CN109948656A - A kind of information processing method, device and storage medium - Google Patents
A kind of information processing method, device and storage medium Download PDFInfo
- Publication number
- CN109948656A CN109948656A CN201910130653.4A CN201910130653A CN109948656A CN 109948656 A CN109948656 A CN 109948656A CN 201910130653 A CN201910130653 A CN 201910130653A CN 109948656 A CN109948656 A CN 109948656A
- Authority
- CN
- China
- Prior art keywords
- lane change
- sample
- feedback samples
- vehicle
- sample set
- 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
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/217—Validation; Performance evaluation; Active pattern learning techniques
- G06F18/2178—Validation; Performance evaluation; Active pattern learning techniques based on feedback of a supervisor
-
- 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
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- 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
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0637—Strategic management or analysis, e.g. setting a goal or target of an organisation; Planning actions based on goals; Analysis or evaluation of effectiveness of goals
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Human Resources & Organizations (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Entrepreneurship & Innovation (AREA)
- Educational Administration (AREA)
- Strategic Management (AREA)
- Economics (AREA)
- Software Systems (AREA)
- Game Theory and Decision Science (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Computing Systems (AREA)
- Biomedical Technology (AREA)
- Development Economics (AREA)
- Molecular Biology (AREA)
- Health & Medical Sciences (AREA)
- Mathematical Physics (AREA)
- General Health & Medical Sciences (AREA)
- Marketing (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Traffic Control Systems (AREA)
Abstract
The embodiment of the present invention proposes information processing method, device and storage medium, wherein the described method includes: obtaining the second sample set according to sample extraction strategy and first sample set;Second sample set is clustered into strategy according to lane change and carries out clustering processing, in the case where obtaining at least two clustering informations, at least two clustering information is compared at least two first data obtained in the practical driving of vehicle, the first feedback samples are obtained according to comparison result;According to first feedback samples, orient and failure scene inconsistent in the practical road test driven of vehicle;According to the failure scene that first feedback samples are oriented, existing vehicle lane change model is modified, lane change selection is carried out according to revised vehicle lane change model.The accuracy rate of lane change can be improved using the embodiment of the present invention.
Description
Technical field
The present invention relates to technical field of information processing more particularly to a kind of information processing methods, device and storage medium.
Background technique
A kind of application scenarios of information processing are in Vehicular automatic driving, in order to realize the movement rule of vehicle automatic running
It draws, needs to evaluate the track of vehicle movement.Vehicle lane change is a ring important in motion profile.In the related technology can not
Automatic positioning tests inconsistent failure scene with real road, that is to say, that cannot effectively position (bad where ging wrong
Example) and chosen.If cannot position where ging wrong, vehicle lane change decision can be made inaccurate, then become based on the vehicle
Road decision carries out lane change accuracy rate when lane change selection and declines, to finally be difficult to ensure the lane change result in practical driving procedure
Feasibility and safety.
Summary of the invention
The embodiment of the present invention provides a kind of information processing method, is asked with solving one or more technologies in the prior art
Topic.
In a first aspect, the embodiment of the invention provides a kind of information processing methods, which comprises
The second sample set is obtained according to sample extraction strategy and first sample set;
Second sample set is clustered into strategy according to lane change and carries out clustering processing, obtains the feelings of at least two clustering informations
Under condition, at least two clustering information is compared at least two first data obtained in the practical driving of vehicle, root
The first feedback samples are obtained according to comparison result;
According to first feedback samples, orient and failure field inconsistent in the practical road test driven of vehicle
Scape;
According to the failure scene that first feedback samples are oriented, existing vehicle lane change model is modified, root
Lane change selection is carried out according to revised vehicle lane change model.
In a kind of embodiment, the second sample set is obtained according to sample extraction strategy and first sample set, comprising:
The neural network for constituting the existing vehicle lane change model is obtained, the neural network includes input layer, middle layer
And output layer;
Whole sample extractions of the corresponding input layer and the output layer are concentrated to come out the first sample, as the
One subsample;
The part sample extraction of the corresponding middle layer is concentrated to come out the first sample, as the second subsample;
Second sample set is obtained according to first subsample and second subsample.
In a kind of embodiment, by obtained in the practical driving of at least two clustering information and vehicle at least two the
One data are compared, and obtain the first feedback samples according to comparison result, comprising:
From at least two first data, the target data not matched that with each clustering information is inquired;
Using the target data as the first feedback samples.
In a kind of embodiment, the method also includes:
Second sample set is clustered into strategy according to lane change and carries out clustering processing, in the case where not obtaining clustering information,
Current scene is the boundary scene in the practical driving of vehicle;
Using the boundary scene as the second feedback samples;
Second feedback samples are added into first feedback samples.
In a kind of embodiment, the method also includes:
First feedback samples are inputted into the existing vehicle lane change model, operation obtains the probability point of feedback samples
Cloth;
Feedback entropy is obtained according to the probability distribution of the feedback samples;
Judge whether to meet the practical road test driven of vehicle according to the feedback entropy.
In a kind of embodiment, the method also includes:
First feedback samples are added into second sample set, third sample set is obtained;
The third sample set is clustered into strategy according to lane change and carries out clustering processing, obtains for characterizing lane change to the left
One subclass, the second subclass for characterizing lane change to the right and the third subclass for characterizing straight trip;
According to first subclass, second subclass and the third subclass, to the composition existing vehicle lane change mould
Each corresponding sub-network in the neural network of type is adjusted.
Second aspect, the embodiment of the invention provides a kind of information processing unit, described device includes:
Sample process unit, for obtaining the second sample set according to sample extraction strategy and first sample set;
First feedback samples processing unit carries out at cluster for second sample set to be clustered strategy according to lane change
Reason will be obtained in the practical driving of at least two clustering information and vehicle in the case where obtaining at least two clustering informations
At least two first data are compared, and obtain the first feedback samples according to comparison result;
Positioning unit, for according to first feedback samples, orient in the practical road test driven of vehicle not
The failure scene being consistent;
Lane change selecting unit, the failure scene for being oriented according to first feedback samples become existing vehicle
Road model is modified, and carries out lane change selection according to revised vehicle lane change model.
In a kind of embodiment, the sample process unit is further used for:
The neural network for constituting the existing vehicle lane change model is obtained, the neural network includes input layer, middle layer
And output layer;
Whole sample extractions of the corresponding input layer and the output layer are concentrated to come out the first sample, as the
One subsample;
The part sample extraction of the corresponding middle layer is concentrated to come out the first sample, as the second subsample;
Second sample set is obtained according to first subsample and second subsample.
In a kind of embodiment, the first feedback samples processing unit is further used for:
From at least two first data, the target data not matched that with each clustering information is inquired;
Using the target data as the first feedback samples.
In a kind of embodiment, described device further include:
Clustering processing unit carries out clustering processing for second sample set to be clustered strategy according to lane change, does not obtain
In the case where clustering information, current scene is the boundary scene in the practical driving of vehicle;
Second feedback samples processing unit, for using the boundary scene as the second feedback samples;
First sample adding unit, for second feedback samples to be added into first feedback samples.
In a kind of embodiment, described device further include:
First arithmetic element, for first feedback samples to be inputted the existing vehicle lane change model, operation is obtained
The probability distribution of feedback samples;
Second arithmetic element, for obtaining feedback entropy according to the probability distribution of the feedback samples;
Judging unit, for judging whether to meet the practical road test driven of vehicle according to the feedback entropy.
In a kind of embodiment, described device further include:
Second sample adding unit obtains for first feedback samples to be added into second sample set
Three sample sets;
Subclassing unit carries out clustering processing for the third sample set to be clustered strategy according to lane change, is used
In the first subclass of characterization lane change to the left, the second subclass for characterizing lane change to the right and the third subclass for characterizing straight trip;
Adjustment unit, for according to first subclass, second subclass and the third subclass, to constitute it is described
There is each corresponding sub-network in the neural network of vehicle lane change model to be adjusted.
The third aspect, the embodiment of the invention provides a kind of information processing unit, the function of described device can be by hard
Part is realized, corresponding software realization can also be executed by hardware.The hardware or software include one or more and above-mentioned function
It can corresponding module.
It include processor and memory in the structure of described device in a possible design, the memory is used for
Storage supports described device to execute the program of any above- mentioned information processing method, the processor is configured to described for executing
The program stored in memory.Described device can also include communication interface, be used for and other equipment or communication.
Fourth aspect, the embodiment of the invention provides a kind of computer readable storage mediums, for storing information processing apparatus
Set computer software instructions used comprising for executing program involved in any above- mentioned information processing method.
A technical solution in above-mentioned technical proposal have the following advantages that or the utility model has the advantages that
The embodiment of the present invention obtains the second sample set according to sample extraction strategy and first sample set, by second sample
Collection clusters strategy according to lane change and carries out clustering processing, and in the case where obtaining at least two clustering informations, described at least two are gathered
Category information is compared at least two first data obtained in the practical driving of vehicle, obtains the first feedback according to comparison result
Sample.According to first feedback samples, orient with failure scene inconsistent in the practical road test driven of vehicle,
According to the failure scene that first feedback samples are oriented, existing vehicle lane change model is modified, after amendment
Vehicle lane change model carry out lane change selection.Since clustering processing is obtained the data ratio in clustering information and the practical driving of vehicle
Right, available first feedback samples (bad example) can orient and the practical road driven of vehicle according to the first feedback samples
Inconsistent failure scene in test, (bad example) where finding problem are simultaneously chosen.Pass through the failure scene oriented
Existing vehicle lane change model is modified, lane change selection is carried out according to revised vehicle lane change model, vehicle change can be improved
The accuracy rate in road.
Above-mentioned general introduction is merely to illustrate that the purpose of book, it is not intended to be limited in any way.Except foregoing description
Schematical aspect, except embodiment and feature, by reference to attached drawing and the following detailed description, the present invention is further
Aspect, embodiment and feature, which will be, to be readily apparent that.
Detailed description of the invention
In the accompanying drawings, unless specified otherwise herein, otherwise indicate the same or similar through the identical appended drawing reference of multiple attached drawings
Component or element.What these attached drawings were not necessarily to scale.It should be understood that these attached drawings depict only according to the present invention
Disclosed some embodiments, and should not serve to limit the scope of the present invention.
Fig. 1 shows the flow chart of information processing method according to an embodiment of the present invention.
Fig. 2 shows the schematic diagrames of screening sample according to an embodiment of the present invention.
Fig. 3 shows the flow chart of information processing method according to an embodiment of the present invention.
Fig. 4 shows the flow chart of information processing method according to an embodiment of the present invention.
Fig. 5 shows the flow chart of information processing method according to an embodiment of the present invention.
Fig. 6 shows the structural block diagram of information processing unit according to an embodiment of the present invention.
Fig. 7 shows the structural block diagram of information processing unit according to an embodiment of the present invention.
Specific embodiment
Hereinafter, certain exemplary embodiments are simply just described.As one skilled in the art will recognize that
Like that, without departing from the spirit or scope of the present invention, described embodiment can be modified by various different modes.
Therefore, attached drawing and description are considered essentially illustrative rather than restrictive.
In the related technology, lane change decision model (such as relationship class model and emulation class based on machine learning and deep learning
Model) in real road test there will necessarily be the scene of failure, because machine learning and deep learning can not be identified effectively
Bad example (bad case), according to handmarking method be suitable for adapter tube class problem (between such as vehicle it is impinging one another, by barrier
The scene of blocking or speed-flow breakdown), being further adapted to can be by rule come clearly identifiable behavior (such as compacting line).
In addition, lane change decision model can not learn bad case be because scene difference or characteristic information expression deficiency cause its
There is deviation in study.Using the following embodiment of the present invention, field can be found by the optimization to lane change decision model automatically
Scape difference and bad case to form feedback samples, and then are not consistent according to feedback samples automatic positioning with real road test
The failure scene of conjunction continues to improve and promoted the accuracy using model lane change according to failure scene.
Fig. 1 shows the flow chart of information processing method according to an embodiment of the present invention.As shown in Figure 1, the process includes:
Step 101 obtains the second sample set according to sample extraction strategy and first sample set.
Second sample set is clustered strategy progress clustering processing according to lane change by step 102, obtains at least two clusters
In the case where information, at least two first data obtained in the practical driving of at least two clustering information and vehicle are carried out
It compares, the first feedback samples is obtained according to comparison result.
Step 103, according to first feedback samples, orient with it is inconsistent in the practical road test driven of vehicle
Failure scene.
Step 104, the failure scene oriented according to first feedback samples carry out existing vehicle lane change model
Amendment carries out lane change selection according to revised vehicle lane change model.
In one embodiment, it is contemplated that the first scene information obtained in true man's driving procedure can only reflect the true field of the mankind
Scape, and the second scene information obtained in machine analogue simulation driving procedure can only reflect Dynamic Simulation Results, can both not
Represent true drive, it is impossible to 100% all data of covering, therefore, it is necessary to which the two is merged, then first sample set, by acquiring
Second scene letter obtained in first scene information obtained in true man's driving procedure and acquisition machine analogue simulation driving procedure
Breath is constituted.The possible inaccuracy of first sample set that the first scene information and the second scene information are constituted is merged, because imitative
It is perceived in true simulation and is originally likely to be inaccurate, for example, the sensing module in possible analogue simulation itself has exception, if any
Noise, that possible noise just will affect sensing results, lead to perception inaccuracy.For another example, in emulation, barrier is far to be seen
It arrives, can be in the true world, the sensor perception on far barrier vehicle is less than this part and practical Driving Scene are not
It is consistent, needs to filter out this part scene, that is to say, that for perception inaccuracy, requires to filter out, to match reality
It is relatively accurate to filter out the first sample rally obtained after these practical Driving Scenes of mismatch for Driving Scene.
It is that the actual travel situation to vehicle on traveling lane carries out Image Acquisition for the first scene information, according to
The first scene information that collection result obtains is to be acquired as unit of every frame and travelled obtained letter on a vehicle lane in office
Breath.Specifically, the information in addition to include the environmental informations of vehicle-surroundings, route or travel by vehicle, obstacle information (either statically or dynamically
Obstacle information), further include the driving behavior of professional driver.Wherein, feature is carried out for the driving behavior in driving procedure
Available characteristic information is extracted, for example, acceleration, speed, speed limit, corresponding timestamp and location information etc..To such as acceleration
The characteristic informations such as degree, speed, speed limit, corresponding timestamp and location information tag after being identified, obtain such as Zuo Biandao, the right side
The label information of lane change or straight trip.First scene information includes at least the environmental information of vehicle-surroundings, route or travel by vehicle, obstacle
Object information (obstacle information either statically or dynamically) and the obtained characteristic information of driving behavior and mark by analyzing professional driver
Sign information.Here speed limit is that current vehicle (such as main vehicle) discovery has other vehicles (as before on current vehicle travel route
Vehicle), if the speed of main vehicle be 80 step, in order to avoid collision, need speed limit be 50 step, with pull open between main vehicle and front truck away from
From, it is ensured that mutual driving safety avoids knocking into the back.
It is to carry out image after travel situations to vehicle on traveling lane carry out analog simulation for the second scene information
Acquisition, obtains the second scene information according to collection result, including at least the environmental information of vehicle-surroundings, route or travel by vehicle, barrier
Hinder object information (obstacle information either statically or dynamically), needs to utilize the obtained feature of driving behavior by analyzing professional driver
Information and label information, this feature information and label information are also covered by the second scene information as a result, and analog simulation is to be based on
These information simulate the driving behavior of professional driver, and seeing can export which data obtained, since one be true man's driving, one
It is the emulation that true man are driven, there is some difference for the two.
Filtering out the first sample set obtained after the practical Driving Scene of above-mentioned mismatch can even if relatively accurate
Still and practical Driving Scene there are deviations.For example, cannot but orient the failure scene in time when there is failure scene, this is just
It is easy to appear deviation, this failure scene needs is found in time, first according to sample extraction strategy and above-mentioned first sample set
Obtain the second sample set.Specifically, obtaining the neural network for constituting existing vehicle lane change model, the neural network includes input
Layer, middle layer and output layer.As shown in Fig. 2, Fig. 2 is the schematic diagram of screening sample, inputted corresponding in the first sample set 11
Whole sample extractions of layer 121 and output layer 123 come out, as the first subsample 131, in corresponding in first sample set 131
The part sample extraction of interbed 122 comes out, as the second subsample 132.According to the first subsample 131 and the second subsample 132
Obtain the second sample set 14.Second sample set is clustered into strategy according to lane change and carries out clustering processing, obtains at least two clusters
In the case where information, at least two clustering informations and at least two first data obtained in the practical driving of vehicle are compared
It is right, the first feedback samples are obtained according to comparison result.According to the first feedback samples, orient and the practical road drive test driven of vehicle
Inconsistent failure scene in examination.The failure scene oriented according to the first feedback samples, to existing vehicle lane change model
It is modified, lane change selection is carried out according to revised vehicle lane change model.Using the embodiment of the present invention, due to can be according to anti-
Feedback sample automatic positioning tests inconsistent failure scene with real road, continues to improve and promoted utilization according to failure scene
The accuracy of model lane change.
Fig. 3 shows the flow chart of information processing method according to an embodiment of the present invention.As shown in figure 3, the process includes:
Step 201 obtains the second sample set according to sample extraction strategy and first sample set.
Second sample set is clustered strategy progress clustering processing according to lane change by step 202, obtains at least two clustering informations
In the case where, at least two clustering informations are compared at least two first data obtained in the practical driving of vehicle, from
In at least two first data, the target data not matched that with each clustering information is inquired, using target data as first
Feedback samples.
In one example, according to the second sample set, respectively straight trip, lane change to the left, to the right carry out in lane change three categories (from
Line) small categorical clusters, to obtain cluster result, specific clustering method can choose K-means or hierarchical clustering etc..It can also be right
Sample in second sample set is numbered, in sample number, input feature vector layer, crucial middle layer output, output probability layer, vehicle
Association index is established between road direction, subclass number.
In one example, (such as existed according to the first data of clustering information (such as offline clustering information) and vehicle actual travel
The data of line) it is compared, to obtain the subordinate relation of the second sample intensive data and existing classification, and it is not belonging to any son
The abnormal new data of class.The first data according to clustering information (such as offline clustering information) and actual travel are (such as online
Data) purpose that is compared is: it sees whether some scene meets the clustering information that cluster obtains, includes at least 2 judgements
Branch: first, if meeting clustering information, preferably example, good example are not used as feeding back;Second, if not meeting clustering information,
It is used as feedback for bad case, bad case.By comparing bad case become reconciled example and feature space index pass through clustering algorithm
On-line prediction part is not present distinguishing sample type or boundary scene or other situations.Using the bad case as feedback
Sample, then further rise and be trained again with the data that have in the second sample set before.
Step 203, according to the first feedback samples, orient and mistake inconsistent in the practical road test driven of vehicle
Imitate scene.
Step 204, the failure scene oriented according to the first feedback samples, repair existing vehicle lane change model
Just, lane change selection is carried out according to revised vehicle lane change model.
In one example, as shown in Fig. 2, the whole samples that input layer 121 and output layer 123 will be corresponded in the first sample set 11
Originally it extracts, as the first subsample 131, the part sample extraction that middle layer 122 is corresponded in first sample set 131 is gone out
Come, as the second subsample 132.The second sample set 14 is obtained according to the first subsample 131 and the second subsample 132.According to this
Second sample set carries out multiple first data obtained in the practical driving of multiple clustering informations and vehicle after carrying out clustering processing
It compares, including in multiple first data further includes the data not matched that with clustering information, the number that will be matched with clustering information
According to as target data, using target data as the first feedback samples.According to the first feedback samples, orient and vehicle is practical drives
Inconsistent failure scene in the road test sailed.The failure scene oriented according to the first feedback samples, to existing vehicle
Lane change model is modified, and carries out lane change selection according to revised vehicle lane change model.
Fig. 4 shows the flow chart of information processing method according to an embodiment of the present invention.As shown in figure 4, the process includes:
Step 301 obtains the second sample set according to sample extraction strategy and first sample set.
Second sample set is clustered strategy progress clustering processing according to lane change by step 302, obtains at least two clustering informations
In the case where, at least two clustering informations are compared at least two first data obtained in the practical driving of vehicle, from
In at least two first data, the target data not matched that with each clustering information is inquired, using target data as first
Feedback samples.
Second sample set is clustered the case where strategy carries out clustering processing, do not obtain clustering information according to lane change by step 303
Under, current scene is the boundary scene in the practical driving of vehicle, using the boundary scene as the second feedback samples.
Step 304, according to the first feedback samples and the second feedback samples, orient and the practical road test driven of vehicle
In inconsistent failure scene.
It for above-mentioned cluster and obtains in an example of more than one feedback samples, according to the second sample set, respectively straight
Row, carries out to the right (offline) small categorical clusters in lane change three categories at lane change to the left, to obtain cluster result, specific clustering method
It can choose K-means or hierarchical clustering etc..Sample in second sample set can also be numbered, in sample number, defeated
Enter and establishes association index between characteristic layer, crucial middle layer output, output probability layer, track direction, subclass number.Believed according to cluster
Breath (such as offline clustering information) and the first data (such as online data) of vehicle actual travel are compared, to obtain second
The subordinate relation of sample intensive data and existing classification, and it is not belonging to the abnormal new data of any subclass.According to clustering information
The purpose that the first data (such as online data) of (such as offline clustering information) and actual travel are compared is: seeing some
Whether scene meets the clustering information that cluster obtains, and includes at least 2 and judges branch: first, if meeting clustering information,
Preferably example, good example are not used as feeding back;Second, being used as feedback if not meeting clustering information for bad case, bad case.
Further, a bad case special case is that have scene there be not the case where clustering information, usually corresponds to boundary scene.For this
Situation, and feedback can be used as.Example and feature space index are become reconciled by clustering algorithm on-line prediction by comparing bad case
Part is not present distinguishing sample type or boundary scene or other situations.By the bad case and/or bad case
Special case is all used as feedback samples, and feedback samples are inputted in the second sample set.Due in the second sample set with data can not
Positioning failure scene, cannot cover true man's Driving Scene comprehensively, and therefore, the feedback samples by inputting the second sample set are counted
According to repair and polishing after as new training data, new training data is used for model (for example, to existing vehicle lane change mould
Type) optimization, it can export to obtain more accurate processing result, so that the lane change carried out according to the vehicle lane change model after optimization
Accuracy can greatly improve.
Step 305, the failure scene oriented according to the first feedback samples and the second feedback samples become existing vehicle
Road model is modified, and carries out lane change selection according to revised vehicle lane change model.
Fig. 5 shows the flow chart of information processing method according to an embodiment of the present invention.As shown in figure 5, the process includes:
Step 401 obtains the second sample set according to sample extraction strategy and first sample set.
Second sample set is clustered strategy progress clustering processing according to lane change by step 402, obtains at least two clustering informations
In the case where, at least two clustering informations are compared at least two first data obtained in the practical driving of vehicle, from
In at least two first data, the target data not matched that with each clustering information is inquired, using target data as first
Feedback samples.
Step 403, according to the first feedback samples, orient and mistake inconsistent in the practical road test driven of vehicle
Imitate scene.
Step 404, the failure scene oriented according to the first feedback samples, repair existing vehicle lane change model
Just, lane change selection is carried out according to revised vehicle lane change model.
First feedback samples are inputted existing vehicle lane change model by step 405, and operation obtains the probability point of feedback samples
Cloth.
Step 406 obtains feedback entropy according to the probability distribution of feedback samples, is judged whether to meet vehicle reality according to feedback entropy
The road test that border drives.
In one example, having vehicle lane change model can be a basic decision model, and the first feedback samples are inputted
The decision model simultaneously optimizes, and improves modelling effect.Specifically, the first feedback samples are inputted in the decision model, operation
The probability distribution of feedback samples is obtained, feedback entropy is obtained according to the probability distribution of feedback samples, entropy is bigger to have illustrated that example is more preferable.
In one embodiment, different sub-network structures can also be adjusted by clustering obtained different clustering informations.Tool
Body, the first feedback samples are added into second sample set, third sample set is obtained.By third sample set according to lane change
Cluster strategy carries out clustering processing, obtains the first subclass for characterizing lane change to the left, the second son for characterizing lane change to the right
Class and for characterize straight trip third subclass.According to the first subclass, the second subclass and third subclass, has vehicle lane change to constituting
Each corresponding sub-network in the neural network of model is adjusted.The same neural network and data are by all lane models
(relationship class model, emulation class model and decision class model etc.) is shared.
Fig. 6 shows a kind of structural block diagram of information processing unit, and described device includes: sample process unit 21, is used for root
The second sample set is obtained according to sample extraction strategy and first sample set;First feedback samples processing unit 22, for by described the
Two sample sets cluster strategy according to lane change and carry out clustering processing, in the case where obtaining at least two clustering informations, by described at least
Two clustering informations and vehicle are practical drive obtained at least two first data be compared, obtain the according to comparison result
One feedback samples;Positioning unit 23, for orienting and the practical road test driven of vehicle according to first feedback samples
In inconsistent failure scene;Lane change selecting unit 24, the failure field for being oriented according to first feedback samples
Scape is modified existing vehicle lane change model, carries out lane change selection according to revised vehicle lane change model.
In one embodiment, the sample process unit is further used for: obtaining and constitutes the existing vehicle lane change model
Neural network, the neural network includes input layer, middle layer and output layer;Concentrate correspondence described defeated the first sample
The whole sample extractions for entering layer and the output layer come out, as the first subsample;The first sample is concentrated described in corresponding to
The part sample extraction of middle layer comes out, as the second subsample;It is obtained according to first subsample and second subsample
To second sample set.
In one embodiment, the first feedback samples processing unit is further used for: from least two first number
In, the target data not matched that with each clustering information is inquired;Using the target data as the first feedback samples.
In one embodiment, described device further include: clustering processing unit is used for second sample set according to lane change
Cluster strategy carries out clustering processing, and in the case where not obtaining clustering information, current scene is the boundary field in the practical driving of vehicle
Scape;Second feedback samples processing unit, for using the boundary scene as the second feedback samples;First sample adding unit,
For second feedback samples to be added into first feedback samples.
In one embodiment, described device further include: the first arithmetic element, for first feedback samples to be inputted institute
Existing vehicle lane change model is stated, operation obtains the probability distribution of feedback samples;Second arithmetic element, for according to the feedback sample
This probability distribution obtains feedback entropy;Judging unit, for judging whether to meet according to the feedback entropy, vehicle is practical drives
Road test.
In one embodiment, described device further include: the second sample adding unit, for first feedback samples to be added
It is added in second sample set, obtains third sample set;Subclassing unit is used for the third sample set according to lane change
Cluster strategy carries out clustering processing, obtains the first subclass for characterizing lane change to the left, the second son for characterizing lane change to the right
Class and for characterize straight trip third subclass;Adjustment unit, for according to first subclass, second subclass and described the
Three subclasses are adjusted each corresponding sub-network in the neural network for constituting the existing vehicle lane change model.
The function of each module in each device of the embodiment of the present invention may refer to the corresponding description in the above method, herein not
It repeats again.
Fig. 7 shows the structural block diagram of information processing unit according to an embodiment of the present invention.As shown in fig. 7, the device includes:
Memory 910 and processor 920 are stored with the computer program that can be run on processor 920 in memory 910.Processor
The automatic Pilot method in above-described embodiment is realized when 920 execution computer program.The quantity of memory 910 and processor 920
It can be one or more.
The device further include: communication interface 930 carries out data interaction for being communicated with external device.
Memory 910 may include high speed RAM memory, it is also possible to further include nonvolatile memory (non-
Volatile memory), a for example, at least magnetic disk storage.
If memory 910, processor 920 and the independent realization of communication interface 930, memory 910,920 and of processor
Communication interface 930 can be connected with each other by bus and complete mutual communication.Bus can be industry standard architecture
(ISA, Industry Standard Architecture) bus, external equipment interconnection (PCI, Peripheral
Component) bus or extended industry-standard architecture (EISA, Extended Industry Standard
Component) bus etc..Bus can be divided into address bus, data/address bus, control bus etc..For convenient for indicating, in Fig. 7 only
It is indicated with a thick line, it is not intended that an only bus or a type of bus.
Optionally, in specific implementation, if memory 910, processor 920 and communication interface 930 are integrated in one piece of core
On piece, then memory 910, processor 920 and communication interface 930 can complete mutual communication by internal interface.
The embodiment of the invention provides a kind of computer readable storage mediums, are stored with computer program, the program quilt
Processor realizes any method in above-described embodiment when executing.
In the description of this specification, reference term " one embodiment ", " some embodiments ", " example ", " specifically show
The description of example " or " some examples " etc. means specific features, structure, material or spy described in conjunction with this embodiment or example
Point is included at least one embodiment or example of the invention.Moreover, particular features, structures, materials, or characteristics described
It may be combined in any suitable manner in any one or more of the embodiments or examples.In addition, without conflicting with each other, this
The technical staff in field can be by the spy of different embodiments or examples described in this specification and different embodiments or examples
Sign is combined.
In addition, term " first ", " second " are used for descriptive purposes only and cannot be understood as indicating or suggesting relative importance
Or implicitly indicate the quantity of indicated technical characteristic." first " is defined as a result, the feature of " second " can be expressed or hidden
It include at least one this feature containing ground.In the description of the present invention, the meaning of " plurality " is two or more, unless otherwise
Clear specific restriction.
Any process described otherwise above or method description are construed as in flow chart or herein, and expression includes
It is one or more for realizing specific logical function or process the step of executable instruction code module, segment or portion
Point, and the range of the preferred embodiment of the present invention includes other realization, wherein can not press shown or discussed suitable
Sequence, including according to related function by it is basic simultaneously in the way of or in the opposite order, Lai Zhihang function, this should be of the invention
Embodiment person of ordinary skill in the field understood.
Expression or logic and/or step described otherwise above herein in flow charts, for example, being considered use
In the order list for the executable instruction for realizing logic function, may be embodied in any computer-readable medium, for
Instruction execution system, device or equipment (such as computer based system, including the system of processor or other can be held from instruction
The instruction fetch of row system, device or equipment and the system executed instruction) it uses, or combine these instruction execution systems, device or set
It is standby and use.For the purpose of this specification, " computer-readable medium ", which can be, any may include, stores, communicates, propagates or pass
Defeated program is for instruction execution system, device or equipment or the dress used in conjunction with these instruction execution systems, device or equipment
It sets.The more specific example (non-exhaustive list) of computer-readable medium include the following: there is the electricity of one or more wirings
Interconnecting piece (electronic device), portable computer diskette box (magnetic device), random access memory (RAM), read-only memory
(ROM), erasable edit read-only storage (EPROM or flash memory), fiber device and portable read-only memory
(CDROM).In addition, computer-readable medium can even is that the paper that can print described program on it or other suitable Jie
Matter, because can then be edited, be interpreted or when necessary with other for example by carrying out optical scanner to paper or other media
Suitable method is handled electronically to obtain described program, is then stored in computer storage.
It should be appreciated that each section of the invention can be realized with hardware, software, firmware or their combination.Above-mentioned
In embodiment, software that multiple steps or method can be executed in memory and by suitable instruction execution system with storage
Or firmware is realized.It, and in another embodiment, can be under well known in the art for example, if realized with hardware
Any one of column technology or their combination are realized: having a logic gates for realizing logic function to data-signal
Discrete logic, with suitable combinational logic gate circuit specific integrated circuit, programmable gate array (PGA), scene
Programmable gate array (FPGA) etc..
Those skilled in the art are understood that realize all or part of step that above-described embodiment method carries
It suddenly is that relevant hardware can be instructed to complete by program, the program can store in a kind of computer-readable storage medium
In matter, which when being executed, includes the steps that one or a combination set of embodiment of the method.
It, can also be in addition, each functional unit in each embodiment of the present invention can integrate in a processing module
It is that each unit physically exists alone, can also be integrated in two or more units in a module.Above-mentioned integrated mould
Block both can take the form of hardware realization, can also be realized in the form of software function module.The integrated module is such as
Fruit is realized and when sold or used as an independent product in the form of software function module, also can store in a computer
In readable storage medium storing program for executing.The storage medium can be read-only memory, disk or CD etc..
The above description is merely a specific embodiment, but scope of protection of the present invention is not limited thereto, any
Those familiar with the art in the technical scope disclosed by the present invention, can readily occur in its various change or replacement,
These should be covered by the protection scope of the present invention.Therefore, protection scope of the present invention should be with the guarantor of the claim
It protects subject to range.
Claims (14)
1. a kind of information processing method, which is characterized in that the described method includes:
The second sample set is obtained according to sample extraction strategy and first sample set;
Second sample set is clustered into the case where strategy carries out clustering processing, obtains at least two clustering informations according to lane change
Under, at least two clustering information is compared at least two first data obtained in the practical driving of vehicle, according to
Comparison result obtains the first feedback samples;
According to first feedback samples, orient and failure scene inconsistent in the practical road test driven of vehicle;
According to the failure scene that first feedback samples are oriented, existing vehicle lane change model is modified, according to repairing
Vehicle lane change model after just carries out lane change selection.
2. the method according to claim 1, wherein obtaining second according to sample extraction strategy and first sample set
Sample set, comprising:
Obtain the neural network for constituting the existing vehicle lane change model, the neural network includes input layer, middle layer and defeated
Layer out;
Whole sample extractions of the corresponding input layer and the output layer are concentrated to come out the first sample, as the first son
Sample;
The part sample extraction of the corresponding middle layer is concentrated to come out the first sample, as the second subsample;
Second sample set is obtained according to first subsample and second subsample.
3. the method according to claim 1, wherein by the practical driving of at least two clustering information and vehicle
Obtained at least two first data be compared, the first feedback samples are obtained according to comparison result, comprising:
From at least two first data, the target data not matched that with each clustering information is inquired;
Using the target data as the first feedback samples.
4. the method according to claim 1, wherein the method also includes:
Second sample set is clustered into strategy according to lane change and carries out clustering processing, in the case where not obtaining clustering information, currently
Scene is the boundary scene in the practical driving of vehicle;
Using the boundary scene as the second feedback samples;
Second feedback samples are added into first feedback samples.
5. method according to any one of claims 1 to 4, which is characterized in that the method also includes:
First feedback samples are inputted into the existing vehicle lane change model, operation obtains the probability distribution of feedback samples;
Feedback entropy is obtained according to the probability distribution of the feedback samples;
Judge whether to meet the practical road test driven of vehicle according to the feedback entropy.
6. according to the method described in claim 5, it is characterized in that, the method also includes:
First feedback samples are added into second sample set, third sample set is obtained;
The third sample set is clustered into strategy according to lane change and carries out clustering processing, obtains the first son for characterizing lane change to the left
Class, the second subclass for characterizing lane change to the right and the third subclass for characterizing straight trip;
According to first subclass, second subclass and the third subclass, to the composition existing vehicle lane change model
Each corresponding sub-network in neural network is adjusted.
7. a kind of information processing unit, which is characterized in that described device includes:
Sample process unit, for obtaining the second sample set according to sample extraction strategy and first sample set;
First feedback samples processing unit carries out clustering processing for second sample set to be clustered strategy according to lane change, obtains
It, will be at least two obtained in the practical driving of at least two clustering information and vehicle in the case where at least two clustering informations
A first data are compared, and obtain the first feedback samples according to comparison result;
Positioning unit, for orienting and not being consistent with the practical road test driven of vehicle according to first feedback samples
The failure scene of conjunction;
Lane change selecting unit, the failure scene for being oriented according to first feedback samples, to existing vehicle lane change mould
Type is modified, and carries out lane change selection according to revised vehicle lane change model.
8. device according to claim 7, which is characterized in that the sample process unit is further used for:
Obtain the neural network for constituting the existing vehicle lane change model, the neural network includes input layer, middle layer and defeated
Layer out;
Whole sample extractions of the corresponding input layer and the output layer are concentrated to come out the first sample, as the first son
Sample;
The part sample extraction of the corresponding middle layer is concentrated to come out the first sample, as the second subsample;
Second sample set is obtained according to first subsample and second subsample.
9. device according to claim 7, which is characterized in that the first feedback samples processing unit is further used for:
From at least two first data, the target data not matched that with each clustering information is inquired;
Using the target data as the first feedback samples.
10. device according to claim 7, which is characterized in that described device further include:
Clustering processing unit carries out clustering processing for second sample set to be clustered strategy according to lane change, is not clustered
In the case where information, current scene is the boundary scene in the practical driving of vehicle;
Second feedback samples processing unit, for using the boundary scene as the second feedback samples;
First sample adding unit, for second feedback samples to be added into first feedback samples.
11. according to the described in any item devices of claim 7 to 10, which is characterized in that described device further include:
First arithmetic element, for first feedback samples to be inputted the existing vehicle lane change model, operation is fed back
The probability distribution of sample;
Second arithmetic element, for obtaining feedback entropy according to the probability distribution of the feedback samples;
Judging unit, for judging whether to meet the practical road test driven of vehicle according to the feedback entropy.
12. device according to claim 11, which is characterized in that described device further include:
Second sample adding unit obtains third sample for first feedback samples to be added into second sample set
This collection;
Subclassing unit carries out clustering processing for the third sample set to be clustered strategy according to lane change, obtains for table
Levy the first subclass, the second subclass for characterizing lane change to the right and the third subclass for characterizing straight trip of lane change to the left;
Adjustment unit is used for according to first subclass, second subclass and the third subclass, to the composition existing vehicle
Each corresponding sub-network in the neural network of lane change model is adjusted.
13. a kind of information processing unit, which is characterized in that described device includes:
One or more processors;
Storage device, for storing one or more programs;
When one or more of programs are executed by one or more of processors, so that one or more of processors
Realize such as method described in any one of claims 1 to 6.
14. a kind of computer readable storage medium, is stored with computer program, which is characterized in that the program is held by processor
Such as method described in any one of claims 1 to 6 is realized when row.
Priority Applications (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210485715.5A CN114861793A (en) | 2019-02-21 | 2019-02-21 | Information processing method, device and storage medium |
CN201910130653.4A CN109948656B (en) | 2019-02-21 | 2019-02-21 | Information processing method, device and storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910130653.4A CN109948656B (en) | 2019-02-21 | 2019-02-21 | Information processing method, device and storage medium |
Related Child Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202210485715.5A Division CN114861793A (en) | 2019-02-21 | 2019-02-21 | Information processing method, device and storage medium |
Publications (2)
Publication Number | Publication Date |
---|---|
CN109948656A true CN109948656A (en) | 2019-06-28 |
CN109948656B CN109948656B (en) | 2022-11-25 |
Family
ID=67006910
Family Applications (2)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910130653.4A Active CN109948656B (en) | 2019-02-21 | 2019-02-21 | Information processing method, device and storage medium |
CN202210485715.5A Pending CN114861793A (en) | 2019-02-21 | 2019-02-21 | Information processing method, device and storage medium |
Family Applications After (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202210485715.5A Pending CN114861793A (en) | 2019-02-21 | 2019-02-21 | Information processing method, device and storage medium |
Country Status (1)
Country | Link |
---|---|
CN (2) | CN109948656B (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111783226A (en) * | 2020-06-29 | 2020-10-16 | 北京百度网讯科技有限公司 | Method and device for generating automatic driving scene measurement parameters, electronic equipment and storage medium |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20160152239A1 (en) * | 2013-07-19 | 2016-06-02 | Nissan Motor Co., Ltd | Driving state estimation device |
CN105894359A (en) * | 2016-03-31 | 2016-08-24 | 百度在线网络技术(北京)有限公司 | Order pushing method, device and system |
CN106557963A (en) * | 2015-09-30 | 2017-04-05 | 百度在线网络技术(北京)有限公司 | Process method, device and the server for using car order |
CN109060370A (en) * | 2018-06-29 | 2018-12-21 | 奇瑞汽车股份有限公司 | The method and device of vehicle testing is carried out to automatic driving vehicle |
CN109085837A (en) * | 2018-08-30 | 2018-12-25 | 百度在线网络技术(北京)有限公司 | Control method for vehicle, device, computer equipment and storage medium |
CN109241929A (en) * | 2018-09-20 | 2019-01-18 | 北京海纳川汽车部件股份有限公司 | Method for detecting lane lines, device and the automatic driving vehicle of automatic driving vehicle |
CN109271858A (en) * | 2018-08-13 | 2019-01-25 | 武汉中海庭数据技术有限公司 | A kind of crossing recognition methods and system based on wheelpath and vision lane side line number evidence |
-
2019
- 2019-02-21 CN CN201910130653.4A patent/CN109948656B/en active Active
- 2019-02-21 CN CN202210485715.5A patent/CN114861793A/en active Pending
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20160152239A1 (en) * | 2013-07-19 | 2016-06-02 | Nissan Motor Co., Ltd | Driving state estimation device |
CN106557963A (en) * | 2015-09-30 | 2017-04-05 | 百度在线网络技术(北京)有限公司 | Process method, device and the server for using car order |
CN105894359A (en) * | 2016-03-31 | 2016-08-24 | 百度在线网络技术(北京)有限公司 | Order pushing method, device and system |
CN109060370A (en) * | 2018-06-29 | 2018-12-21 | 奇瑞汽车股份有限公司 | The method and device of vehicle testing is carried out to automatic driving vehicle |
CN109271858A (en) * | 2018-08-13 | 2019-01-25 | 武汉中海庭数据技术有限公司 | A kind of crossing recognition methods and system based on wheelpath and vision lane side line number evidence |
CN109085837A (en) * | 2018-08-30 | 2018-12-25 | 百度在线网络技术(北京)有限公司 | Control method for vehicle, device, computer equipment and storage medium |
CN109241929A (en) * | 2018-09-20 | 2019-01-18 | 北京海纳川汽车部件股份有限公司 | Method for detecting lane lines, device and the automatic driving vehicle of automatic driving vehicle |
Non-Patent Citations (4)
Title |
---|
SHUANG SU等: "预测自动驾驶车辆变道意图", 《HTTPS://CLOUD.TENCENT.COM/DEVELOPER/NEWS/289224》 * |
VOKHIDOV, HUSAN等: "Recognition of Damaged Arrow-Road Markings by Visible Light Camera Sensor Based on Convolutional Neural Network.", 《SENSORS (BASEL, SWITZERLAND)》 * |
贾会群等: "基于神经网络与最小二乘法的车道线检测算法研究", 《汽车工程》 * |
黄鑫等: "监控视频中的车辆异常行为检测", 《计算机***应用》 * |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111783226A (en) * | 2020-06-29 | 2020-10-16 | 北京百度网讯科技有限公司 | Method and device for generating automatic driving scene measurement parameters, electronic equipment and storage medium |
CN111783226B (en) * | 2020-06-29 | 2024-03-19 | 北京百度网讯科技有限公司 | Method, device, electronic equipment and storage medium for generating automatic driving scene measurement parameters |
Also Published As
Publication number | Publication date |
---|---|
CN114861793A (en) | 2022-08-05 |
CN109948656B (en) | 2022-11-25 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
DE102017213217A1 (en) | Test scenario database system for realistic virtual test driving scenarios | |
US10699167B1 (en) | Perception visualization tool | |
DE102019102205A1 (en) | SYSTEM AND METHOD FOR THE END TO END VALIDATION OF AUTONOMOUS VEHICLES | |
CN108319907A (en) | A kind of vehicle identification method, device and storage medium | |
CN109017797A (en) | Driver's Emotion identification method and the vehicular control unit for implementing this method | |
DE102018100192A1 (en) | Generation of training data for automatic leak detection in vehicles | |
DE102014206623A1 (en) | Localization of a head-mounted display (HMD) in the vehicle | |
KR102481817B1 (en) | Method for Estimating Driver's Gaze Zone by Transfer Learning | |
WO2017021318A1 (en) | Apparatus, vehicle, method and computer program for computing at least one video signal or control signal | |
EP3392111A1 (en) | Method and device for determining a visual deflection of a driver of a vehicle | |
CN110928620B (en) | Evaluation method and system for driving distraction caused by automobile HMI design | |
DE102019208788A1 (en) | Motor vehicle | |
CN109703465A (en) | The control method and device of vehicle-mounted imaging sensor | |
CN113327479A (en) | Motor vehicle driving intelligent training system based on MR technology | |
CN109886198B (en) | Information processing method, device and storage medium | |
CN110930811B (en) | System suitable for unmanned decision learning and training | |
del Egio et al. | Self-driving a car in simulation through a CNN | |
CN109948656A (en) | A kind of information processing method, device and storage medium | |
DE102021201445A1 (en) | Computer-implemented method for testing conformance between real and synthetic images for machine learning | |
CN109948449A (en) | A kind of information processing method, device and storage medium | |
DE102018219125A1 (en) | Method for classifying objects using an automated motor vehicle and automated motor vehicle | |
DE102021213538A1 (en) | Simulation to validate an automated driving function for a vehicle | |
EP3754544A1 (en) | Detection system, work method and training method | |
DE102013001747A1 (en) | Method for providing a rating of drive along traffic route traveled by learner-driver of motor car, involves detecting route data and evaluation data along busy route so as to assign the data to current geographical position of car | |
DE102019204348A1 (en) | Method for route guidance to a target person, electronic device of the target person and electronic device of the person collecting it, as well as a motor vehicle |
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 |