CN109760681A - A kind of lane-change control method and device - Google Patents
A kind of lane-change control method and device Download PDFInfo
- Publication number
- CN109760681A CN109760681A CN201711064941.1A CN201711064941A CN109760681A CN 109760681 A CN109760681 A CN 109760681A CN 201711064941 A CN201711064941 A CN 201711064941A CN 109760681 A CN109760681 A CN 109760681A
- Authority
- CN
- China
- Prior art keywords
- lane
- change
- vehicle
- data
- movement
- 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
Landscapes
- Control Of Driving Devices And Active Controlling Of Vehicle (AREA)
Abstract
The embodiment of the invention discloses a kind of lane-change control method and device, this method, comprising: obtains the current environment data of current time controlled vehicle ambient enviroment；According to the scoring that the action evaluation model and current environment data being previously obtained, acquisition current time act lane-change, lane-change movement includes acceleration and steering angle；Highest corresponding lane-change movement of scoring is determined as the selected lane-change movement at current time；It is acted according to selected lane-change, controls controlled vehicle movement.The embodiment of the present invention to during lane-change acceleration and steering angle comprehensively consider, avoid by vertical and horizontal planning by force decouple caused by stability problem, increase robustness and the comfort level of passenger.
Description
Technical field
The present invention relates to automatic Pilot technical field more particularly to a kind of lane-change control method and device.
Background technique
Automatic lane-change be realize one of the key technology of Vehicular automatic driving, but by the nonlinear characteristic of automobile dynamics with
And how the restriction of the driving environment of actual complex guarantees vehicle safety, quick, stable lane-change during automatic Pilot
It is the key points and difficulties of current automatic Pilot technical research again.
During the control of automatic lane-change, on the adjacent lane line and adjacent lane line by identifying controlled vehicle
Vehicle location and speed calculate local route, and controlled vehicle is made to move to the lane center that adjacent vehicle arrives through the local route.
Existing lane-change control technology has the disadvantage in that the vertically and horizontally controlling planning that will be intercoupled decouples by force, to longitudinal acceleration
With laterally turn to separately planning, cause that autonomous driving vehicle robustness during lane-change is not high, passenger comfort is bad.
Summary of the invention
In view of this, being able to solve existing automatic Pilot vapour the present invention provides a kind of lane-change control method and device
The vehicle problem that robustness is not high during lane-change, comfort is bad.
A kind of lane-change control method provided in an embodiment of the present invention, comprising:
Obtain the current environment data of current time controlled vehicle ambient enviroment；
According to the action evaluation model and the current environment data being previously obtained, obtain what current time acted lane-change
Scoring, the lane-change movement includes acceleration and steering angle；
Highest corresponding lane-change movement of scoring is determined as the selected lane-change movement at current time；
It is acted according to the selected lane-change, controls the controlled vehicle movement.
Optionally, the method for building up of the action evaluation model, specifically includes:
Lane-change action training collection is obtained, the lane-change action training collection includes multiple groups lane-change data, every group of lane-change data packet
First described in including when driver during a lane-change controls each lane-change movement that the first vehicle executes and executes lane-change movement
The vehicle speed data of vehicle and its environmental data of ambient enviroment；
According to the lane-change action training collection, based on convolutional neural networks to training objective functionBe trained, training convergence after, obtain the action evaluation model q (s,
a)；
Wherein, a is the movement of current lane-change, and s is current environment data, and a' is that the lane-change of subsequent time acts, and s' is next
The environmental data at moment, γ are learning rate, and R (s) is the instant return at current time.
Optionally,
During lane-change, the instant return is positively correlated with the first comfortable data and the second comfortable data, described instant
Return is also positively correlated with the first secure data and/or the second secure data；
Wherein, the variation degree of the described first comfortable data and first vehicle transverse acceleration within a preset period of time is negative
It is related；The variation degree negative of the second comfortable data and first vehicle longitudinal acceleration in the preset time period
It closes；First secure data and first distance and the speed of the second vehicle are positively correlated, and second vehicle is in first front side
And it is located at lane to be transferred to, the first distance is the fore-and-aft distance of second vehicle and first vehicle, first safety
Data also with the speed of first vehicle, maximum braking deceleration and react time delay negative correlation；Second secure data and
Two distances and the speed of first vehicle are positively correlated, and the second distance is the fore-and-aft distance of third vehicle and first vehicle, institute
Third vehicle is stated at first vehicle rear and is located at the lane to be transferred to, second secure data also with the third vehicle
Speed, maximum braking deceleration and reaction time delay are negatively correlated；
At the end of lane-change, the instant return is 100.
Optionally, during lane-change, the instant return specifically:
Wherein, [f0,f1] it is transverse acceleration frequency spectrum, [f2,f3] it is longitudinal acceleration frequency spectrum, axIt is first vehicle in institute
State the transverse acceleration in preset time period, ayFor longitudinal acceleration of first vehicle in the preset time period, y1For institute
State first distance, y2For the second distance, v1For the speed of first vehicle, v2For the speed of second vehicle, v3It is described
The speed of third vehicle,
amaxThe maximum braking deceleration of vehicle is corresponded to for k,
τ is the reaction time delay that k corresponds to vehicle.
Optionally, the action evaluation model and the current environment data that the basis is previously obtained obtain current time
Scoring to lane-change movement, specifically includes:
The current environment is inputted the action evaluation model q, and (s, a), the lane-change for obtaining current time act scoring letter
Number q (a)；
It is described that highest corresponding lane-change movement of scoring is determined as the selected lane-change at current time and acts, it specifically includes:
According to formulaDetermine the selected lane-change movement aIt is selected。
A kind of lane-change control device provided in an embodiment of the present invention, comprising: acquiring unit, scoring unit, determination unit and
Control unit；
The acquiring unit, for obtaining the current environment data of current time controlled vehicle ambient enviroment；
The scoring unit, for being worked as according to the action evaluation model and the current environment data being previously obtained
The scoring that the preceding moment acts lane-change, the lane-change movement includes acceleration and steering angle；
The determination unit, for highest corresponding lane-change movement of scoring to be determined as the selected lane-change at current time and moves
Make；
Described control unit controls the controlled vehicle movement for acting according to the selected lane-change.
Optionally, described device, further includes: model training unit；The model training unit, is specifically used for:
Lane-change action training collection is obtained, the lane-change action training collection includes multiple groups lane-change data, every group of lane-change data packet
First described in including when driver during a lane-change controls each lane-change movement that the first vehicle executes and executes lane-change movement
The vehicle speed data of vehicle and its environmental data of ambient enviroment；
According to the lane-change action training collection, based on convolutional neural networks to training objective functionBe trained, training convergence after, obtain the action evaluation model q (s,
a)；
Wherein, a is the movement of current lane-change, and s is current environment data, and a' is that the lane-change of subsequent time acts, and s' is next
The environmental data at moment, γ are learning rate, and R (s) is the instant return at current time.
Optionally,
During lane-change, the instant return is positively correlated with the first comfortable data and the second comfortable data, described instant
Return is also positively correlated with the first secure data and/or the second secure data；
Wherein, the variation degree of the described first comfortable data and first vehicle transverse acceleration within a preset period of time is negative
It is related；The variation degree negative of the second comfortable data and first vehicle longitudinal acceleration in the preset time period
It closes；First secure data and first distance and the speed of the second vehicle are positively correlated, and second vehicle is in first front side
And it is located at lane to be transferred to, the first distance is the fore-and-aft distance of second vehicle and first vehicle, first safety
Data also with the speed of first vehicle, maximum braking deceleration and react time delay negative correlation；Second secure data and
Two distances and the speed of first vehicle are positively correlated, and the second distance is the fore-and-aft distance of third vehicle and first vehicle, institute
Third vehicle is stated at first vehicle rear and is located at the lane to be transferred to, second secure data also with the third vehicle
Speed, maximum braking deceleration and reaction time delay are negatively correlated；
At the end of lane-change, the instant return is 100.
Optionally, during lane-change, the instant return specifically:
Wherein, [f0,f1] it is transverse acceleration frequency spectrum, [f2,f3] it is longitudinal acceleration frequency spectrum, axIt is first vehicle in institute
State the transverse acceleration in preset time period, ayFor longitudinal acceleration of first vehicle in the preset time period, y1For institute
State first distance, y2For the second distance, v1For the speed of first vehicle, v2For the speed of second vehicle, v3It is described
The speed of third vehicle,
amaxThe maximum braking deceleration of vehicle is corresponded to for k,
τ is the reaction time delay that k corresponds to vehicle.
Optionally, the scoring unit, is specifically used for:
The current environment is inputted the action evaluation model q, and (s, a), the lane-change for obtaining current time act scoring letter
Number q (a)；
It is described that highest corresponding lane-change movement of scoring is determined as the selected lane-change at current time and acts, it specifically includes:
According to formulaDetermine the selected lane-change movement aIt is selected。
The embodiment of the invention also provides a kind of computer readable storage mediums, are stored thereon with computer program, described
Computer program is performed, and is performed the steps of
Obtain the current environment data of current time controlled vehicle ambient enviroment；
According to the action evaluation model and the current environment data being previously obtained, obtain what current time acted lane-change
Scoring, the lane-change movement includes acceleration and steering angle；
Highest corresponding lane-change movement of scoring is determined as the selected lane-change movement at current time；
It is acted according to the selected lane-change, controls the controlled vehicle movement.
The embodiment of the invention also provides a kind of entire car controllers, comprising: memory and processor；
The memory, for being stored with computer program, the computer program can when being executed by the processor
It performs the steps of
Obtain the current environment data of current time controlled vehicle ambient enviroment；
According to the action evaluation model and the current environment data being previously obtained, obtain what current time acted lane-change
Scoring, the lane-change movement includes acceleration and steering angle；
Highest corresponding lane-change movement of scoring is determined as the selected lane-change movement at current time；
It is acted according to the selected lane-change, controls the controlled vehicle movement.
Compared with prior art, the present invention has at least the following advantages:
In embodiments of the present invention, the environmental data around controlled vehicle is acquired first, according to the environmental data and preparatory instruction
The action evaluation model got obtains the combination scoring of current time acceleration and steering angle to get under current state
It is dynamic to be controlled vehicle using the corresponding lane-change action control of highest scoring for the similarity of the acceleration and steering angle and true driver behavior
Make, so that the true lane-change movement under current state close to driver of the lane-change process of controlled vehicle, ensure that passenger's is comfortable
Degree.The embodiment of the present invention to during lane-change acceleration and steering angle comprehensively consider, avoid and advise vertical and horizontal
Stability problem caused by decoupling by force is drawn, robustness and the comfort level of passenger are increased.
Detailed description of the invention
In order to illustrate the technical solutions in the embodiments of the present application or in the prior art more clearly, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
The some embodiments recorded in application, for those of ordinary skill in the art, without creative efforts,
It can also be obtained according to these attached drawings other attached drawings.
Fig. 1 is the schematic diagram of vehicle lane-change to the left；
Fig. 2 is a kind of flow diagram of lane-change control method provided in an embodiment of the present invention；
Fig. 3 is the flow diagram of another lane-change control method provided in an embodiment of the present invention；
Fig. 4 is the training process schematic diagram of convolutional neural networks in the specific embodiment of the invention；
Fig. 5 is a kind of structural schematic diagram of lane-change control device provided in an embodiment of the present invention.
Specific embodiment
In order to enable those skilled in the art to better understand the solution of the present invention, below in conjunction in the embodiment of the present invention
Attached drawing, technical scheme in the embodiment of the invention is clearly and completely described, it is clear that described embodiment is only this
Invention a part of the embodiment, instead of all the embodiments.Based on the embodiments of the present invention, those of ordinary skill in the art exist
Every other embodiment obtained under the premise of creative work is not made, shall fall within the protection scope of the present invention.
Before introducing specific embodiments of the present invention, the detailed process of lower vehicle lane-changing is introduced first.Vehicle changes to the left
The process in road has the rear vehicle B to have vehicle C as shown in Figure 1, vehicle A will carry out lane-change to the left in front of the adjacent lane of left,
Vehicle A and lane line L1、L2And L3Distance be respectively x1、x2And x3.The condition that lane-change terminates is | x1-x2| < Kw and | h-hm|
< C.Wherein, w is the lane width of left-hand lane, and K is constant, characterizes degree of the vehicle driving among lane, the smaller vehicle of K
Closer to lane center LmTraveling, it is vehicle course angle, h that general K, which takes 0-0.1, h,mAngle is navigated by water for left-lane, C is constant, one
As take 0-0.1.
Referring to fig. 2, which is a kind of flow diagram of lane-change control method provided in an embodiment of the present invention.In the present invention
In embodiment, be illustrated by taking controlled vehicle to the left lane-change as an example, the control method for being controlled vehicle lane-change to the right is similar, referring specifically to
The related description of left lane-change, repeats no more in embodiments of the present invention.
Lane-change control method provided in this embodiment, includes the following steps S201-S204.
S201: the current environment data of current time controlled vehicle ambient enviroment are obtained.
In the present embodiment, the environmental data of controlled vehicle can be obtained by environmental perception module, wherein environmental data packet
Include information of vehicles, the lane line information etc. around controlled vehicle.
For example, environmental perception module is responsible for obtaining the overhead view image around controlled vehicle, can specifically be swashed by multi-thread
The point cloud data of optical radar scanning constitutes the overhead view image, alternatively, can also pass through front camera, rear camera, left and right
The video image for looking around camera acquisition splices to obtain the overhead view image.It should be noted that the peace in order to guarantee automatic lane-change
Entirely, the detecting distance of camera needs to be greater than 100 meters.
S202: according to the action evaluation model and current environment data being previously obtained, current time is obtained to lane-change movement
Scoring.
What needs to be explained here is that lane-change movement includes the acceleration and lane-change angle (or steering wheel angle) of vehicle, it can be with
Lane-change movement is indicated with a (α, θ), and α is acceleration, and θ is lane-change angle.For example, a (1, -2) indicates that controlled vehicle is accelerated with 1m/s, together
When controlled vehicle lane-change angle be to turn left 2 °；A (- 1,2) indicates that controlled vehicle is slowed down with 1m/s, while the lane-change angle of controlled vehicle is to turn right
2°。
It is understood that existing automatic lane-change technology, carries out to longitudinal (i.e. acceleration) and lateral (turning to)
When controlling planning, planning control is carried out to the two respectively, and is intercoupled between longitudinal direction and crosswise joint, the two influences each other,
The robustness for causing lane-change to control is bad, passenger comfort level is not high.And in the embodiment of the present invention, by longitudinally controlled and laterally control
It is made as an entirety (i.e. lane-change movement), integrated planning longitudinal direction of car and lateral planning improve robustness and comfort level.
In the present embodiment, action evaluation model acts phase with the environmental data around controlled vehicle and its lane-change of execution
It closes, it is similar between the vehicle lane-change movement executed controlled under current state and the true driver behavior of driver for evaluating
Degree, the higher similarity that scores are higher.Action evaluation model can actually drive lane-change process according to driver gathered in advance
Lane-change movement, using convolutional neural networks training obtain.
S203: highest corresponding lane-change movement of scoring is determined as the selected lane-change at current time and acted.
S204: acting according to selected lane-change, controls controlled vehicle movement.
S201 and S202 through the above steps has obtained being controlled vehicle execution lane-change movement score function under the environmental data,
I.e. currently controlled vehicle executes the similarity between lane-change movement and the true driver behavior of driver, and scores between similarity just
Therefore highest corresponding lane-change movement of scoring is determined as selected lane-change movement by correlation, and control controlled vehicle and execute this and selected change
Road movement can be that there is the lane-change process of controlled vehicle class people to drive effect, have preferable comfort level.
In the present embodiment, the environmental data around controlled vehicle is acquired first, according to the environmental data and trained in advance
The action evaluation model arrived, the combination scoring for obtaining current time acceleration and steering angle should add to get under current state
The similarity of speed and steering angle and true driver behavior is controlled vehicle movement using the corresponding lane-change action control of highest scoring,
So that the true lane-change movement under current state close to driver of the lane-change process of controlled vehicle, ensure that the comfort level of passenger.
The embodiment of the present invention to during lane-change acceleration and lane-change angle comprehensively consider, avoid vertical and horizontal planning is strong
Stability problem caused by row decouples, increases robustness and the comfort level of passenger.
It is exemplified below and specifically how to establish action evaluation model.
Referring to Fig. 3, in embodiments of the present invention, action evaluation model can specifically pass through following steps S301-S302 institute
The method training stated obtains.
S301: lane-change action training collection is obtained.
Wherein, lane-change action training collection includes multiple groups lane-change data, and every group of lane-change data include that a lane-change is driven in the process
The vehicle speed data and its surrounding ring of first vehicle when the person of sailing controls each lane-change movement that the first vehicle executes and executes lane-change movement
The environmental data in border.
It, can be with it is understood that with the movement of the practical lane-change of driver for according to being trained to action evaluation model
So that the highest lane-change that scores acts and driver is in the practical lane-change operation similarity executed of current environment when practical control
It is higher, improve the robustness of lane-change process and the comfort level of passenger.
What needs to be explained here is that in practical applications, due to the difference of vehicle, will lead to the lane-change movement actually executed
It is distinct, in order to guarantee the accuracy and comfort of lane-change, when training action evaluation model used lane-change action training
Collection is needed with actual controlled vehicle correlation, i.e. the first vehicle is identical as the vehicle of controlled vehicle (or size is similar).
S302: according to lane-change action training collection, following formula (1) is shown based on convolutional neural networks training objective function into
Row training, training convergence after, obtain action evaluation model q (s, a)；
Wherein, a is the movement of current lane-change, and s is current environment data, and a' is that the lane-change of subsequent time acts, and s' is next
The environmental data at moment, γ are learning rate, and R (s) is the instant return at current time.
The training process of convolutional neural networks is as shown in figure 4, those skilled in the art can specifically set according to the actual situation
Determine the parameter of convolutional layer, pond layer and full articulamentum, no longer the specific training process of convolutional neural networks is repeated here.
In practical applications, learning rate can be equal to 0.9 with γ.
In the possible implementation of the present embodiment, during lane-change, return immediately is relaxed with the first comfortable data and second
Suitable data are positively correlated, and return is also positively correlated with the first secure data and/or the second secure data immediately；At the end of lane-change, immediately
Return is 100.
Wherein, the variation journey of the first comfortable data and the first vehicle (A vehicle as shown in figure 1) transverse acceleration within a preset period of time
Degree is negatively correlated；The variation degree of second comfortable data and the first vehicle longitudinal acceleration within a preset period of time is negatively correlated；First peace
Total to be positively correlated according to first distance and the speed of the second vehicle (B vehicle as shown in figure 1), the second vehicle is in the first front side and is located at wait turn
Enter lane, first distance is the fore-and-aft distance of the second vehicle and the first vehicle, and the first secure data is also made with the speed of the first vehicle, maximum
Dynamic deceleration and reaction time delay are negatively correlated；The speed of second secure data and second distance and the first vehicle is positively correlated, second distance
For the fore-and-aft distance of third vehicle (C vehicle as shown in figure 1) and the first vehicle, third vehicle is at the first vehicle rear and is located at lane to be transferred to, the
Two secure datas also with the speed of third vehicle, maximum braking deceleration and react time delay negative correlation.
As an example, during lane-change, return is specifically obtained by following formula (2) immediately:
Wherein, [f0,f1] it is transverse acceleration frequency spectrum, [f2,f3] it is longitudinal acceleration frequency spectrum, axIt is the first vehicle when default
Between transverse acceleration in section, ayFor the longitudinal acceleration of the first vehicle within a preset period of time, y1For first distance, y2It is second
Distance, v1For the speed of the first vehicle, v2For the speed of the second vehicle, v3For the speed of third vehicle, amaxThe maximum braking of vehicle is corresponded to for k
Deceleration, τ are the reaction time delay that k corresponds to vehicle,
Then, the step S202-S203 in above-described embodiment, can specifically include:
By current environment input action evaluation model q, (s, a), the lane-change for obtaining current time act score function q (a)；
Highest corresponding lane-change movement of scoring is determined as the selected lane-change movement at current time, is specifically included: public according to following formula (4)
Formula determines that selected lane-change acts aIt is selected。
The lane-change control method provided based on the above embodiment, the embodiment of the invention also provides a kind of controls of lane-change to fill
It sets.
Referring to Fig. 5, which is a kind of structural schematic diagram of lane-change control device provided in an embodiment of the present invention.
A kind of lane-change control device provided in this embodiment, comprising: acquiring unit 100, scoring unit 200, determination unit
300 and control unit 400；
Acquiring unit 100, for obtaining the current environment data of current time controlled vehicle ambient enviroment；
Score unit 200, for obtaining current time according to the action evaluation model and current environment data being previously obtained
Scoring to lane-change movement, lane-change movement include acceleration and steering angle；
Determination unit 300, for highest corresponding lane-change movement of scoring to be determined as the selected lane-change at current time and acts；
Control unit 400, for controlling controlled vehicle movement according to lane-change movement is chosen.
In some possible implementations of the present embodiment, further includes: model training unit；
Model training unit, is specifically used for:
Lane-change action training collection is obtained, lane-change action training collection includes multiple groups lane-change data, and every group of lane-change data include one
Driver controls the movement of each lane-change and the speed for executing the first vehicle when the lane-change acts that the first vehicle executes during a lane-change
The environmental data of data and its ambient enviroment；
According to lane-change action training collection, training objective function shown in following formula (1) is instructed based on convolutional neural networks
Practice, training convergence after, obtain action evaluation model q (s, a)；
Wherein, a is the movement of current lane-change, and s is current environment data, and a' is that the lane-change of subsequent time acts, and s' is next
The environmental data at moment, γ are learning rate, and R (s) is the instant return at current time.
In some possible implementations of the present embodiment, during lane-change, immediately return with the first comfortable data and
Second comfortable data are positively correlated, and return is also positively correlated with the first secure data and/or the second secure data immediately；
Wherein, the variation degree of the first comfortable data and the first vehicle transverse acceleration within a preset period of time is negatively correlated；The
The variation degree of two comfortable data and the first vehicle longitudinal acceleration within a preset period of time is negatively correlated；First secure data and first
Distance and the speed of the second vehicle are positively correlated, and the second vehicle is in the first front side and is located at lane to be transferred to, and first distance is the second vehicle
With the fore-and-aft distance of the first vehicle, the first secure data also with the speed of the first vehicle, maximum braking deceleration and react time delay negative
It closes；The speed of second secure data and second distance and the first vehicle is positively correlated, and second distance is the longitudinal direction of third vehicle and the first vehicle
Distance, third vehicle is at the first vehicle rear and is located at lane to be transferred to, and the second secure data is also braked with the speed of third vehicle, maximum
Deceleration and reaction time delay are negatively correlated；
At the end of lane-change, return immediately is 100.
In some possible implementations of the present embodiment, during lane-change, return is specifically obtained by following formula (2) immediately
Out:
Wherein, [f0,f1] it is transverse acceleration frequency spectrum, [f2,f3] it is longitudinal acceleration frequency spectrum, axIt is the first vehicle when default
Between transverse acceleration in section, ayFor the longitudinal acceleration of the first vehicle within a preset period of time, y1For first distance, y2It is second
Distance, v1For the speed of the first vehicle, v2For the speed of the second vehicle, v3For the speed of third vehicle, amaxThe maximum braking of vehicle is corresponded to for k
Deceleration, τ are the reaction time delay that k corresponds to vehicle,
In some possible implementations of the present embodiment, score unit, is specifically used for:
By current environment input action evaluation model q, (s, a), the lane-change for obtaining current time act score function q (a)；
Highest corresponding lane-change movement of scoring is determined as the selected lane-change movement at current time, is specifically included:
According to following formula (4), determine that selected lane-change acts aIt is selected。
In the present embodiment, the environmental data around controlled vehicle is acquired first, according to the environmental data and trained in advance
The action evaluation model arrived, the combination scoring for obtaining current time acceleration and steering angle should add to get under current state
The similarity of speed and steering angle and true driver behavior is controlled vehicle movement using the corresponding lane-change action control of highest scoring,
So that the true lane-change movement under current state close to driver of the lane-change process of controlled vehicle, ensure that the comfort level of passenger.
The embodiment of the present invention to during lane-change acceleration and lane-change angle comprehensively consider, avoid vertical and horizontal planning is strong
Stability problem caused by row decouples, increases robustness and the comfort level of passenger.
The lane-change control method and device provided based on the above embodiment, the embodiment of the invention also provides a kind of computers
Readable storage medium storing program for executing.Computer program is stored on the computer readable storage medium, the computer program is performed, real
Existing following steps:
Obtain the current environment data of current time controlled vehicle ambient enviroment；
According to the action evaluation model and the current environment data being previously obtained, obtain what current time acted lane-change
Scoring, the lane-change movement includes acceleration and steering angle；
Highest corresponding lane-change movement of scoring is determined as the selected lane-change movement at current time；
It is acted according to the selected lane-change, controls the controlled vehicle movement.
The lane-change control method and device provided based on the above embodiment, the embodiment of the invention also provides a kind of vehicle controls
Device processed.The entire car controller, comprising: memory and processor；
The memory, for being stored with computer program, the computer program can when being executed by the processor
It performs the steps of
Obtain the current environment data of current time controlled vehicle ambient enviroment；
According to the action evaluation model and the current environment data being previously obtained, obtain what current time acted lane-change
Scoring, the lane-change movement includes acceleration and steering angle；
Highest corresponding lane-change movement of scoring is determined as the selected lane-change movement at current time；
It is acted according to the selected lane-change, controls the controlled vehicle movement.
It should be noted that each embodiment in this specification is described in a progressive manner, each embodiment emphasis is said
Bright is the difference from other embodiments, and the same or similar parts in each embodiment may refer to each other.For reality
For applying device disclosed in example, since it is corresponded to the methods disclosed in the examples, so description is fairly simple, related place ginseng
See method part illustration.
It should also be noted that, herein, relational terms such as first and second and the like are used merely to one
Entity or operation are distinguished with another entity or operation, without necessarily requiring or implying between these entities or operation
There are any actual relationship or orders.Moreover, the terms "include", "comprise" or its any other variant are intended to contain
Lid non-exclusive inclusion, so that the process, method, article or equipment including a series of elements is not only wanted including those
Element, but also including other elements that are not explicitly listed, or further include for this process, method, article or equipment
Intrinsic element.In the absence of more restrictions, the element limited by sentence "including a ...", it is not excluded that
There is also other identical elements in process, method, article or equipment including the element.
The step of method described in conjunction with the examples disclosed in this document or algorithm, can directly be held with hardware, processor
The combination of capable software module or the two is implemented.Software module can be placed in random access memory (RAM), memory, read-only deposit
Reservoir (ROM), electrically programmable ROM, electrically erasable ROM, register, hard disk, moveable magnetic disc, CD-ROM or technology
In any other form of storage medium well known in field.
The above described is only a preferred embodiment of the present invention, being not intended to limit the present invention in any form.Though
So the present invention has been disclosed as a preferred embodiment, and however, it is not intended to limit the invention.It is any to be familiar with those skilled in the art
Member, without departing from the scope of the technical proposal of the invention, all using the methods and technical content of the disclosure above to the present invention
Technical solution makes many possible changes and modifications or equivalent example modified to equivalent change.Therefore, it is all without departing from
The content of technical solution of the present invention, according to the technical essence of the invention any simple modification made to the above embodiment, equivalent
Variation and modification, all of which are still within the scope of protection of the technical scheme of the invention.
Claims (12)
1. a kind of lane-change control method, which is characterized in that the method, comprising:
Obtain the current environment data of current time controlled vehicle ambient enviroment；
According to the action evaluation model and the current environment data being previously obtained, acquisition current time comments lane-change movement
Point, the lane-change movement includes acceleration and steering angle；
Highest corresponding lane-change movement of scoring is determined as the selected lane-change movement at current time；
It is acted according to the selected lane-change, controls the controlled vehicle movement.
2. lane-change control method according to claim 1, which is characterized in that the method for building up of the action evaluation model,
It specifically includes:
Lane-change action training collection is obtained, the lane-change action training collection includes multiple groups lane-change data, and every group of lane-change data include one
First vehicle described in when driver controls each lane-change movement of the first vehicle execution and executes lane-change movement during a lane-change
The environmental data of vehicle speed data and its ambient enviroment；
According to the lane-change action training collection, based on convolutional neural networks to training objective functionBe trained, training convergence after, obtain the action evaluation model q (s,
a)；
Wherein, a is the movement of current lane-change, and s is current environment data, and a' is that the lane-change of subsequent time acts, and s' is subsequent time
Environmental data, γ is learning rate, and R (s) is the instant return at current time.
3. lane-change control method according to claim 2, which is characterized in that
During lane-change, the instant return is positively correlated with the first comfortable data and the second comfortable data, the instant return
Also it is positively correlated with the first secure data and/or the second secure data；
Wherein, the variation degree negative of the described first comfortable data and first vehicle transverse acceleration within a preset period of time
It closes；The variation degree of the second comfortable data and the first vehicle longitudinal acceleration in the preset time period is negatively correlated；
First secure data and first distance and the speed of the second vehicle are positively correlated, and second vehicle is in first front side and position
In lane to be transferred to, the first distance is the fore-and-aft distance of second vehicle and first vehicle, first secure data
Also with the speed of first vehicle, maximum braking deceleration and react time delay negative correlation；Second secure data and second away from
From with the positive correlation of the speed of first vehicle, the second distance is the fore-and-aft distance of third vehicle and first vehicle, described the
Three vehicles are at first vehicle rear and are located at the lane to be transferred to, vehicle of second secure data also with the third vehicle
Speed, maximum braking deceleration and reaction time delay are negatively correlated；
At the end of lane-change, the instant return is 100.
4. lane-change control method according to claim 3, which is characterized in that during lane-change, the instant return tool
Body are as follows:
Wherein, [f0,f1] it is transverse acceleration frequency spectrum, [f2,f3] it is longitudinal acceleration frequency spectrum, axIt is first vehicle described pre-
If the transverse acceleration in the period, ayFor longitudinal acceleration of first vehicle in the preset time period, y1It is described
One distance, y2For the second distance, v1For the speed of first vehicle, v2For the speed of second vehicle, v3For the third
The speed of vehicle,
amaxThe maximum braking deceleration of vehicle, τ k are corresponded to for k
The reaction time delay of corresponding vehicle.
5. lane-change control method according to claim 2, which is characterized in that the action evaluation mould that the basis is previously obtained
Type and the current environment data, the scoring that acquisition current time acts lane-change, specifically include:
The current environment is inputted the action evaluation model q, and (s, a), the lane-change for obtaining current time act score function q
(a)；
It is described that highest corresponding lane-change movement of scoring is determined as the selected lane-change at current time and acts, it specifically includes:
According to formulaDetermine the selected lane-change movement aIt is selected。
6. a kind of lane-change control device, which is characterized in that described device, comprising: acquiring unit, scoring unit, determination unit and
Control unit；
The acquiring unit, for obtaining the current environment data of current time controlled vehicle ambient enviroment；
The scoring unit, the action evaluation model being previously obtained for basis and the current environment data, when obtaining current
The scoring acted to lane-change is carved, the lane-change movement includes acceleration and steering angle；
The determination unit, for highest corresponding lane-change movement of scoring to be determined as the selected lane-change at current time and acts；
Described control unit controls the controlled vehicle movement for acting according to the selected lane-change.
7. lane-change control device according to claim 6, which is characterized in that described device, further includes: model training list
Member；The model training unit, is specifically used for:
Lane-change action training collection is obtained, the lane-change action training collection includes multiple groups lane-change data, and every group of lane-change data include one
First vehicle described in when driver controls each lane-change movement of the first vehicle execution and executes lane-change movement during a lane-change
The environmental data of vehicle speed data and its ambient enviroment；
According to the lane-change action training collection, based on convolutional neural networks to training objective functionBe trained, training convergence after, obtain the action evaluation model q (s,
a)；
Wherein, a is the movement of current lane-change, and s is current environment data, and a' is that the lane-change of subsequent time acts, and s' is subsequent time
Environmental data, γ is learning rate, and R (s) is the instant return at current time.
8. lane-change control device according to claim 7, which is characterized in that
During lane-change, the instant return is positively correlated with the first comfortable data and the second comfortable data, the instant return
Also it is positively correlated with the first secure data and/or the second secure data；
Wherein, the variation degree negative of the described first comfortable data and first vehicle transverse acceleration within a preset period of time
It closes；The variation degree of the second comfortable data and the first vehicle longitudinal acceleration in the preset time period is negatively correlated；
First secure data and first distance and the speed of the second vehicle are positively correlated, and second vehicle is in first front side and position
In lane to be transferred to, the first distance is the fore-and-aft distance of second vehicle and first vehicle, first secure data
Also with the speed of first vehicle, maximum braking deceleration and react time delay negative correlation；Second secure data and second away from
From with the positive correlation of the speed of first vehicle, the second distance is the fore-and-aft distance of third vehicle and first vehicle, described the
Three vehicles are at first vehicle rear and are located at the lane to be transferred to, vehicle of second secure data also with the third vehicle
Speed, maximum braking deceleration and reaction time delay are negatively correlated；
At the end of lane-change, the instant return is 100.
9. lane-change control device according to claim 8, which is characterized in that during lane-change, the instant return tool
Body are as follows:
Wherein, [f0,f1] it is transverse acceleration frequency spectrum, [f2,f3] it is longitudinal acceleration frequency spectrum, axIt is first vehicle described pre-
If the transverse acceleration in the period, ayFor longitudinal acceleration of first vehicle in the preset time period, y1It is described
One distance, y2For the second distance, v1For the speed of first vehicle, v2For the speed of second vehicle, v3For the third
The speed of vehicle,
amaxThe maximum braking deceleration of vehicle, τ k are corresponded to for k
The reaction time delay of corresponding vehicle.
10. lane-change control device according to claim 7, which is characterized in that the scoring unit is specifically used for:
The current environment is inputted the action evaluation model q, and (s, a), the lane-change for obtaining current time act score function q
(a)；
It is described that highest corresponding lane-change movement of scoring is determined as the selected lane-change at current time and acts, it specifically includes:
According to formulaDetermine the selected lane-change movement aIt is selected。
11. a kind of computer readable storage medium, which is characterized in that be stored thereon with computer program, the computer program
It is performed, performs the steps of
Obtain the current environment data of current time controlled vehicle ambient enviroment；
According to the action evaluation model and the current environment data being previously obtained, acquisition current time comments lane-change movement
Point, the lane-change movement includes acceleration and steering angle；
Highest corresponding lane-change movement of scoring is determined as the selected lane-change movement at current time；
It is acted according to the selected lane-change, controls the controlled vehicle movement.
12. a kind of entire car controller characterized by comprising memory and processor；
The memory, for being stored with computer program, the computer program can be realized when being executed by the processor
Following steps:
Obtain the current environment data of current time controlled vehicle ambient enviroment；
According to the action evaluation model and the current environment data being previously obtained, acquisition current time comments lane-change movement
Point, the lane-change movement includes acceleration and steering angle；
Highest corresponding lane-change movement of scoring is determined as the selected lane-change movement at current time；
It is acted according to the selected lane-change, controls the controlled vehicle movement.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711064941.1A CN109760681B (en) | 2017-11-02 | 2017-11-02 | Channel changing control method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711064941.1A CN109760681B (en) | 2017-11-02 | 2017-11-02 | Channel changing control method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109760681A true CN109760681A (en) | 2019-05-17 |
| CN109760681B CN109760681B (en) | 2020-09-29 |
Family
ID=66448667
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711064941.1A Active CN109760681B (en) | 2017-11-02 | 2017-11-02 | Channel changing control method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109760681B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111081046A (en) * | 2020-01-03 | 2020-04-28 | 北京百度网讯科技有限公司 | Vehicle lane changing method, device, electronic equipment and medium |
| CN112124310A (en) * | 2019-06-24 | 2020-12-25 | 初速度（苏州）科技有限公司 | Vehicle path transformation method and device |
| CN112672942A (en) * | 2020-03-23 | 2021-04-16 | 华为技术有限公司 | Vehicle lane changing method and related equipment |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7363140B2 (en) * | 2004-06-17 | 2008-04-22 | Robert Bosch Gmbh | Lane changing assistant for motor vehicles |
| CN106874597A (en) * | 2017-02-16 | 2017-06-20 | 北理慧动（常熟）车辆科技有限公司 | A kind of highway passing behavior decision-making technique for being applied to automatic driving vehicle |
| CN107134173A (en) * | 2017-06-15 | 2017-09-05 | 长安大学 | A kind of lane-change early warning system and method recognized with driving habit |
| CN107161155A (en) * | 2017-04-27 | 2017-09-15 | 大连理工大学 | A kind of vehicle collaboration lane-change method and its system based on artificial neural network |
| CN107215339A (en) * | 2017-06-26 | 2017-09-29 | 地壳机器人科技有限公司 | The lane-change control method and device of automatic driving vehicle |
-
2017
- 2017-11-02 CN CN201711064941.1A patent/CN109760681B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7363140B2 (en) * | 2004-06-17 | 2008-04-22 | Robert Bosch Gmbh | Lane changing assistant for motor vehicles |
| CN106874597A (en) * | 2017-02-16 | 2017-06-20 | 北理慧动（常熟）车辆科技有限公司 | A kind of highway passing behavior decision-making technique for being applied to automatic driving vehicle |
| CN107161155A (en) * | 2017-04-27 | 2017-09-15 | 大连理工大学 | A kind of vehicle collaboration lane-change method and its system based on artificial neural network |
| CN107134173A (en) * | 2017-06-15 | 2017-09-05 | 长安大学 | A kind of lane-change early warning system and method recognized with driving habit |
| CN107215339A (en) * | 2017-06-26 | 2017-09-29 | 地壳机器人科技有限公司 | The lane-change control method and device of automatic driving vehicle |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112124310A (en) * | 2019-06-24 | 2020-12-25 | 初速度（苏州）科技有限公司 | Vehicle path transformation method and device |
| CN112124310B (en) * | 2019-06-24 | 2021-12-24 | 魔门塔(苏州)科技有限公司 | Vehicle path transformation method and device |
| CN111081046A (en) * | 2020-01-03 | 2020-04-28 | 北京百度网讯科技有限公司 | Vehicle lane changing method, device, electronic equipment and medium |
| CN111081046B (en) * | 2020-01-03 | 2022-01-14 | 阿波罗智能技术(北京)有限公司 | Method, device, electronic equipment and medium for automatically changing lane of driving vehicle |
| CN112672942A (en) * | 2020-03-23 | 2021-04-16 | 华为技术有限公司 | Vehicle lane changing method and related equipment |
| WO2021189210A1 (en) * | 2020-03-23 | 2021-09-30 | 华为技术有限公司 | Vehicle lane changing method and related device |
| CN112672942B (en) * | 2020-03-23 | 2022-01-14 | 华为技术有限公司 | Vehicle lane changing method and related equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109760681B (en) | 2020-09-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113291308B (en) | Vehicle self-learning lane-changing decision-making system and method considering driving behavior characteristics | |
| CN109598934B (en) | Rule and learning model-based method for enabling unmanned vehicle to drive away from high speed | |
| CN107792065B (en) | Method for planning road vehicle track | |
| CN108919795B (en) | Automatic driving automobile lane change decision method and device | |
| US20200189597A1 (en) | Reinforcement learning based approach for sae level-4 automated lane change | |
| EP3133455B1 (en) | System for autonomously or partially autonomously driving a vehicle with a communication module for obtaining additional information from a vehicle driver and corresponding method | |
| JP7020097B2 (en) | Vehicle control device, risk map generator, and program | |
| CN101054092B (en) | Driver workload-based vehicle stability enhancement control | |
| CN105857306A (en) | Vehicle autonomous parking path programming method used for multiple parking scenes | |
| Lefevre et al. | Lane keeping assistance with learning-based driver model and model predictive control | |
| CN106184223A (en) | A kind of automatic Pilot control method, device and automobile | |
| CN108995655A (en) | A kind of driver's driving intention recognition methods and system | |
| CN108944943B (en) | Bend following model based on risk dynamic balance theory | |
| EP3725627A1 (en) | Method and apparatus for generating vehicle control command, and vehicle controller and storage medium | |
| JP7172287B2 (en) | Autonomous driving system | |
| CN108717268A (en) | Automatic Pilot minimum time maneuver control system and its control method based on optimum control and safe distance | |
| CN108137046A (en) | The driving assist system of vehicle | |
| CN109760681A (en) | A kind of lane-change control method and device | |
| CN113147752B (en) | Unmanned method and system | |
| CN107132840B (en) | Cross-country electrically-driven unmanned vehicle longitudinal/transverse/vertical personification cooperative control method | |
| EP3725609B1 (en) | Calibrating method for vehicle anti-collision parameters, vehicle controller and storage medium | |
| CN115158308A (en) | Intelligent vehicle active obstacle avoidance control method and device, storage medium and terminal | |
| WO2019003302A1 (en) | Vehicle control apparatus | |
| CN116118780A (en) | Vehicle obstacle avoidance track planning method, system, vehicle and storage medium | |
| CN114889625A (en) | Method for establishing vehicle following model at road curve based on safe potential field theory |
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 |