the position for the position being presently in
Confidence breath, driving direction and travel speed, the first obstacle vehicle are that the distance between first vehicle is less than pre-determined distance threshold value
Vehicle.
Optionally, the section where the target position is crossing;
First determination unit 802, is specifically used for:
By at the traveling task and driving information of the first vehicle, the driving information of the first obstacle vehicle and the crossing with should
Input of the corresponding signal lamp state in target position as the target nerve network model, it is true by the target nerve network model
The driving trace of fixed first vehicle.
Optionally, which refers to appoints with the traveling in section and the first vehicle where the target position
The corresponding neural network model of business, and multiple second vehicle refers to that current time passes through the road where the target position before
Section and traveling task vehicle identical with the traveling task of first vehicle.
In embodiments of the present invention, server can be according within a preset period of time by the section where the target position
All second vehicles driving trace, training initialization neural network model, obtain and the section where the target position
Corresponding target nerve network model, in order to which the first vehicle determines driving trace according to the target nerve network model.That is,
When the first vehicle is when target location determines the driving trace of itself by target nerve network model, not only allow for this
The driving information of one vehicle and the first obstacle vehicle reference is also made to pass through the section where the target position before current time
The driving trace of multiple second vehicles, to reduce accident rate of first vehicle according to the driving trace determined when driving,
That is, improving the feasibility of the driving trace determined.
It should be understood that the device of planning driving trace provided by the above embodiment is in the traveling rail for planning the first vehicle
When mark, only the example of the division of the above functional modules, in practical application, it can according to need and by above-mentioned function
Distribution is completed by different functional modules, i.e., the internal structure of the first vehicle is divided into different functional modules, with complete with
The all or part of function of upper description.In addition, the device and above-described embodiment of planning driving trace provided by the above embodiment
In the method for planning driving trace belong to same design, specific implementation process is detailed in embodiment of the method, and which is not described herein again.
Fig. 9 A is the device 900 of another planning driving trace provided in an embodiment of the present invention, is applied to clothes shown in FIG. 1
It is engaged in device, as shown in Figure 9 A, which includes the second determination unit 901 and transmission unit 902:
Second determination unit 901, for executing step 601 or embodiment shown in Fig. 7 in embodiment shown in Fig. 6 A
In step 706;
Transmission unit 902, for executing in step 602 or embodiment shown in Fig. 7 in embodiment shown in Fig. 6 A
Step 707.
Optionally, referring to Fig. 9 B, which further includes third determination unit 903, selecting unit 904 and training unit
905:
Third determination unit 903, for executing the step 701 in embodiment shown in Fig. 7 to step 703;
Selecting unit 904, for executing the step 704 in embodiment shown in Fig. 7;
Training unit 905, for executing the step 705 in embodiment shown in Fig. 7.
Optionally, which includes the first determining subelement and training subelement:
First determine subelement, for determine N number of second vehicle traveling task and driving information and it is N number of second barrier
Hinder the driving information of vehicle;
Wherein, which is the corresponding vehicle of the N number of driving trace, and N number of second obstacle vehicle is N number of with this
Second vehicle corresponds, and the second obstacle vehicle is that the distance between corresponding second vehicle is less than pre-determined distance threshold value
Vehicle, the traveling task include keeping straight on, turn left, turn right and turning around, which includes the position being presently in, driving direction
And travel speed;
Training subelement, for traveling task and driving information according to N number of second vehicle, N number of second obstacle vehicle
Driving information and N number of second vehicle driving trace, the neural network model of initialization is trained, is somebody's turn to do
Target nerve network model.
Optionally, the section where the target position is crossing;
The training unit 905 further includes the second determining subelement:
Second determines subelement, for determining N number of signal lamp state, N number of signal lamp state and N number of second vehicle one
One is corresponding, and each signal lamp state refers to corresponding second vehicle corresponding signal lamp shape at crossing when by the crossing
State;
Correspondingly, the training subelement, is specifically used for:
By the traveling task of N number of second vehicle and driving information, N number of second obstacle vehicle driving information and should
Input of N number of signal lamp state as the neural network model of the initialization, using the driving trace of N number of second vehicle as this
The output of the neural network model of initialization is trained the neural network model of the initialization, obtains the target nerve net
Network model.
Optionally, third determination unit 903, specifically for executing step 702 and step in embodiment shown in Fig. 7
703。
Optionally, second determination unit 901, is specifically used for:
According to the traveling task of the location information of the target position and first vehicle, from the neural network model of storage
Determine target nerve network model corresponding with the traveling task in section and first vehicle where the target position;
Correspondingly, multiple second vehicle refers to the section where passing through the target position before current time and traveling is appointed
Business vehicle identical with the traveling task of first vehicle.
In embodiments of the present invention, server can be according within a preset period of time by the section where the target position
All second vehicles driving trace, training initialization neural network model, obtain and the section where the target position
Corresponding target nerve network model, in order to which the first vehicle determines driving trace according to the target nerve network model.That is,
When the first vehicle is when target location determines the driving trace of itself by target nerve network model, not only allow for this
The driving information of one vehicle and the first obstacle vehicle reference is also made to pass through the section where the target position before current time
The driving trace of multiple second vehicles, to reduce accident rate of first vehicle according to the driving trace determined when driving,
That is, improving the feasibility of the driving trace determined.
It should be understood that the device of planning driving trace provided by the above embodiment is in the traveling rail for planning the first vehicle
When mark, only the example of the division of the above functional modules, in practical application, it can according to need and by above-mentioned function
Distribution is completed by different functional modules, i.e., the internal structure of server is divided into different functional modules, more than completing
The all or part of function of description.In addition, in the device and above-described embodiment of planning driving trace provided by the above embodiment
The method of planning driving trace belong to same design, specific implementation process is detailed in embodiment of the method, and which is not described herein again.
In the above-described embodiments, can come wholly or partly by software, hardware, firmware or its any combination real
It is existing.When implemented in software, it can entirely or partly realize in the form of a computer program product.The computer program produces
Product include one or more computer instructions.When loading and execute on computers the computer instruction, all or part of real estate
Life is according to process or function described in the embodiment of the present invention.The computer can be general purpose computer, special purpose computer, calculating
Machine network or other programmable devices.The computer instruction may be stored in a computer readable storage medium, Huo Zhecong
One computer readable storage medium is transmitted to another computer readable storage medium, for example, the computer instruction can be with
From a web-site, computer, server or data center by it is wired (such as: coaxial cable, optical fiber, digital subscriber line
(Digital Subscriber Line, DSL)) or wireless (such as: infrared, wireless, microwave etc.) mode to another website station
Point, computer, server or data center are transmitted.The computer readable storage medium can be computer and can access
Any usable medium either include that the data storages such as one or more usable mediums integrated server, data center are set
It is standby.The usable medium can be magnetic medium (such as: floppy disk, hard disk, tape), optical medium (such as: digital versatile disc
(Digital Versatile Disc, DVD)) or semiconductor medium (such as: solid state hard disk (Solid State Disk,
SSD)) etc..
Those of ordinary skill in the art will appreciate that realizing that all or part of the steps of above-described embodiment can pass through hardware
It completes, relevant hardware can also be instructed to complete by program, the program can store in a kind of computer-readable
In storage medium, storage medium mentioned above can be read-only memory, disk or CD etc..
The above is embodiment provided by the present application, all in spirit herein and original not to limit the application
Within then, any modification, equivalent replacement, improvement and so on be should be included within the scope of protection of this application.
Claims (18)
1. a kind of method for planning driving trace, is applied to the first vehicle, which is characterized in that the described method includes:
The target nerve network model that server is sent is received, the target nerve network model refers to works as with first vehicle
The corresponding neural network model in section where preceding locating target position, and the target nerve network model is the service
Device is obtained according to the training of the driving trace of multiple second vehicles, and the multiple second vehicle, which refers to, passes through institute before current time
State the vehicle in the section where target position;
According to the traveling task of first vehicle and driving information and the driving information of the first obstacle vehicle, by described
Target nerve network model determines the driving trace of first vehicle;
Wherein, the traveling task includes keeping straight on, turn left, turn right and turning around, and the driving information includes the position being presently in
Location information, driving direction and travel speed, the first obstacle vehicle is to be less than with the distance between described first vehicle
The vehicle of pre-determined distance threshold value.
2. the method according to claim 1, wherein the section where the target position is crossing;
It is described according to the traveling task and driving information of first vehicle and the driving information of the first obstacle vehicle, pass through
The target nerve network model determines the driving trace of first vehicle, comprising:
By the traveling task of first vehicle and driving information, the driving information of the first obstacle vehicle and the crossing
Locate input of the signal lamp state corresponding with the target position as the target nerve network model, passes through the target mind
The driving trace of first vehicle is determined through network model.
3. method according to claim 1 or 2, which is characterized in that the target nerve network model refers to and the mesh
The corresponding neural network model of traveling task in section and first vehicle where cursor position, and the multiple second vehicle
Refer to before current time by the section where the target position and travels the traveling of task and first vehicle and appoint
It is engaged in identical vehicle.
4. a kind of method for planning driving trace, is applied to server, which is characterized in that the described method includes:
From mesh corresponding with the section where the target position that the first vehicle is presently in determining in the neural network model of storage
Neural network model is marked, the target nerve network model is obtained according to the training of the driving trace of multiple second vehicles, institute
It states multiple second vehicles and refers to the vehicle for passing through the section where the target position before current time;
The target nerve network model is sent to first vehicle, so that first vehicle passes through the target nerve net
Network model determines the driving trace of first vehicle.
5. according to the method described in claim 4, it is characterized in that, being determined and first in the neural network model from storage
Before the corresponding target nerve network model in section where the target position that vehicle is presently in, further includes:
Determine the driving trace of all second vehicles in preset time period by the section where the target position and each
The scoring of the driving trace of second vehicle;
Scoring is selected to be greater than N number of driving trace of default scoring from all driving traces got, the N is greater than 1 and small
In or equal to the total quantity of driving trace that gets;
The neural network model of initialization is trained by N number of driving trace, obtains the target nerve network mould
Type.
6. according to the method described in claim 5, it is characterized in that, it is described by N number of driving trace to the mind of initialization
It is trained through network model, obtains the target nerve network model, comprising:
Determine the traveling task of N number of second vehicle and the driving information of driving information and N number of second obstacle vehicle;
Wherein, N number of second vehicle is the corresponding vehicle of the N number of driving trace, N number of second obstacle vehicle with it is described
N number of second vehicle corresponds, and the second obstacle vehicle is that the distance between corresponding second vehicle is less than pre-determined distance threshold
The vehicle of value, the traveling task include keep straight on, turn left, turn right and turn around, the driving information include the position being presently in,
Driving direction and travel speed;
According to the traveling task of N number of second vehicle and driving information, N number of second obstacle vehicle driving information and
The driving trace of N number of second vehicle, is trained the neural network model of initialization, obtains the target nerve network
Model.
7. according to the method described in claim 6, it is characterized in that, the section where the target position is crossing;
The driving information of the traveling task and driving information of N number of second vehicle of determination and N number of second obstacle vehicle it
Afterwards, further includes:
Determine N number of signal lamp state, N number of signal lamp state and N number of second vehicle correspond, each signal lamp shape
State refers to corresponding second vehicle corresponding signal lamp state at crossing when by the crossing;
Correspondingly, the traveling task and driving information according to N number of second vehicle, N number of second obstacle vehicle
The driving trace of driving information and N number of second vehicle, is trained the neural network model of initialization, obtains described
Target nerve network model, comprising:
By the traveling task of N number of second vehicle and driving information, the driving information of N number of second obstacle vehicle and institute
Input of N number of signal lamp state as the neural network model of the initialization is stated, by the driving trace of N number of second vehicle
The output of neural network model as the initialization is trained the neural network model of the initialization, obtains institute
State target nerve network model.
8. the method according to the description of claim 7 is characterized in that the driving trace of each second vehicle of the determination is commented
Point, comprising:
For any second vehicle in all second vehicles, second vehicle is determined according to the driving trace of second vehicle
Whether driving condition in the process of moving, the driving condition include the number to collide, observe traffic rules and regulations, lane change
Number, traveling duration and whether be smooth ride;
According to the driving condition of second vehicle in the process of moving, the scoring of the driving trace of second vehicle is determined.
9. according to any method of claim 4 to 8, which is characterized in that in the neural network model from storage really
Fixed target nerve network model corresponding with the section where the target position that the first vehicle is presently in, comprising:
According to the traveling task of the location information of the target position and first vehicle, from the neural network model of storage
It determines and the corresponding target nerve network model of traveling task in the section and first vehicle where the target position;
Correspondingly, the multiple second vehicle refers to and appoints before current time by the section where the target position and traveling
Business vehicle identical with the traveling task of first vehicle.
10. a kind of device for planning driving trace, is applied to the first vehicle, which is characterized in that described device includes:
Receiving unit, for receive server transmission target nerve network model, the target nerve network model refer to
The corresponding neural network model in section where the target position that first vehicle is presently in, and the target nerve network
Model is that the server is obtained according to the training of the driving trace of multiple second vehicles, and the multiple second vehicle refers to currently
By the vehicle in the section where the target position before time;
First determination unit, for according to the traveling task and driving information of first vehicle and the first obstacle vehicle
Driving information determines the driving trace of first vehicle by the target nerve network model;
Wherein, the traveling task includes keeping straight on, turn left, turn right and turning around, and the driving information includes the position being presently in
Location information, driving direction and travel speed, the first obstacle vehicle is to be less than with the distance between described first vehicle
The vehicle of pre-determined distance threshold value.
11. device according to claim 10, which is characterized in that the section where the target position is crossing;
First determination unit, is specifically used for:
By the traveling task of first vehicle and driving information, the driving information of the first obstacle vehicle and the crossing
Locate input of the signal lamp state corresponding with the target position as the target nerve network model, passes through the target mind
The driving trace of first vehicle is determined through network model.
12. device described in 0 or 11 according to claim 1, which is characterized in that the target nerve network model refer to it is described
The corresponding neural network model of traveling task in section and first vehicle where target position, and the multiple second
Vehicle refers to the section where passing through the target position before current time and travels the traveling of task and first vehicle
The identical vehicle of task.
13. a kind of device for planning driving trace, is applied to server, which is characterized in that described device includes:
Second determination unit, for the target position institute being presently in the first vehicle determining from the neural network model of storage
The corresponding target nerve network model in section, the target nerve network model is the traveling rail according to multiple second vehicles
Mark training obtains, and the multiple second vehicle refers to the vehicle for passing through the section where the target position before current time
?;
Transmission unit, for sending the target nerve network model to first vehicle, so that first vehicle passes through
The target nerve network model determines the driving trace of first vehicle.
14. device according to claim 13, which is characterized in that described device further include:
Third determination unit, for determining in preset time period by all second vehicles in the section where the target position
Driving trace and each second vehicle driving trace scoring;
Selecting unit, for N number of driving trace of the selection scoring greater than default scoring, institute from all driving traces got
N is stated greater than 1 and is less than or equal to the total quantity of the driving trace got;
Training unit obtains the mesh for being trained by N number of driving trace to the neural network model of initialization
Mark neural network model.
15. device according to claim 14, which is characterized in that the training unit includes:
First determines subelement, for determining the traveling task and driving information and N number of second obstacle vehicle of N number of second vehicle
Driving information;
Wherein, N number of second vehicle is the corresponding vehicle of the N number of driving trace, N number of second obstacle vehicle with it is described
N number of second vehicle corresponds, and the second obstacle vehicle is that the distance between corresponding second vehicle is less than pre-determined distance threshold
The vehicle of value, the traveling task include keep straight on, turn left, turn right and turn around, the driving information include the position being presently in,
Driving direction and travel speed;
Training subelement, for traveling task and driving information according to N number of second vehicle, N number of second obstacle vehicle
Driving information and N number of second vehicle driving trace, the neural network model of initialization is trained, is obtained
The target nerve network model.
16. device according to claim 15, which is characterized in that the section where the target position is crossing;
The training unit further include:
Second determines subelement, for determining N number of signal lamp state, N number of signal lamp state and N number of second vehicle one
One is corresponding, and each signal lamp state refers to corresponding second vehicle corresponding signal lamp at crossing when by the crossing
State;
Correspondingly, the trained subelement, is specifically used for:
By the traveling task of N number of second vehicle and driving information, the driving information of N number of second obstacle vehicle and institute
Input of N number of signal lamp state as the neural network model of the initialization is stated, by the driving trace of N number of second vehicle
The output of neural network model as the initialization is trained the neural network model of the initialization, obtains institute
State target nerve network model.
17. device according to claim 16, which is characterized in that the third determination unit is specifically used for:
For any second vehicle in all second vehicles, second vehicle is determined according to the driving trace of second vehicle
Whether driving condition in the process of moving, the driving condition include the number to collide, observe traffic rules and regulations, lane change
Number, traveling duration and whether be smooth ride;
According to the driving condition of second vehicle in the process of moving, the scoring of the driving trace of second vehicle is determined.
18. 3 to 17 any device according to claim 1, which is characterized in that second determination unit is specifically used for:
According to the traveling task of the location information of the target position and first vehicle, from the neural network model of storage
It determines and the corresponding target nerve network model of traveling task in the section and first vehicle where the target position;
Correspondingly, the multiple second vehicle refers to and appoints before current time by the section where the target position and traveling
Business vehicle identical with the traveling task of first vehicle.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710993734.8A CN109697875B (en) | 2017-10-23 | 2017-10-23 | Method and device for planning driving track |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710993734.8A CN109697875B (en) | 2017-10-23 | 2017-10-23 | Method and device for planning driving track |
Publications (2)
Publication Number | Publication Date |
---|---|
CN109697875A true CN109697875A (en) | 2019-04-30 |
CN109697875B CN109697875B (en) | 2020-11-06 |
Family
ID=66225933
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710993734.8A Active CN109697875B (en) | 2017-10-23 | 2017-10-23 | Method and device for planning driving track |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109697875B (en) |
Cited By (15)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110293968A (en) * | 2019-06-18 | 2019-10-01 | 百度在线网络技术(北京)有限公司 | Control method, device, equipment and the readable storage medium storing program for executing of automatic driving vehicle |
CN111339834A (en) * | 2020-02-04 | 2020-06-26 | 浙江大华技术股份有限公司 | Method for recognizing vehicle traveling direction, computer device, and storage medium |
CN111572562A (en) * | 2020-07-03 | 2020-08-25 | 腾讯科技(深圳)有限公司 | Automatic driving method, device, equipment, system, vehicle and computer readable storage medium |
CN111724598A (en) * | 2020-06-29 | 2020-09-29 | 北京百度网讯科技有限公司 | Method, device, equipment and storage medium for planning path |
CN111832597A (en) * | 2019-08-01 | 2020-10-27 | 北京嘀嘀无限科技发展有限公司 | Vehicle type determination method and device |
CN112078592A (en) * | 2019-06-13 | 2020-12-15 | 初速度(苏州)科技有限公司 | Method and device for predicting vehicle behavior and/or vehicle track |
CN112364847A (en) * | 2021-01-12 | 2021-02-12 | 深圳裹动智驾科技有限公司 | Automatic driving prediction method based on personal big data and computer equipment |
CN112904843A (en) * | 2021-01-14 | 2021-06-04 | 清华大学苏州汽车研究院(吴江) | Automatic driving scene determining method and device, electronic equipment and storage medium |
WO2021147748A1 (en) * | 2020-01-22 | 2021-07-29 | 华为技术有限公司 | Self-driving method and related device |
CN113971885A (en) * | 2020-07-06 | 2022-01-25 | 华为技术有限公司 | Vehicle speed prediction method, device and system |
WO2022016901A1 (en) * | 2020-07-20 | 2022-01-27 | 华为技术有限公司 | Method for planning driving route of vehicle, and intelligent vehicle |
CN114299712A (en) * | 2021-11-26 | 2022-04-08 | 腾讯科技(深圳)有限公司 | Data processing method, device, equipment and readable storage medium |
CN114620072A (en) * | 2022-03-14 | 2022-06-14 | 小米汽车科技有限公司 | Vehicle control method, device, storage medium, electronic device and vehicle |
CN114822050A (en) * | 2022-03-30 | 2022-07-29 | 阿里巴巴(中国)有限公司 | Road condition identification method, electronic equipment and computer program product |
CN111832597B (en) * | 2019-08-01 | 2024-06-11 | 北京嘀嘀无限科技发展有限公司 | Method and device for judging vehicle type |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20090324010A1 (en) * | 2008-06-26 | 2009-12-31 | Billy Hou | Neural network-controlled automatic tracking and recognizing system and method |
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 |
CN106340205A (en) * | 2016-09-30 | 2017-01-18 | 广东中星微电子有限公司 | Traffic monitoring method and traffic monitoring apparatus |
CN106548645A (en) * | 2016-11-03 | 2017-03-29 | 济南博图信息技术有限公司 | Vehicle route optimization method and system based on deep learning |
-
2017
- 2017-10-23 CN CN201710993734.8A patent/CN109697875B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20090324010A1 (en) * | 2008-06-26 | 2009-12-31 | Billy Hou | Neural network-controlled automatic tracking and recognizing system and method |
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 |
CN106340205A (en) * | 2016-09-30 | 2017-01-18 | 广东中星微电子有限公司 | Traffic monitoring method and traffic monitoring apparatus |
CN106548645A (en) * | 2016-11-03 | 2017-03-29 | 济南博图信息技术有限公司 | Vehicle route optimization method and system based on deep learning |
Cited By (22)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112078592A (en) * | 2019-06-13 | 2020-12-15 | 初速度(苏州)科技有限公司 | Method and device for predicting vehicle behavior and/or vehicle track |
CN112078592B (en) * | 2019-06-13 | 2021-12-24 | 魔门塔(苏州)科技有限公司 | Method and device for predicting vehicle behavior and/or vehicle track |
CN110293968A (en) * | 2019-06-18 | 2019-10-01 | 百度在线网络技术(北京)有限公司 | Control method, device, equipment and the readable storage medium storing program for executing of automatic driving vehicle |
CN111832597A (en) * | 2019-08-01 | 2020-10-27 | 北京嘀嘀无限科技发展有限公司 | Vehicle type determination method and device |
CN111832597B (en) * | 2019-08-01 | 2024-06-11 | 北京嘀嘀无限科技发展有限公司 | Method and device for judging vehicle type |
WO2021147748A1 (en) * | 2020-01-22 | 2021-07-29 | 华为技术有限公司 | Self-driving method and related device |
CN111339834A (en) * | 2020-02-04 | 2020-06-26 | 浙江大华技术股份有限公司 | Method for recognizing vehicle traveling direction, computer device, and storage medium |
CN111339834B (en) * | 2020-02-04 | 2023-06-02 | 浙江大华技术股份有限公司 | Method for identifying vehicle driving direction, computer device and storage medium |
CN111724598B (en) * | 2020-06-29 | 2022-04-05 | 北京百度网讯科技有限公司 | Method, device, equipment and storage medium for automatically driving and planning path |
CN111724598A (en) * | 2020-06-29 | 2020-09-29 | 北京百度网讯科技有限公司 | Method, device, equipment and storage medium for planning path |
CN111572562A (en) * | 2020-07-03 | 2020-08-25 | 腾讯科技(深圳)有限公司 | Automatic driving method, device, equipment, system, vehicle and computer readable storage medium |
CN113971885A (en) * | 2020-07-06 | 2022-01-25 | 华为技术有限公司 | Vehicle speed prediction method, device and system |
CN113971885B (en) * | 2020-07-06 | 2023-03-03 | 华为技术有限公司 | Vehicle speed prediction method, device and system |
WO2022016901A1 (en) * | 2020-07-20 | 2022-01-27 | 华为技术有限公司 | Method for planning driving route of vehicle, and intelligent vehicle |
CN112364847A (en) * | 2021-01-12 | 2021-02-12 | 深圳裹动智驾科技有限公司 | Automatic driving prediction method based on personal big data and computer equipment |
CN112904843B (en) * | 2021-01-14 | 2022-04-08 | 清华大学苏州汽车研究院(吴江) | Automatic driving scene determining method and device, electronic equipment and storage medium |
CN112904843A (en) * | 2021-01-14 | 2021-06-04 | 清华大学苏州汽车研究院(吴江) | Automatic driving scene determining method and device, electronic equipment and storage medium |
CN114299712A (en) * | 2021-11-26 | 2022-04-08 | 腾讯科技(深圳)有限公司 | Data processing method, device, equipment and readable storage medium |
CN114299712B (en) * | 2021-11-26 | 2024-03-01 | 腾讯科技(深圳)有限公司 | Data processing method, device, equipment and readable storage medium |
CN114620072A (en) * | 2022-03-14 | 2022-06-14 | 小米汽车科技有限公司 | Vehicle control method, device, storage medium, electronic device and vehicle |
CN114822050A (en) * | 2022-03-30 | 2022-07-29 | 阿里巴巴(中国)有限公司 | Road condition identification method, electronic equipment and computer program product |
CN114822050B (en) * | 2022-03-30 | 2023-07-21 | 阿里巴巴(中国)有限公司 | Road condition identification method, electronic equipment and computer program product |
Also Published As
Publication number | Publication date |
---|---|
CN109697875B (en) | 2020-11-06 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109697875A (en) | Plan the method and device of driving trace | |
US11726477B2 (en) | Methods and systems for trajectory forecasting with recurrent neural networks using inertial behavioral rollout | |
US20220227394A1 (en) | Autonomous Vehicle Operational Management | |
WO2020224434A1 (en) | Driving simulation method and apparatus, electronic device and computer storage medium | |
CN109863513B (en) | Neural network system for autonomous vehicle control | |
US20190378404A1 (en) | Traffic prediction system, vehicle-mounted display apparatus, vehicle, and traffic prediction method | |
US20220051032A1 (en) | Road event detection method, apparatus, device and storage medium | |
Wang et al. | Vehicle collision prediction at intersections based on comparison of minimal distance between vehicles and dynamic thresholds | |
CN113272830A (en) | Trajectory representation in a behavior prediction system | |
CN109993969A (en) | A kind of road conditions determine information acquisition method, device and equipment | |
CN112106124A (en) | System and method for using V2X and sensor data | |
US20210191394A1 (en) | Systems and methods for presenting curated autonomy-system information of a vehicle | |
US11681296B2 (en) | Scenario-based behavior specification and validation | |
KR20210038852A (en) | Method, apparatus, electronic device, computer readable storage medium and computer program for early-warning | |
KR102007181B1 (en) | Method for transmitting route data for traffic telematics | |
US11731652B2 (en) | Systems and methods for reactive agent simulation | |
CN111477028B (en) | Method and device for generating information in automatic driving | |
US8849625B2 (en) | Distributed process simulator | |
CN108171967A (en) | A kind of traffic control method and device | |
CN101469992A (en) | Processing method and apparatus for vehicle navigation and vehicle navigation system | |
CN103065463A (en) | Traffic control simulation system and traffic control simulation method | |
JP2021099877A (en) | Method, device, apparatus and storage medium for reminding travel on exclusive driveway | |
US20220413502A1 (en) | Method, apparatus, and system for biasing a machine learning model toward potential risks for controlling a vehicle or robot | |
CN108399778A (en) | Swarm intelligence congestion reminding method, system and computer readable storage medium | |
CN114829225A (en) | Conditional behavior prediction for autonomous vehicles |
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 |