CN109948656B - Information processing method, device and storage medium - Google Patents
Information processing method, device and storage medium Download PDFInfo
- Publication number
- CN109948656B CN109948656B CN201910130653.4A CN201910130653A CN109948656B CN 109948656 B CN109948656 B CN 109948656B CN 201910130653 A CN201910130653 A CN 201910130653A CN 109948656 B CN109948656 B CN 109948656B
- Authority
- CN
- China
- Prior art keywords
- sample
- lane change
- feedback
- clustering
- vehicle
- 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.)
- Active
Links
Images
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 invention provides an information processing method, an information processing device and a storage medium, wherein the method comprises the following steps: obtaining a second sample set according to the sample extraction strategy and the first sample set; clustering the second sample set according to a lane change clustering strategy to obtain at least two pieces of clustering information, comparing the at least two pieces of clustering information with at least two pieces of first data obtained in actual driving of the vehicle, and obtaining a first feedback sample according to a comparison result; according to the first feedback sample, positioning a failure scene which is not in accordance with the actual driving road test of the vehicle; and correcting the existing vehicle lane change model according to the failure scene positioned by the first feedback sample, and performing lane change selection according to the corrected vehicle lane change model. The method and the device can improve the accuracy of lane changing.
Description
Technical Field
The present invention relates to the field of information processing technologies, and in particular, to an information processing method and apparatus, and a storage medium.
Background
One application scenario of information processing is in vehicle automatic driving, and in order to implement a motion plan for vehicle automatic driving, a trajectory of vehicle motion needs to be evaluated. Vehicle lane change is an important loop in the motion trajectory. In the related art, failure scenes which do not conform to actual road tests cannot be automatically located, that is, the problem (bad case) cannot be effectively located and picked out. If the problem can not be located, the lane change decision of the vehicle is inaccurate, the lane change accuracy rate is reduced when the lane change selection is carried out based on the lane change decision of the vehicle, and finally the feasibility and the safety of the lane change result in the actual driving process are difficult to guarantee.
Disclosure of Invention
Embodiments of the present invention provide an information processing method to solve one or more technical problems in the prior art.
In a first aspect, an embodiment of the present invention provides an information processing method, where the method includes:
obtaining a second sample set according to the sample extraction strategy and the first sample set;
clustering the second sample set according to a lane change clustering strategy to obtain at least two pieces of clustering information, comparing the at least two pieces of clustering information with at least two pieces of first data obtained in actual driving of a vehicle, and obtaining a first feedback sample according to a comparison result;
according to the first feedback sample, positioning a failure scene which is not in accordance with the actual driving road test of the vehicle;
and correcting the existing vehicle lane change model according to the failure scene positioned by the first feedback sample, and performing lane change selection according to the corrected vehicle lane change model.
In one embodiment, obtaining the second set of samples from the sample extraction strategy and the first set of samples comprises:
obtaining a neural network forming the existing vehicle lane change model, wherein the neural network comprises an input layer, a middle layer and an output layer;
extracting all samples corresponding to the input layer and the output layer in the first sample set to be used as first subsamples;
extracting a part of samples corresponding to the middle layer in the first sample set to serve as second subsamples;
obtaining the second set of samples from the first subsample and the second subsample.
In one embodiment, comparing the at least two pieces of cluster information with at least two pieces of first data obtained during actual driving of the vehicle, and obtaining a first feedback sample according to a comparison result includes:
inquiring target data which are not matched with each clustering information from the at least two pieces of first data;
the target data is taken as a first feedback sample.
In one embodiment, the method further comprises:
clustering the second sample set according to a lane change clustering strategy, wherein the current scene is a boundary scene in actual driving of the vehicle under the condition that clustering information is not obtained;
taking the boundary scene as a second feedback sample;
adding the second feedback sample to the first feedback sample.
In one embodiment, the method further comprises:
inputting the first feedback sample into the existing vehicle lane change model, and calculating to obtain the probability distribution of the feedback sample;
obtaining a feedback entropy according to the probability distribution of the feedback sample;
and judging whether the road test of the actual driving of the vehicle is met or not according to the feedback entropy.
In one embodiment, the method further comprises:
adding the first feedback sample into the second sample set to obtain a third sample set;
clustering the third sample set according to a lane change clustering strategy to obtain a first subclass for representing lane change to the left, a second subclass for representing lane change to the right and a third subclass for representing straight line;
and adjusting each corresponding sub-network in the neural network forming the existing vehicle lane change model according to the first sub-class, the second sub-class and the third sub-class.
In a second aspect, an embodiment of the present invention provides an information processing apparatus, including:
the sample processing unit is used for obtaining a second sample set according to the sample extraction strategy and the first sample set;
the first feedback sample processing unit is used for carrying out clustering processing on the second sample set according to a lane change clustering strategy to obtain at least two clustering information, comparing the at least two clustering information with at least two first data obtained in actual driving of a vehicle, and obtaining a first feedback sample according to a comparison result;
the positioning unit is used for positioning a failure scene which is not in accordance with the actual driving road test of the vehicle according to the first feedback sample;
and the lane change selection unit is used for correcting the existing vehicle lane change model according to the failure scene positioned by the first feedback sample and performing lane change selection according to the corrected vehicle lane change model.
In one embodiment, the sample processing unit is further configured to:
obtaining a neural network forming the existing vehicle lane change model, wherein the neural network comprises an input layer, a middle layer and an output layer;
extracting all samples corresponding to the input layer and the output layer in the first sample set to be used as first subsamples;
extracting a part of samples corresponding to the middle layer in the first sample set to serve as second subsamples;
obtaining the second set of samples from the first subsample and the second subsample.
In one embodiment, the first feedback sample processing unit is further configured to:
inquiring target data which are not matched with each clustering information from the at least two pieces of first data;
the target data is taken as a first feedback sample.
In one embodiment, the apparatus further comprises:
the clustering processing unit is used for clustering the second sample set according to a lane change clustering strategy, and under the condition that clustering information is not obtained, the current scene is a boundary scene in the actual driving of the vehicle;
a second feedback sample processing unit, configured to take the boundary scene as a second feedback sample;
a first sample adding unit for adding the second feedback sample to the first feedback sample.
In one embodiment, the apparatus further comprises:
the first operation unit is used for inputting the first feedback sample into the existing vehicle lane change model and obtaining the probability distribution of the feedback sample through operation;
the second operation unit is used for obtaining a feedback entropy according to the probability distribution of the feedback sample;
and the judging unit is used for judging whether the road test according with the actual driving of the vehicle is met according to the feedback entropy.
In one embodiment, the apparatus further comprises:
a second sample adding unit, configured to add the first feedback sample into the second sample set to obtain a third sample set;
the subclass processing unit is used for carrying out clustering processing on the third sample set according to a lane changing clustering strategy to obtain a first subclass for representing lane changing to the left, a second subclass for representing lane changing to the right and a third subclass for representing straight lines;
and the adjusting unit is used for adjusting each corresponding sub-network in the neural network forming the existing vehicle lane-changing model according to the first sub-class, the second sub-class and the third sub-class.
In a third aspect, an embodiment of the present invention provides an information processing apparatus, where functions of the apparatus may be implemented by hardware, or may be implemented by hardware executing corresponding software. The hardware or software includes one or more modules corresponding to the functions described above.
In one possible design, the structure of the apparatus includes a processor and a memory, the memory is used for storing a program supporting the apparatus to execute any one of the above information processing methods, and the processor is configured to execute the program stored in the memory. The apparatus may also include a communication interface for communicating with other devices or a communication network.
In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium for storing computer software instructions for an information processing apparatus, which includes a program for executing any one of the above-described information processing methods.
One of the above technical solutions has the following advantages or beneficial effects:
according to the embodiment of the invention, a second sample set is obtained according to a sample extraction strategy and a first sample set, the second sample set is subjected to clustering processing according to a lane change clustering strategy, under the condition of obtaining at least two pieces of clustering information, the at least two pieces of clustering information are compared with at least two pieces of first data obtained in the actual driving of a vehicle, and a first feedback sample is obtained according to the comparison result. And positioning a failure scene which is not in accordance with the road test of the actual driving of the vehicle according to the first feedback sample, correcting the existing vehicle lane change model according to the failure scene positioned by the first feedback sample, and performing lane change selection according to the corrected vehicle lane change model. The clustering information obtained by clustering processing is compared with the data in the actual driving of the vehicle, so that a first feedback sample (bad case) can be obtained, failure scenes which do not accord with the road test of the actual driving of the vehicle can be positioned according to the first feedback sample, and the problem (bad case) is found and picked out. The existing vehicle lane change model is corrected through the positioned failure scene, and lane change selection is performed according to the corrected vehicle lane change model, so that the accuracy rate of vehicle lane change can be improved.
The foregoing summary is provided for the purpose of description only and is not intended to be limiting in any way. In addition to the illustrative aspects, embodiments, and features described above, further aspects, embodiments, and features of the present invention will be readily apparent by reference to the drawings and following detailed description.
Drawings
In the drawings, like reference numerals refer to the same or similar parts or elements throughout the several views unless otherwise specified. The figures are not necessarily to scale. It is appreciated that these drawings depict only some embodiments in accordance with the disclosure and are therefore not to be considered limiting of its scope.
Fig. 1 shows a flowchart of an information processing method according to an embodiment of the present invention.
FIG. 2 shows a schematic diagram of sample screening according to an embodiment of the invention.
Fig. 3 shows a flowchart of an information processing method according to an embodiment of the present invention.
Fig. 4 shows a flowchart of an information processing method according to an embodiment of the present invention.
Fig. 5 illustrates a flowchart of an information processing method according to an embodiment of the present invention.
Fig. 6 shows a block diagram of the structure of an information processing apparatus according to an embodiment of the present invention.
Fig. 7 shows a block diagram of the structure of an information processing apparatus according to an embodiment of the present invention.
Detailed Description
In the following, only certain exemplary embodiments are briefly described. As those skilled in the art will recognize, the described embodiments may be modified in various different ways, all without departing from the spirit or scope of the present invention. Accordingly, the drawings and description are to be regarded as illustrative in nature, and not as restrictive.
In the related art, lane change decision models (such as a relation model and a simulation model) based on machine learning and deep learning inevitably have failure scenes in an actual road test, because both the machine learning and the deep learning cannot effectively identify bad cases (bad cases), if a manual marking method is adopted, the method is suitable for taking over the problems (such as the scenes that vehicles collide with each other, are blocked by obstacles or are not in a stagnation state), and is also suitable for defining identifiable behaviors (such as pressing solid lines and the like) through rules. In addition, the lane change decision model cannot know whether bad cases are due to scene differences or insufficient feature information expression, so that the bad cases have deviation in learning. By adopting the following embodiment of the invention, the scene difference and bad case can be automatically found through optimizing the lane change decision model to form the feedback sample, so that the failure scene inconsistent with the actual road test can be automatically positioned according to the feedback sample, and the lane change accuracy of the model can be continuously improved and improved according to the failure scene.
Fig. 1 shows a flowchart of an information processing method according to an embodiment of the present invention. As shown in fig. 1, the process includes:
and 101, obtaining a second sample set according to the sample extraction strategy and the first sample set.
And 102, carrying out clustering processing on the second sample set according to a lane change clustering strategy to obtain at least two pieces of clustering information, comparing the at least two pieces of clustering information with at least two pieces of first data obtained in actual driving of the vehicle, and obtaining a first feedback sample according to a comparison result.
And 103, positioning a failure scene which is not consistent with the road test of the actual driving of the vehicle according to the first feedback sample.
And 104, correcting the existing vehicle lane change model according to the failure scene positioned by the first feedback sample, and performing lane change selection according to the corrected vehicle lane change model.
In one embodiment, considering that first scene information obtained in the process of real-person driving can only reflect a human real scene, and second scene information obtained in the process of machine simulation driving can only reflect a dynamic simulation result, both of which do not represent real driving and cannot cover all data by 100%, the first scene information and the second scene information need to be fused, and the first sample set is formed by collecting the first scene information obtained in the process of real-person driving and the second scene information obtained in the process of collecting machine simulation driving. The first sample set formed by fusing the first scene information and the second scene information may be inaccurate, since the sensing in the simulation may be inaccurate, for example, the sensing module in the simulation may have an anomaly, and if there is noise, the noise may affect the sensing result, resulting in inaccurate sensing. For another example, in the simulation, the obstacle can be seen far away, the sensor on the far-away obstacle vehicle cannot sense the obstacle in the real world, the part is not in accordance with the actual driving scene, the part of the scene needs to be filtered, that is, the scene needs to be filtered for inaccurate sensing so as to match the actual driving scene, and the first sample set obtained after the part of the scene is filtered and does not match the actual driving scene is relatively accurate.
The first scene information is obtained by collecting images of actual running conditions of the vehicle on the running lanes, and the first scene information obtained according to the collection result is information obtained by collecting the vehicle running on any lane by taking each frame as a unit. Specifically, the information includes the driving behavior of the professional driver in addition to the environmental information around the vehicle, the vehicle travel route, and the obstacle information (static or dynamic obstacle information). Feature information, such as acceleration, speed limit, corresponding timestamp and position information, can be obtained by performing feature extraction on driving behaviors in the driving process. And identifying characteristic information such as acceleration, speed limit, corresponding timestamp, position information and the like, and labeling to obtain label information such as left lane change, right lane change or straight line. The first scene information includes at least environmental information of the surroundings of the vehicle, a vehicle travel route, obstacle information (static or dynamic obstacle information), and feature information and tag information obtained by analyzing the driving behavior of a professional driver. The speed limit here is that the current vehicle (such as a main vehicle) finds that there is another vehicle (such as a preceding vehicle) on the current vehicle driving route, and if the speed of the main vehicle is 80 steps, the speed limit needs to be 50 steps to avoid collision, so as to pull the distance between the main vehicle and the preceding vehicle apart to ensure the driving safety of each other and avoid rear-end collision.
The second scene information is obtained according to the acquisition result, and at least comprises environmental information, a vehicle driving route and obstacle information (static or dynamic obstacle information) around the vehicle, and the characteristic information and the label information obtained by analyzing the driving behavior of a professional driver are required to be utilized, so that the characteristic information and the label information are also covered in the second scene information.
The first sample set obtained after the actual driving scene is mismatched is filtered, and even if the first sample set is relatively accurate, the first sample set can still deviate from the actual driving scene. For example, when a failure scenario exists but cannot be located in time, a deviation is easy to occur, the failure scenario needs to be located in time, and a second sample set is obtained according to a sample extraction strategy and the first sample set. Specifically, a neural network forming an existing vehicle lane change model is obtained, and the neural network comprises an input layer, a middle layer and an output layer. As shown in fig. 2, fig. 2 is a schematic diagram of sample screening, where all samples corresponding to the input layer 121 and the output layer 123 in the first sample set 11 are extracted as a first sub-sample 131, and a part of samples corresponding to the intermediate layer 122 in the first sample set 131 are extracted as a second sub-sample 132. A second set of samples 14 is derived from the first subsample 131 and the second subsample 132. And carrying out clustering processing on the second sample set according to a lane changing clustering strategy to obtain at least two pieces of clustering information, comparing the at least two pieces of clustering information with at least two pieces of first data obtained in actual driving of the vehicle, and obtaining a first feedback sample according to a comparison result. And according to the first feedback sample, positioning a failure scene which is not consistent with the road test of the actual driving of the vehicle. And correcting the existing vehicle lane change model according to the failure scene positioned by the first feedback sample, and performing lane change selection according to the corrected vehicle lane change model. By adopting the embodiment of the invention, because the failure scene which is not in accordance with the actual road test can be automatically positioned according to the feedback sample, the accuracy of changing the road by utilizing the model is continuously improved and promoted according to the failure scene.
Fig. 3 shows a flowchart of an information processing method according to an embodiment of the present invention. As shown in fig. 3, the process includes:
In one example, according to the second sample set, performing (offline) small-class clustering in three major classes of straight-going, left-lane changing and right-lane changing respectively to obtain a clustering result, and a specific clustering method can select K-means or hierarchical clustering, etc. And numbering the samples in the second sample set, and establishing an associated index among the sample number, the input characteristic layer, the key intermediate layer output, the output probability layer, the lane direction and the subclass number.
In one example, the cluster information (e.g., offline cluster information) is compared with the first data (e.g., online data) actually driven by the vehicle to obtain the membership of the data in the second sample set to the existing category, and the abnormal new data that does not belong to any subclass. The purpose of comparing the cluster information (such as off-line cluster information) with the first data (such as on-line data) of the actual driving of the vehicle is to: whether a certain scene accords with the clustering information obtained by clustering or not is judged, and the clustering information at least comprises 2 judgment branches: firstly, if the clustering information is met, the good example is a good example, and the good example is not used as feedback; the second step is as follows: if the clustering information is not met, the result is bad case, and the bad case is used as feedback. The bad cases and good cases are compared with the feature space index, and whether the sample type does not exist or the boundary scene or other situations are distinguished through the online prediction part of the clustering algorithm. The bad case is used as a feedback sample and retrained with the data in the second sample set.
And step 203, positioning a failure scene which is not consistent with the road test of the actual driving of the vehicle according to the first feedback sample.
And 204, correcting the existing vehicle lane change model according to the failure scene positioned by the first feedback sample, and performing lane change selection according to the corrected vehicle lane change model.
In one example, as shown in fig. 2, all samples corresponding to the input layer 121 and the output layer 123 in the first sample set 11 are extracted as a first sub-sample 131, and a part of samples corresponding to the intermediate layer 122 in the first sample set 131 are extracted as a second sub-sample 132. A second set of samples 14 is derived from the first subsample 131 and the second subsample 132. And after clustering processing is carried out according to the second sample set, comparing the plurality of clustering information with a plurality of first data obtained in the actual driving of the vehicle, wherein the plurality of first data also comprise data which are not matched with the clustering information, taking the data which are matched with the clustering information as target data, and taking the target data as a first feedback sample. And according to the first feedback sample, positioning a failure scene which is not consistent with the road test of the actual driving of the vehicle. And correcting the existing vehicle lane change model according to the failure scene positioned by the first feedback sample, and performing lane change selection according to the corrected vehicle lane change model.
Fig. 4 illustrates a flowchart of an information processing method according to an embodiment of the present invention. As shown in fig. 4, the process includes:
and 301, obtaining a second sample set according to the sample extraction strategy and the first sample set.
And 302, clustering the second sample set according to a lane change clustering strategy to obtain at least two pieces of clustering information, comparing the at least two pieces of clustering information with at least two pieces of first data obtained in the actual driving of the vehicle, inquiring target data which are not matched with each piece of clustering information from the at least two pieces of first data, and taking the target data as a first feedback sample.
And 303, clustering the second sample set according to a lane change clustering strategy, wherein the current scene is a boundary scene in the actual driving of the vehicle under the condition that clustering information is not obtained, and the boundary scene is used as a second feedback sample.
And step 304, positioning a failure scene which is not consistent with the road test of the actual driving of the vehicle according to the first feedback sample and the second feedback sample.
For the above example of clustering and obtaining more than one feedback sample, according to the second sample set, performing (offline) small-class clustering on three main classes of straight, left lane changing and right lane changing respectively to obtain a clustering result, and the specific clustering method may select K-means or hierarchical clustering, etc. And numbering the samples in the second sample set, and establishing an associated index among the sample number, the input characteristic layer, the key intermediate layer output, the output probability layer, the lane direction and the subclass number. And comparing the clustering information (such as offline clustering information) with first data (such as online data) actually driven by the vehicle to obtain the membership of the data in the second sample set and the existing classes and abnormal new data which do not belong to any subclass. The purpose of comparing the cluster information (such as off-line cluster information) with the first data (such as on-line data) of the actual driving of the vehicle is to: whether a certain scene accords with the clustering information obtained by clustering or not is judged, and the clustering information at least comprises 2 judgment branches: firstly, if the clustering information is met, the method is a good example, and the good example is not used as feedback; and the second step is as follows: if the clustering information is not met, the result is bad case, and the bad case is used as feedback. Further, a bad case is a case where there is no clustering information in a scene, and usually corresponds to a boundary scene. For this case too, it will be used as feedback. And distinguishing whether the sample type does not exist or a boundary scene or other conditions through an online prediction part of a clustering algorithm by comparing the bad case and the good case with the feature space index. And taking the bad case and/or the special case of the bad case as feedback samples, and inputting the feedback samples into the second sample set. Because the matched data in the second sample set cannot locate a failure scene and cannot fully cover a real-person driving scene, the feedback samples of the second sample set are input to be used as new training data after data restoration and completion, the new training data is used for model optimization (for example, an existing vehicle lane-changing model), a more accurate processing result can be output, and lane-changing accuracy according to the optimized vehicle lane-changing model can be greatly improved.
And 305, correcting the existing vehicle lane change model according to the failure scenes positioned by the first feedback sample and the second feedback sample, and performing lane change selection according to the corrected vehicle lane change model.
Fig. 5 illustrates a flowchart of an information processing method according to an embodiment of the present invention. As shown in fig. 5, the process includes:
And step 402, carrying out clustering processing on the second sample set according to a lane change clustering strategy to obtain at least two pieces of clustering information, comparing the at least two pieces of clustering information with at least two pieces of first data obtained in actual driving of the vehicle, inquiring target data which are not matched with each piece of clustering information from the at least two pieces of first data, and taking the target data as a first feedback sample.
And step 403, positioning a failure scene which is not consistent with the road test of the actual driving of the vehicle according to the first feedback sample.
And step 404, correcting the existing vehicle lane change model according to the failure scene positioned by the first feedback sample, and performing lane change selection according to the corrected vehicle lane change model.
And 405, inputting the first feedback sample into the existing vehicle lane change model, and calculating to obtain the probability distribution of the feedback sample.
And 406, obtaining a feedback entropy according to the probability distribution of the feedback sample, and judging whether the road test according with the actual driving of the vehicle is met according to the feedback entropy.
In one example, the existing vehicle lane change model may be a basic decision model, and the first feedback sample is input into the decision model and optimized to improve the model effect. Specifically, a first feedback sample is input into the decision model, the probability distribution of the feedback sample is obtained through operation, the feedback entropy is obtained according to the probability distribution of the feedback sample, and the larger the entropy is, the better the example is.
In an embodiment, different subnet structures may also be adjusted by different clustering information obtained by clustering. Specifically, a first feedback sample is added to the second sample set to obtain a third sample set. And clustering the third sample set according to a lane changing clustering strategy to obtain a first subclass for representing lane changing to the left, a second subclass for representing lane changing to the right and a third subclass for representing straight rows. And adjusting each corresponding sub-network in the neural network forming the existing vehicle lane change model according to the first sub-network, the second sub-network and the third sub-network. The same neural network and data are shared by all lane models (relationship model, simulation model, decision model, etc.).
Fig. 6 shows a block diagram of a configuration of an information processing apparatus including: a sample processing unit 21, configured to obtain a second sample set according to the sample extraction policy and the first sample set; the first feedback sample processing unit 22 is configured to perform clustering processing on the second sample set according to a lane change clustering strategy to obtain at least two pieces of clustering information, compare the at least two pieces of clustering information with at least two pieces of first data obtained in actual driving of a vehicle, and obtain a first feedback sample according to a comparison result; the positioning unit 23 is configured to position a failure scene that does not conform to a road test of actual driving of the vehicle according to the first feedback sample; and the lane change selection unit 24 is configured to modify the existing vehicle lane change model according to the failure scene located by the first feedback sample, and perform lane change selection according to the modified vehicle lane change model.
In one embodiment, the sample processing unit is further configured to: obtaining a neural network forming the existing vehicle lane change model, wherein the neural network comprises an input layer, a middle layer and an output layer; extracting all samples corresponding to the input layer and the output layer in the first sample set to be used as first subsamples; extracting a part of samples corresponding to the middle layer in the first sample set to be used as second sub-samples; obtaining the second sample set according to the first subsample and the second subsample.
In one embodiment, the first feedback sample processing unit is further configured to: inquiring target data which are not matched with each clustering information from the at least two pieces of first data; the target data is taken as a first feedback sample.
In one embodiment, the apparatus further comprises: the clustering processing unit is used for clustering the second sample set according to a lane change clustering strategy, and under the condition that clustering information is not obtained, the current scene is a boundary scene in the actual driving of the vehicle; a second feedback sample processing unit, configured to take the boundary scene as a second feedback sample; a first sample adding unit for adding the second feedback sample to the first feedback sample.
In one embodiment, the apparatus further comprises: the first operation unit is used for inputting the first feedback sample into the existing vehicle lane change model and obtaining the probability distribution of the feedback sample through operation; the second operation unit is used for obtaining feedback entropy according to the probability distribution of the feedback sample; and the judging unit is used for judging whether the road test of the actual driving of the vehicle is met or not according to the feedback entropy.
In one embodiment, the apparatus further comprises: a second sample adding unit, configured to add the first feedback sample into the second sample set to obtain a third sample set; the subclass processing unit is used for carrying out clustering processing on the third sample set according to a lane change clustering strategy to obtain a first subclass for representing lane change to the left, a second subclass for representing lane change to the right and a third subclass for representing straight lines; and the adjusting unit is used for adjusting each corresponding sub-network in the neural network forming the existing vehicle lane-changing model according to the first subclass, the second subclass and the third subclass.
The functions of the modules in the apparatuses according to the embodiments of the present invention may refer to the corresponding descriptions in the above methods, and are not described herein again.
Fig. 7 shows a block diagram of the structure of an information processing apparatus according to an embodiment of the present invention. As shown in fig. 7, the apparatus includes: a memory 910 and a processor 920, the memory 910 having stored therein computer programs operable on the processor 920. The processor 920 implements the automatic driving method in the above-described embodiment when executing the computer program. The number of the memory 910 and the processor 920 may be one or more.
The device also includes: and a communication interface 930 for communicating with an external device to perform data interactive transmission.
If the memory 910, the processor 920 and the communication interface 930 are implemented independently, the memory 910, the processor 920 and the communication interface 930 may be connected to each other through a bus and perform communication with each other. The bus may be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, an Extended ISA (Extended Industry Standard Component) bus, or the like. The bus may be divided into an address bus, a data bus, a control bus, etc. For ease of illustration, only one thick line is shown in FIG. 7, but this is not intended to represent only one bus or type of bus.
Optionally, in an implementation, if the memory 910, the processor 920 and the communication interface 930 are integrated on a chip, the memory 910, the processor 920 and the communication interface 930 may complete communication with each other through an internal interface.
An embodiment of the present invention provides a computer-readable storage medium, which stores a computer program, and the computer program is used for implementing the method of any one of the above embodiments when being executed by a processor.
In the description herein, references to the description of the term "one embodiment," "some embodiments," "an example," "a specific example," or "some examples," etc., mean that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the invention. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples. Furthermore, various embodiments or examples and features of different embodiments or examples described in this specification can be combined and combined by one skilled in the art without contradiction.
Furthermore, the terms "first", "second" and "first" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one such feature. In the description of the present invention, "a plurality" means two or more unless specifically defined otherwise.
Any process or method descriptions in flow charts or otherwise described herein may be understood as representing modules, segments, or portions of code which include one or more executable instructions for implementing specific logical functions or steps of the process, and alternate implementations are included within the scope of the preferred embodiment of the present invention in which functions may be executed out of order from that shown or discussed, including substantially concurrently or in reverse order, depending on the functionality involved, as would be understood by those reasonably skilled in the art of the present invention.
The logic and/or steps represented in the flowcharts or otherwise described herein, e.g., an ordered listing of executable instructions that can be considered to implement logical functions, can be embodied in any computer-readable medium for use by or in connection with an instruction execution system, apparatus, or device, such as a computer-based system, processor-containing system, or other system that can fetch the instructions from the instruction execution system, apparatus, or device and execute the instructions. For the purposes of this description, a "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transport the program for use by or in connection with the instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of the computer-readable medium would include the following: an electrical connection (electronic device) having one or more wires, a portable computer diskette (magnetic device), a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable read-only memory (CDROM). Additionally, the computer-readable medium could even be paper or another suitable medium upon which the program is printed, as the program can be electronically captured, via for instance optical scanning of the paper or other medium, then compiled, interpreted or otherwise processed in a suitable manner if necessary, and then stored in a computer memory.
It should be understood that portions of the present invention may be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, various steps or methods may be implemented in software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, any one or combination of the following techniques, which are known in the art, may be used: a discrete logic circuit having a logic gate circuit for implementing a logic function on a data signal, an application specific integrated circuit having an appropriate combinational logic gate circuit, a Programmable Gate Array (PGA), a Field Programmable Gate Array (FPGA), or the like.
It will be understood by those skilled in the art that all or part of the steps carried by the method for implementing the above embodiments may be implemented by hardware related to instructions of a program, which may be stored in a computer readable storage medium, and when the program is executed, the program includes one or a combination of the steps of the method embodiments.
In addition, functional units in the embodiments of the present invention may be integrated into one processing module, or each unit may exist alone physically, or two or more units are integrated into one module. The integrated module can be realized in a hardware mode, and can also be realized in a software functional module mode. The integrated module, if implemented in the form of a software functional module and sold or used as a separate product, may also be stored in a computer readable storage medium. The storage medium may be a read-only memory, a magnetic or optical disk, or the like.
The above description is only for the specific embodiment of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art can easily conceive various changes or substitutions within the technical scope of the present invention, and these should be covered by the scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the appended claims.
Claims (12)
1. An information processing method, characterized in that the method comprises:
obtaining a second sample set according to the sample extraction strategy and the first sample set;
clustering the second sample set according to a lane change clustering strategy to obtain at least two pieces of clustering information, comparing the at least two pieces of clustering information with at least two pieces of first data obtained in actual driving of the vehicle, and obtaining a first feedback sample according to a comparison result;
according to the first feedback sample, positioning a failure scene which is not in accordance with the actual driving road test of the vehicle;
correcting the existing vehicle lane change model according to the failure scene positioned by the first feedback sample, and performing lane change selection according to the corrected vehicle lane change model;
the method further comprises the following steps:
inputting the first feedback sample into the existing vehicle lane change model, and calculating to obtain the probability distribution of the feedback sample;
obtaining a feedback entropy according to the probability distribution of the feedback sample;
and judging whether the road test of the actual driving of the vehicle is met or not according to the feedback entropy.
2. The method of claim 1, wherein deriving the second set of samples from the sample extraction strategy and the first set of samples comprises:
obtaining a neural network forming the existing vehicle lane change model, wherein the neural network comprises an input layer, a middle layer and an output layer;
extracting all samples corresponding to the input layer and the output layer in the first sample set to be used as first sub-samples;
extracting a part of samples corresponding to the middle layer in the first sample set to be used as second sub-samples;
obtaining the second set of samples from the first subsample and the second subsample.
3. The method of claim 1, wherein comparing the at least two pieces of cluster information with at least two pieces of first data obtained during actual driving of the vehicle, and obtaining a first feedback sample according to a comparison result comprises:
inquiring target data which are not matched with each clustering information from the at least two pieces of first data;
the target data is taken as a first feedback sample.
4. The method of claim 1, further comprising:
clustering the second sample set according to a lane change clustering strategy, wherein the current scene is a boundary scene in the actual driving of the vehicle under the condition that clustering information is not obtained;
taking the boundary scene as a second feedback sample;
adding the second feedback sample to the first feedback sample.
5. The method of claim 1, further comprising:
adding the first feedback sample into the second sample set to obtain a third sample set;
clustering the third sample set according to a lane change clustering strategy to obtain a first subclass for representing lane change to the left, a second subclass for representing lane change to the right and a third subclass for representing straight line;
and adjusting each corresponding sub-network in the neural network forming the existing vehicle lane change model according to the first sub-class, the second sub-class and the third sub-class.
6. An information processing apparatus characterized in that the apparatus comprises:
the sample processing unit is used for obtaining a second sample set according to the sample extraction strategy and the first sample set;
the first feedback sample processing unit is used for carrying out clustering processing on the second sample set according to a lane change clustering strategy to obtain at least two clustering information, comparing the at least two clustering information with at least two first data obtained in actual driving of a vehicle, and obtaining a first feedback sample according to a comparison result;
the positioning unit is used for positioning a failure scene which is not in accordance with the actual driving road test of the vehicle according to the first feedback sample;
the lane change selection unit is used for correcting the existing vehicle lane change model according to the failure scene positioned by the first feedback sample and performing lane change selection according to the corrected vehicle lane change model;
the device further comprises:
the first operation unit is used for inputting the first feedback sample into the existing vehicle lane change model and obtaining the probability distribution of the feedback sample through operation;
the second operation unit is used for obtaining a feedback entropy according to the probability distribution of the feedback sample;
and the judging unit is used for judging whether the road test according with the actual driving of the vehicle is met according to the feedback entropy.
7. The apparatus of claim 6, wherein the sample processing unit is further configured to:
obtaining a neural network forming the existing vehicle lane change model, wherein the neural network comprises an input layer, a middle layer and an output layer;
extracting all samples corresponding to the input layer and the output layer in the first sample set to be used as first sub-samples;
extracting a part of samples corresponding to the middle layer in the first sample set to be used as second sub-samples;
obtaining the second set of samples from the first subsample and the second subsample.
8. The apparatus of claim 6, wherein the first feedback sample processing unit is further configured to:
querying target data which is not matched with each clustering information from the at least two pieces of first data;
the target data is taken as a first feedback sample.
9. The apparatus of claim 6, further comprising:
the clustering processing unit is used for clustering the second sample set according to a lane change clustering strategy, and under the condition that clustering information is not obtained, the current scene is a boundary scene in the actual driving of the vehicle;
a second feedback sample processing unit, configured to take the boundary scene as a second feedback sample;
a first sample adding unit for adding the second feedback sample to the first feedback sample.
10. The apparatus of claim 6, further comprising:
a second sample adding unit, configured to add the first feedback sample into the second sample set to obtain a third sample set;
the subclass processing unit is used for carrying out clustering processing on the third sample set according to a lane change clustering strategy to obtain a first subclass for representing lane change to the left, a second subclass for representing lane change to the right and a third subclass for representing straight lines;
and the adjusting unit is used for adjusting each corresponding sub-network in the neural network forming the existing vehicle lane-changing model according to the first subclass, the second subclass and the third subclass.
11. An information processing apparatus characterized in that the apparatus comprises:
one or more processors;
storage means for storing one or more programs;
the one or more programs, when executed by the one or more processors, cause the one or more processors to implement the method of any of claims 1-5.
12. A computer-readable storage medium, in which a computer program is stored which, when being executed by a processor, carries out the method according to any one of claims 1 to 5.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910130653.4A CN109948656B (en) | 2019-02-21 | 2019-02-21 | Information processing method, device and storage medium |
| CN202210485715.5A CN114861793A (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 CN109948656A (en) | 2019-06-28 |
| CN109948656B true CN109948656B (en) | 2022-11-25 |
Family
ID=67006910
Family Applications (2)
| 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 |
| CN201910130653.4A Active CN109948656B (en) | 2019-02-21 | 2019-02-21 | Information processing method, device and storage medium |
Family Applications Before (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) | CN114861793A (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| 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 |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| 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 |
| 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 |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105555630B (en) * | 2013-07-19 | 2017-07-25 | 日产自动车株式会社 | Driving condition estimation unit |
| CN106557963B (en) * | 2015-09-30 | 2022-12-16 | 百度在线网络技术（北京）有限公司 | Method, device and server for processing vehicle order |
| CN109060370B (en) * | 2018-06-29 | 2019-12-10 | 奇瑞汽车股份有限公司 | Method and device for vehicle testing of automatically driven vehicle |
| CN109271858B (en) * | 2018-08-13 | 2020-11-17 | 武汉中海庭数据技术有限公司 | Intersection identification method and system based on vehicle path and visual lane sideline data |
| 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 |
-
2019
- 2019-02-21 CN CN202210485715.5A patent/CN114861793A/en active Pending
- 2019-02-21 CN CN201910130653.4A patent/CN109948656B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| 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 |
| 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 |
Non-Patent Citations (2)
| Title |
|---|
| 监控视频中的车辆异常行为检测;黄鑫等;《计算机系统应用》;20180215(第02期);127-133 * |
| 预测自动驾驶车辆变道意图;Shuang Su等;《https://cloud.tencent.com/developer/news/289224》;20180731;第1页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109948656A (en) | 2019-06-28 |
| CN114861793A (en) | 2022-08-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109784391B (en) | Multi-model-based sample labeling method and device | |
| CN109508580B (en) | Traffic signal lamp identification method and device | |
| CN109703569B (en) | Information processing method, device and storage medium | |
| CN109325538B (en) | Object detection method, device and computer-readable storage medium | |
| US20220048536A1 (en) | Method and device for testing a driver assistance system | |
| CN111860255B (en) | Training and using methods, devices, equipment and media of driving detection model | |
| CN109740609B (en) | Track gauge detection method and device | |
| CN112580734B (en) | Target detection model training method, system, terminal equipment and storage medium | |
| CN109886198B (en) | Information processing method, device and storage medium | |
| CN114627394B (en) | Muck vehicle fake plate identification method and system based on unmanned aerial vehicle | |
| CN112561859A (en) | Monocular vision-based steel belt drilling and anchor net identification method and device for anchoring and protecting | |
| CN109948656B (en) | Information processing method, device and storage medium | |
| CN111947669A (en) | Method for using feature-based positioning maps for vehicles | |
| CN109948449B (en) | Information processing method, device and storage medium | |
| CN111352414A (en) | Decoy removal apparatus and method for vehicle and vehicle including the same | |
| US20230237810A1 (en) | Method and system for annotating sensor data | |
| US11720343B2 (en) | Methods and apparatus for automatically labeling data processing events in autonomous driving vehicles via machine learning | |
| CN111143423B (en) | Dynamic scene labeling data mining method and device and terminal | |
| CN115512098A (en) | Electronic bridge inspection system and inspection method | |
| CN115482672A (en) | Vehicle reverse running detection method and device, terminal equipment and storage medium | |
| Ravishankaran | Impact on how AI in automobile industry has affected the type approval process at RDW | |
| CN113085861A (en) | Control method and device for automatic driving vehicle and automatic driving vehicle | |
| CN115393379A (en) | Data annotation method and related product | |
| CN111427037A (en) | Obstacle detection method and device, electronic equipment and vehicle-end equipment | |
| CN112258602B (en) | Stop line generation method and device, electronic equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |