CN112163521A - Vehicle driving behavior identification method, device and equipment - Google Patents
Vehicle driving behavior identification method, device and equipment Download PDFInfo
- Publication number
- CN112163521A CN112163521A CN202011049638.6A CN202011049638A CN112163521A CN 112163521 A CN112163521 A CN 112163521A CN 202011049638 A CN202011049638 A CN 202011049638A CN 112163521 A CN112163521 A CN 112163521A
- Authority
- CN
- China
- Prior art keywords
- driving
- data
- azimuth data
- azimuth
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 94
- 230000006399 behavior Effects 0.000 claims abstract description 215
- 230000008569 process Effects 0.000 claims abstract description 33
- 230000008859 change Effects 0.000 claims description 105
- 238000012549 training Methods 0.000 claims description 11
- 230000011218 segmentation Effects 0.000 claims description 8
- 238000004364 calculation method Methods 0.000 claims description 7
- 238000004590 computer program Methods 0.000 claims description 6
- 238000010586 diagram Methods 0.000 description 7
- 238000005192 partition Methods 0.000 description 4
- 238000012545 processing Methods 0.000 description 4
- 230000009471 action Effects 0.000 description 3
- 239000006185 dispersion Substances 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 238000013145 classification model Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 230000014509 gene expression Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/59—Context or environment of the image inside of a vehicle, e.g. relating to seat occupancy, driver state or inner lighting conditions
-
- 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
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
- G06V10/267—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion by performing operations on regions, e.g. growing, shrinking or watersheds
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Multimedia (AREA)
- Traffic Control Systems (AREA)
Abstract
The embodiment of the application discloses a method, a device and equipment for identifying vehicle driving behaviors, wherein driving azimuth data in the vehicle driving process is collected firstly; and calculating the data characteristics of the driving azimuth data within the preset time length, inputting the data characteristics into a vehicle driving behavior classifier, and obtaining an output classification result. And when the classification result is normal driving, storing the driving azimuth data and re-executing the steps. When the classification result is that the lane changes to drive, determining the driving azimuth data stored in the latest preset duration as first azimuth data, and re-executing the steps of collecting and classifying until the classification result is obtained as normal driving; then, using the driving azimuth data stored within the preset time length as second azimuth data; and finally, determining the type of the driving behavior of the lane according to the first azimuth data and the second azimuth data. Therefore, the type of the driving behavior changing of the lane can be accurately and timely identified.
Description
Technical Field
The application relates to the field of vehicle control, in particular to a method, a device and equipment for identifying vehicle driving behaviors.
Background
The driving behavior is changed in real time during the driving of the vehicle. For example, when a vehicle turns around, overtakes, or turns around, it is necessary to change lanes so that the vehicle travels to the correct position.
For an intelligent automobile, it is important to be able to accurately and quickly determine the driving behavior of the automobile. By determining the driving behavior of the vehicle, the driving path of the vehicle can be predicted, and the driving safety of the intelligent automobile can be ensured. However, current recognition methods for a lane change of a vehicle cannot quickly and accurately recognize these driving behaviors.
Disclosure of Invention
In view of this, embodiments of the present application provide a method, an apparatus, and a device for identifying a driving behavior of a vehicle, which can quickly and accurately identify a type of a driving behavior of a lane change of the vehicle.
In order to solve the above problem, the technical solution provided by the embodiment of the present application is as follows:
a method of identifying a driving behavior of a vehicle, the method comprising:
collecting driving azimuth data in the driving process of a vehicle;
when the acquisition duration reaches a preset duration, calculating the data characteristics of the driving azimuth data in the preset duration;
inputting the data characteristics of the driving azimuth data within the preset duration into a vehicle driving behavior classifier, and obtaining a classification result output by the vehicle driving behavior classifier;
when the classification result is normal driving, storing the driving azimuth data within the preset duration, and re-executing the driving azimuth data in the vehicle driving process and the subsequent steps;
when the classification result is that the lane changes to drive, determining the newly stored driving azimuth data within the preset duration as first azimuth data, and re-executing the acquisition of the driving azimuth data in the driving process of the vehicle and the subsequent steps until the classification result is obtained again as normal driving, and storing the driving azimuth data within the preset duration as second azimuth data;
and determining the type of the driving behavior of the lane according to the first azimuth data and the second azimuth data.
In one possible implementation, the method further includes:
collecting original driving azimuth data of a vehicle when driving behaviors of the vehicle change in lanes of different types;
performing data segmentation on each piece of original driving azimuth data according to a sliding window to obtain driving azimuth data to be trained, wherein the length of the sliding window is preset duration;
calculating the data characteristics of the driving azimuth data to be trained;
and training by using the data characteristics of the driving azimuth data to be trained and the classification result label value corresponding to the driving azimuth data to be trained to obtain a vehicle driving behavior classifier, wherein the classification result label value corresponding to the driving azimuth data to be trained comprises normal driving or lane change driving.
In a possible implementation manner, the calculating the data feature of the driving azimuth data to be trained includes:
and calculating a first variance, a first range and a first average absolute deviation value of the driving azimuth data to be trained as data characteristics of the driving azimuth data to be trained.
In one possible implementation, the vehicle driving behavior classifier is a naive bayes classifier;
the step of inputting the data characteristics of the driving azimuth data within the preset duration into a vehicle driving behavior classifier to obtain the classification result output by the vehicle driving behavior classifier comprises the following steps:
calculating a first posterior probability that the data characteristics of the driving azimuth data in the preset time length belong to a normal driving classification result and a second posterior probability that the data characteristics of the driving azimuth data in the preset time length belong to a lane change driving classification result by using the naive Bayes classifier;
and if the first posterior probability is greater than the second posterior probability, determining that the classification result is normal driving, and if the first posterior probability is less than the second posterior probability, determining that the classification result is lane change driving.
In a possible implementation manner, when the acquisition duration reaches a preset duration, calculating a data feature of the azimuth data within the preset duration includes:
and when the acquisition time length reaches a preset time length, calculating a second variance, a second polar difference and a second average absolute deviation value of the driving azimuth data in the preset time length as data characteristics of the driving azimuth data in the preset time length.
In a possible implementation manner, the determining the type of the lane change driving behavior according to the first azimuth data and the second azimuth data includes:
calculating a first azimuth data mean value corresponding to the first azimuth data;
calculating a second azimuth data mean value corresponding to the second azimuth data;
calculating the absolute value of the difference between the first azimuth data mean value and the second azimuth data mean value to obtain a target numerical value;
if the target value is greater than or equal to 0 and less than or equal to a first threshold value, determining that the type of the lane change driving behavior is lane change driving behavior;
if the target value is larger than or equal to a second threshold value, determining that the type of the lane change driving behavior is a U-turn driving behavior;
and if the target value is greater than or equal to a third threshold value and less than or equal to a fourth threshold value, determining that the type of the lane change driving behavior is the turning driving behavior.
In a possible implementation manner, the calculating a mean value of the first azimuth data corresponding to the first azimuth data includes:
dividing the first azimuth angle data into a plurality of parts, calculating a third variance of each part of the first azimuth angle data, and calculating a mean value of the part of the first azimuth angle data with the minimum third variance as a mean value of the first azimuth angle data;
the calculating a second azimuth data mean value corresponding to the second azimuth data includes:
dividing the second azimuth data into a plurality of parts, calculating a fourth variance of each part of second azimuth data, and calculating a mean value of the part of second azimuth data with the smallest fourth variance as a mean value of the second azimuth data.
An apparatus for recognizing a running behavior of a vehicle, the apparatus comprising:
the acquisition unit is used for acquiring driving azimuth data in the driving process of the vehicle;
the first calculation unit is used for calculating the data characteristics of the driving azimuth data in the preset duration when the acquisition duration reaches the preset duration;
the input unit is used for inputting the data characteristics of the driving azimuth data within the preset duration into a vehicle driving behavior classifier and obtaining a classification result output by the vehicle driving behavior classifier;
the first classification unit is used for storing the driving azimuth data within the preset duration when the classification result is normal driving, and re-executing the driving azimuth data in the driving process of the collected vehicle and the subsequent steps;
the second classification unit is used for determining the latest stored driving azimuth data within the preset duration as the first azimuth data when the classification result indicates that the lane changes and the driving, and re-executing the acquisition of the driving azimuth data in the driving process of the vehicle and the subsequent steps until the classification result indicates that the vehicle normally drives again, and storing the driving azimuth data within the preset duration as the second azimuth data;
and the determining unit is used for determining the type of the driving behavior of the lane change according to the first azimuth angle data and the second azimuth angle data.
An apparatus for recognizing a running behavior of a vehicle, comprising: the vehicle driving behavior recognition method comprises a memory, a processor and a computer program which is stored on the memory and can run on the processor, wherein when the processor executes the computer program, the vehicle driving behavior recognition method is realized.
A computer-readable storage medium having stored therein instructions that, when executed on a terminal device, cause the terminal device to execute the above-described method of recognizing a vehicle travel behavior.
Therefore, the embodiment of the application has the following beneficial effects:
according to the method for identifying the vehicle driving behavior, driving azimuth data in the vehicle driving process are collected; when the acquisition duration reaches a preset duration, calculating the data characteristics of the driving azimuth data within the preset duration; the data characteristics of the driving azimuth data within the preset time length are input into the vehicle driving behavior classifier, so that the classification result output by the vehicle driving behavior classifier can be obtained; when the classification result is normal driving, storing driving azimuth data within a preset time length, and executing the steps again; when the classification result is that the lane changes to drive, determining the driving azimuth data stored in the latest preset time period as first azimuth data, and re-executing the steps of collecting the driving azimuth data and the subsequent steps until the classification result is obtained to be normal drive; then, using the driving azimuth data stored within the preset time length as second azimuth data; and finally, determining the type of the driving behavior of the lane according to the first azimuth data and the second azimuth data.
Therefore, the vehicle driving behaviors can be classified according to the data characteristics by acquiring the driving azimuth data of the vehicle driving and correspondingly calculating the data characteristics of the driving azimuth data within the preset duration. And when the classification result is that the lane is changed to drive, respectively taking the latest stored driving azimuth data within the preset duration and driving azimuth data within the preset duration of normal driving, which are acquired again later, as the first azimuth data and the second azimuth data. The type of the driving behavior of the lane change can be accurately obtained through the first azimuth data before the lane change and the second azimuth data after the lane change. And the driving azimuth data is collected, stored and analyzed in real time, so that the type of the driving behavior of the lane change can be identified and determined in time according to the driving azimuth data.
Drawings
Fig. 1 is a schematic diagram of a framework of an exemplary application scenario provided in an embodiment of the present application;
fig. 2 is a flowchart of a method for identifying a driving behavior of a vehicle according to an embodiment of the present application;
FIG. 3 is a schematic diagram of three vehicle lane change driving modes provided by the embodiment of the present application;
FIG. 4 is a flowchart of a training method for a vehicle driving behavior classifier according to an embodiment of the present disclosure;
FIG. 5 is a flowchart of a method for determining a type of lane change driving behavior according to an embodiment of the present disclosure;
fig. 6 is a schematic structural diagram of an identification device for vehicle driving behavior according to an embodiment of the present application.
Detailed Description
In order to make the aforementioned objects, features and advantages of the present application more comprehensible, embodiments accompanying the drawings are described in detail below.
In order to facilitate understanding and explaining the technical solutions provided by the embodiments of the present application, the following description will first describe the background art of the present application.
After researching the traditional method for identifying the driving behavior of the vehicle, the inventor finds that the existing method for identifying the driving behavior of the vehicle is obtained by analyzing the camera data or the vehicle driving data monitored by a radar, has higher requirements on storage and computing equipment, and is difficult to meet the requirements on intelligent vehicle control. And the conventional recognition method of the traveling behavior of the vehicle cannot realize recognition of a specific type of the traveling behavior of the vehicle. Based on the above, the embodiment of the application provides a method for identifying vehicle driving behaviors, by collecting driving azimuth data in a vehicle driving process, when the collection duration reaches a preset duration, calculating data characteristics of the driving azimuth data within the preset duration; the data characteristics of the driving azimuth data within the preset time length are input into the vehicle driving behavior classifier, so that the classification result output by the vehicle driving behavior classifier can be obtained; when the classification result is normal driving, storing driving azimuth data within a preset time length, and executing the steps again; when the classification result is that the lane changes to drive, determining the driving azimuth data stored in the latest preset time period as first azimuth data, and re-executing the steps of collecting the driving azimuth data and the subsequent steps until the classification result is obtained to be normal drive; then, using the driving azimuth data stored within the preset time length as second azimuth data; and finally, determining the type of the driving behavior of the lane according to the first azimuth data and the second azimuth data.
In order to facilitate understanding of the method for identifying the vehicle driving behavior provided in the embodiment of the present application, the following description is made with reference to a scene example shown in fig. 1. Referring to fig. 1, the figure is a schematic diagram of a framework of an exemplary application scenario provided in an embodiment of the present application.
In practical application, the driving azimuth data of the vehicle in the driving process is collected firstly. And when the acquisition time reaches the preset time, calculating the data characteristics of the driving azimuth data acquired in the preset time, and inputting the data characteristics into a vehicle driving behavior classifier to obtain a classification result. And if the classification result is normal driving, the vehicle lane is not changed at the moment, the driving azimuth data in the current preset time length is stored, and the driving azimuth data is continuously acquired, extracted and classified. And if the classification result continues to be normal driving, storing the driving azimuth data within the current preset duration again, and continuing to collect, extract and classify the driving azimuth data, and so on.
If the classification result indicates that the lane is changed to drive, the lane of the vehicle is changed at the moment, the newly stored driving azimuth data within the preset time length (namely the driving azimuth data within the preset time length stored when the last classification result is normal driving) is obtained as the first azimuth data, and the driving azimuth data is continuously collected, extracted and classified. And when the classification result is changed into normal driving again, the vehicle lane change behavior is finished, the vehicle lane change behavior is changed back to normal driving, and the driving azimuth data within the current preset time length is stored as second azimuth data. And finally, determining the type of the driving behavior of the lane according to the first azimuth data and the second azimuth data.
Those skilled in the art will appreciate that the block diagram shown in fig. 1 is only one example in which embodiments of the present application may be implemented. The scope of applicability of the embodiments of the present application is not limited in any way by this framework.
In order to facilitate understanding of the present application, a method for identifying a vehicle driving behavior provided in an embodiment of the present application is described below with reference to the accompanying drawings.
Referring to fig. 2, which is a flowchart illustrating a method for identifying a driving behavior of a vehicle according to an embodiment of the present disclosure, as shown in fig. 2, the method may include S201 to S206:
s201: and collecting the driving azimuth angle data in the driving process of the vehicle.
During the running process of the vehicle, the running route of the vehicle is changed by adjusting the running azimuth. When lane changing, turning around or turning is required, the driving azimuth of the vehicle is required to be controlled to adjust the driving of the vehicle.
Wherein, the driving azimuth angle can be defined as an included angle between the driving direction of the vehicle and the due north direction. A clockwise direction may be defined as a positive direction and a counter-clockwise direction may be defined as a negative direction. And the range of values for the heading angle data may be between plus 180 degrees and minus 180 degrees.
The azimuth data may be collected by a device mounted on the vehicle, such as a global positioning system device or a magnetic-stylus device. And in the running process of the vehicle, collecting the running azimuth data of the vehicle so as to identify the running behavior of the vehicle according to the running azimuth data.
S202: and when the acquisition time reaches the preset time, calculating the data characteristics of the driving azimuth data in the preset time.
It can be understood that the driving behavior of the vehicle needs to be identified by the driving azimuth data for a period of time, and the driving behavior of the vehicle cannot be accurately judged only by the driving azimuth data at one moment.
When collecting the data of the driving azimuth, a preset duration can be set. When the acquisition time reaches the preset time, the driving azimuth data within the preset time can be processed. The preset time period may be determined according to an average operation time period of the driving behaviors such as lane change, steering, and the like during the driving of the vehicle. In one possible implementation, the preset duration may be set to 3 seconds.
After the driving azimuth data within the preset duration is obtained, the data characteristics of the driving azimuth data within the preset duration can be calculated. The data characteristic of the driving azimuth data within the preset time period may be a time domain characteristic of the driving azimuth data, including but not limited to variance, range, average, and the like. The data characteristic of the driving azimuth data within the preset time length can represent the driving behavior of the vehicle within the preset time length.
In one possible implementation, the fluctuation of the azimuth data within the preset time period may be calculated as the data characteristic of the azimuth data. The embodiment of the present application further provides a method for calculating data characteristics, and please refer to the following detailed description.
S203: and inputting the data characteristics of the driving azimuth data within the preset time length into the vehicle driving behavior classifier, and obtaining a classification result output by the vehicle driving behavior classifier.
And inputting the obtained data characteristics into a vehicle driving behavior classifier, determining the driving behavior of the vehicle within a preset time length according to a classification result output by the vehicle driving behavior classifier, and performing corresponding processing on different driving behaviors.
The classification result output by the vehicle driving behavior classifier is normal driving or lane change driving. The normal driving means that the driving behavior of the vehicle represented by the data characteristics of the driving azimuth data within the preset time length is normal driving of the vehicle, and no other driving behavior exists. The lane change driving means that the driving behavior of the vehicle represented by the data characteristics of the driving azimuth data within the preset time length is the lane change of the vehicle. The driving behavior of the vehicle can be preliminarily judged according to the classification result output by the vehicle driving behavior classifier, and the driving azimuth data is correspondingly processed based on the driving behavior of the vehicle.
The vehicle driving behavior classifier is a classification model capable of realizing data feature classification. The vehicle driving behavior classifier may be trained in advance according to data features representing different vehicle driving behaviors.
The embodiment of the application also provides a training method for the vehicle driving behavior classifier. And when the vehicle driving behavior classifier is a naive bayes classifier, a corresponding method for classifying the vehicle driving behavior according to the data characteristics of the driving azimuth data is provided, which is specifically referred to in the following detailed description.
S204: and when the classification result is normal driving, storing the driving azimuth data within a preset time length, and re-executing the steps of collecting the driving azimuth data in the driving process of the vehicle and the follow-up steps.
When the classification result is normal driving, the vehicle lane is not changed, and the type of the driving behavior of the vehicle lane is not required to be judged according to the driving azimuth data.
And storing the driving azimuth data within the preset time length so as to identify the type of the driving behavior of the lane change by using the stored driving azimuth data after the lane change behavior of the vehicle is determined.
In one possible implementation, the driving azimuth data within the preset time period may be stored in a storage device, such as a data server.
After the driving azimuth data within the preset time length is stored, in order to identify the driving behavior of the vehicle in real time, the step S201 of collecting the driving azimuth data during the driving process of the vehicle is executed again, and the subsequent steps are executed again. And further obtaining a classification result corresponding to the data characteristic of the driving azimuth data of the next preset duration, and carrying out corresponding processing on the driving azimuth data according to the classification result. And if the classification result corresponding to the data feature of the driving azimuth data of the next preset time length is still normal driving, storing the driving azimuth data of the preset time length, and continuing to perform S201 acquisition and subsequent steps. And if the classification result corresponding to the data feature of the driving azimuth data of the next preset duration is that the lane is changed for driving, executing S205 a subsequent step corresponding to the lane change for driving.
S205: and when the classification result is that the lane changes to drive, determining the newly stored driving azimuth data within the preset duration as first azimuth data, and re-executing the steps of collecting the driving azimuth data in the driving process of the vehicle until the classification result is obtained again as normal driving, and storing the driving azimuth data within the preset duration as second azimuth data.
When the classification result is that the lane change driving occurs, it is necessary to further determine the type of the lane change driving behavior.
When the vehicle lane changes to run, the running azimuth data of the vehicle changes correspondingly. And after lane change driving, the driving azimuth data corresponding to the driving behavior of different types of lane change driving are different. For example, referring to fig. 3, the embodiment of the present application provides a schematic diagram of three types of vehicle lane change driving. As shown in fig. 3a, when the vehicle performs the normal lane change driving behavior, the driving directions of the vehicles before and after the normal lane change are substantially the same, and the driving azimuth data of the corresponding vehicles before and after the normal lane change are substantially the same. And as shown in fig. 3b, the driving azimuth data of the vehicle should be reversed before and after the vehicle is driven to turn around. As shown in fig. 3c, the data of the traveling azimuth angle of the vehicle before and after the vehicle turns is different and has a certain angle difference.
Therefore, after it is determined that the vehicle lane change travel has occurred, the type of the vehicle lane change travel behavior may be determined from the travel azimuth data before and after the vehicle lane change travel.
The driving azimuth data before the vehicle lane is changed may be the latest stored driving azimuth data within the preset time. And determining the newly stored driving azimuth data within the preset duration as first azimuth data.
For the driving azimuth data after the vehicle lane is changed, after it is determined that the classification result of the driving azimuth data within the current preset time is that the lane is changed, S201 is executed again to collect the driving azimuth data during the vehicle driving process and the subsequent steps until the obtained classification result is normal driving, and at this time, the corresponding driving azimuth data within the preset time is the driving azimuth data after the vehicle lane is changed. And storing the driving azimuth data within the preset time length of normal driving as a classification result after re-executing the acquisition of the driving azimuth data in the driving process of the vehicle and the subsequent steps, and taking the driving azimuth data as second azimuth data.
S206: and determining the type of the driving behavior of the lane according to the first azimuth data and the second azimuth data.
According to the obtained first azimuth data and the second azimuth data, the change of the azimuth of the vehicle before and after the lane change driving behavior can be determined, and the type of the lane change driving behavior can be determined.
The lane change driving behavior comprises normal lane change driving, turning driving and the like.
The embodiment of the present application further provides a method for determining a type of a driving behavior of a lane change according to the first azimuth data and the second azimuth data, which is described in detail below.
In the embodiment of the application, by acquiring the driving azimuth data of the vehicle and calculating the data characteristics of the driving azimuth data within the preset time length, the classification result corresponding to the driving azimuth data can be obtained through the vehicle driving behavior classifier. By collecting the driving azimuth data of the vehicle in time and classifying according to the data characteristics, the driving behaviors of the vehicle can be classified in time, and the type of the driving behavior of the lane can be conveniently determined. When the lane change driving is performed, the type of the lane change driving behavior can be determined by the driving azimuth data before the lane change driving and the driving azimuth data after the lane change driving, that is, the first azimuth data and the second azimuth data. Through the driving azimuth data before and after the lane is changed, the type of the driving behavior of the vehicle lane can be accurately identified.
Before the data characteristics of the driving azimuth data within the input preset duration are classified by using the vehicle driving behavior classifier, the vehicle driving behavior classifier needs to be trained.
Referring to fig. 4, the figure is a flowchart of a training method for a vehicle driving behavior classifier according to an embodiment of the present application. In the embodiment of the present application, the method for identifying a driving behavior of a vehicle further includes S401 to S404:
s401: the method comprises the steps of collecting original driving azimuth data of a vehicle when driving behaviors of the vehicle change in lanes belonging to different types.
And aiming at the change of the driving behaviors of different types of lanes, acquiring the original driving azimuth data of the vehicle. The raw driving azimuth data may include all driving azimuth data of the vehicle in the driving process of lane change, and the acquisition duration of the raw driving azimuth data may include a time length of the vehicle when the vehicle completes one lane change driving. As an example, the acquisition time of the raw azimuth data may be 15 seconds.
By collecting the original driving azimuth data of the driving behavior changing of the lanes belonging to different types, the driving azimuth data corresponding to the driving behavior changing of the lanes of different types can be obtained, so that the collected original driving azimuth data can be used for extracting corresponding data characteristics.
In order to improve the accuracy of classification of the trained vehicle driving behavior classifier, a plurality of groups of data can be acquired according to original driving azimuth data of driving behaviors changed in lanes of different types. In one possible implementation, the raw driving azimuth data corresponding to the changed driving behavior of the different types of lanes may be at least 50 sets.
S402: and performing data segmentation on each piece of original driving azimuth data according to a sliding window to obtain driving azimuth data to be trained, wherein the length of the sliding window is preset duration.
And performing data segmentation on each piece of original driving azimuth data, specifically adopting a sliding window mode, wherein the length of the sliding window can be preset duration. Within the preset time, the data in the process of changing the driving behavior of the vehicle lane can be covered. The preset duration may be less than the acquisition duration of the original driving azimuth data. For example, the preset time period may be 3 seconds.
In addition, in order to prevent the original driving azimuth data from being lost when data division is performed, data partially overlapping between adjacent sliding windows may exist when the sliding windows are moved. In one possible implementation, there may be 50% overlapping data between adjacent sliding windows.
Taking the example that the acquisition time of the original driving azimuth data can be 15 seconds, the preset time can be 3 seconds, and 50% of coincident data exists between adjacent sliding windows, 9 driving azimuth data to be trained can be obtained after data segmentation is carried out on each group of original driving azimuth data. The corresponding calculation formula is 15/(3 × 50%) -1 ═ 9.
And obtaining the driving azimuth data to be trained corresponding to each sliding window through the sliding windows, and realizing data segmentation of the original driving azimuth data.
S403: and calculating the data characteristics of the driving azimuth data to be trained.
And after the data are segmented to obtain the driving azimuth data to be trained, calculating the data characteristics of the driving azimuth data to be trained.
The data characteristics of the driving azimuth data to be trained can reflect the characteristics of the driving azimuth data to be trained under the driving behaviors, can reflect the characteristics of the driving azimuth data to be trained under different driving behaviors, and train the vehicle driving behavior classifier by using the data characteristics.
The data features may be time domain features of the driving azimuth data to be trained. The embodiment of the application also provides a method for calculating the time domain characteristics of the driving azimuth data to be trained, please refer to the following.
S404: and training by using the data characteristics of the driving azimuth data to be trained and the classification result label value corresponding to the driving azimuth data to be trained to obtain the vehicle driving behavior classifier, wherein the classification result label value corresponding to the driving azimuth data to be trained comprises normal driving or lane change driving.
And training to generate a vehicle driving behavior classifier by using the obtained data characteristics of the driving azimuth data to be trained and the classification result label values corresponding to the driving azimuth data to be trained.
The classification result label value corresponding to the driving azimuth data to be trained comprises two types of driving, namely normal driving or lane change driving, and can be determined according to the driving behavior of the vehicle corresponding to the driving azimuth data to be trained. The driving azimuth data to be trained with the classification result label value of normal driving can be the driving azimuth data to be trained before the vehicle changes the lane or after the lane change is completed. The driving azimuth data to be trained for driving with lane change can be the driving azimuth data to be trained when the vehicle changes lanes.
Based on the above content, in the embodiment of the application, the data characteristics of the segmented driving azimuth data to be trained can be obtained by collecting the original driving azimuth data when the driving behaviors of the vehicles change in lanes of different types and segmenting the original driving azimuth data. And training to generate a corresponding vehicle driving behavior classifier by using the data characteristics of the driving azimuth data to be trained and the classification result label value corresponding to the driving azimuth data to be trained. The original driving azimuth data are segmented in a sliding window mode, and therefore the driving azimuth data to be trained, which are in different driving behaviors in the driving process of changing the vehicle lane, can be obtained. Therefore, the vehicle driving behavior classifier can be trained by using the driving azimuth data to be trained of different driving behaviors, and the accurate vehicle driving behavior classifier is obtained.
Further, the data characteristics of the driving azimuth data to be trained may include variance, range, and mean absolute deviation.
Calculating the data characteristics of the driving azimuth data to be trained, comprising the following steps:
and calculating a first variance, a first range and a first average absolute deviation value of the driving azimuth data to be trained as data characteristics of the driving azimuth data to be trained.
Under different driving behaviors, the numerical value of the driving azimuth angle data to be trained is different in variation. During normal driving, as the vehicle keeps driving in a certain direction, the numerical value of the corresponding driving azimuth data to be trained can fluctuate randomly with a small amplitude up and down at a certain numerical value. However, when the lane changes, the vehicle may continuously adjust the driving direction, and the corresponding value of the driving azimuth data to be trained may fluctuate significantly according to a certain development trend.
In order to characterize the fluctuation of the driving azimuth data to be trained, the data characteristics of the driving azimuth data to be trained can be calculated. The data characteristic is a three-dimensional characteristic vector consisting of a first variance, a first range and a first average absolute deviation value of the driving azimuth data to be trained.
Wherein the first variance is the average of the squared values of the difference between the value of each of the driving azimuth data to be trained and the average of the driving azimuth data to be trained. The first variance may characterize a degree of dispersion of the heading angle data to be trained. The first extreme difference is a difference between a maximum value and a minimum value in the driving azimuth data to be trained, and the dispersion degree of the driving azimuth data to be trained can be represented as well. The first average absolute deviation value is an average value of absolute values of deviations of the single driving azimuth data to be trained and the arithmetic average value of the driving azimuth data to be trained. The first average absolute deviation value can avoid mutual offset of errors and accurately reflect the magnitude of data errors.
By calculating the first variance, the first range and the first average absolute deviation value of the driving azimuth data to be trained, the fluctuation condition of the driving azimuth data to be trained can be reflected more accurately from multiple aspects. The data characteristics of the driving azimuth data to be trained are utilized to train the vehicle driving behavior classifier, so that the obtained vehicle driving behavior classifier is more accurate, the accuracy of vehicle driving behavior classification is improved, and the recognition of the driving behavior of lane change can be accurately realized.
When the data features of the driving azimuth data of the vehicle are classified, the classification real-time performance needs to be ensured, so that the classification result corresponding to the data features of the driving azimuth data of the vehicle can be determined quickly, and the driving behavior of the vehicle can be identified.
In a possible implementation manner, the vehicle driving behavior classifier can adopt a naive Bayes classifier with shorter execution time, so that the data characteristics of the driving azimuth data of the vehicle can be classified on line in time.
Based on the training method for the vehicle driving behavior classifier, when the vehicle driving behavior classifier is a naive Bayes classifier, the embodiment of the application also provides a method for obtaining the classification result by using the naive Bayes classifier.
Inputting the data characteristics of the driving azimuth data within the preset duration into a vehicle driving behavior classifier to obtain a classification result output by the vehicle driving behavior classifier, wherein the method comprises the following two steps:
a1: and calculating a first posterior probability that the data characteristics of the driving azimuth data in the preset time length belong to the normal driving classification result and a second posterior probability that the data characteristics of the driving azimuth data in the preset time length belong to the lane change driving classification result by using a naive Bayes classifier.
And inputting the data characteristics of the driving azimuth data within the preset time length into a naive Bayes classifier, and calculating a first posterior probability that the data characteristics of the driving azimuth data within the preset time length respectively belong to the normal driving classification result and a second posterior probability that the data characteristics belong to the lane change driving classification result.
By using a naive Bayes classifier, the posterior probabilities of the data characteristics of the driving azimuth data corresponding to the normal driving classification results and the driving classification results with changed lanes are calculated, and the probabilities of the data characteristics of the driving azimuth data corresponding to two different driving classification results can be obtained. And then determining a classification result according to the magnitude relation between the first posterior probability and the second posterior probability.
A2: and if the first posterior probability is greater than the second posterior probability, determining that the classification result is normal driving, and if the first posterior probability is less than the second posterior probability, determining that the classification result is lane change driving.
When the first posterior probability is greater than the second posterior probability, the probability that the data characteristic of the driving azimuth data corresponds to the normal driving classification result is greater, and the data characteristic of the driving azimuth data corresponds to the normal driving classification result. On the contrary, when the second posterior probability is greater than the first posterior probability, the data characteristic of the driving azimuth data corresponding to the result of lane change driving classification is more likely to correspond to the result of lane change driving classification.
Based on the above, by using the naive bayes classifier as the vehicle driving behavior classifier, the classification efficiency of the data features of the driving azimuth data can be improved. When the classification result is determined, the classification result to which the data feature of the driving azimuth data in the preset duration belongs can be determined quickly by calculating the first posterior probability and the second posterior probability corresponding to the data feature of the driving azimuth data in the preset duration and then comparing the first posterior probability and the second posterior probability. Therefore, the data of the driving azimuth angle can be classified in time, and the efficiency of determining the driving behavior is improved.
In one possible implementation, the data characteristics of the driving azimuth data within the preset time period may include a variance, a range, and a mean absolute deviation value.
In the method for calculating the data characteristics of the data of the driving azimuth provided by the embodiment of the application, when the acquisition duration reaches the preset duration, the data characteristics of the data of the driving azimuth within the preset duration are calculated, and the method includes the following steps:
and when the acquisition time reaches the preset time, calculating a second variance, a second polar difference and a second average absolute deviation value of the driving azimuth data in the preset time as the data characteristics of the driving azimuth data in the preset time.
And when the acquisition time reaches the preset time, calculating a second variance, a second polar difference and a second average absolute deviation value of the driving azimuth data in the preset time as data characteristics to form a three-dimensional characteristic vector corresponding to the driving azimuth data in the preset time so as to input the obtained data characteristics to a vehicle driving behavior classifier and obtain a corresponding classification result.
The method for calculating the second variance, the second range and the second average absolute deviation of the driving azimuth data within the preset time duration is similar to the method for calculating the first variance, the first range and the first average absolute deviation of the driving azimuth data to be trained, and is not repeated herein.
When the type of the driving behavior of the lane is determined according to the first azimuth data and the second azimuth data, the predetermined threshold corresponding to the type of the driving behavior of different lanes can be used for determining.
Specifically, referring to fig. 5, this is a flowchart of a method for determining a type of lane change driving behavior according to an embodiment of the present application. Determining the type of the lane change driving behavior according to the first azimuth data and the second azimuth data, including S501-S506:
s501: and calculating a first azimuth data mean value corresponding to the first azimuth data.
After the first azimuth angle data are obtained, a first azimuth angle data mean value corresponding to the first azimuth angle data is calculated. And the first azimuth angle mean value is the mean value of the running azimuth angle data of the vehicle. The average azimuth angle of the first azimuth angle data can be reflected through the first azimuth angle average value, and the azimuth angle of the corresponding vehicle before the lane change driving is determined according to the first azimuth angle average value.
The first azimuth data may have a portion in which fluctuation is large. In this case, the first azimuth data mean determined from the first azimuth data may be affected by the partial data, and the obtained first azimuth data mean may not relatively accurately represent the traveling direction of the vehicle.
In order to solve the above problem, the first azimuth data may be divided before calculating the first azimuth data average value corresponding to the first azimuth data, and the first azimuth data average value may be calculated with the smallest variation width.
Calculating a first azimuth data mean value corresponding to the first azimuth data, including:
dividing the first azimuth angle data into a plurality of parts, calculating a third variance of each part of the first azimuth angle data, and calculating a mean value of the part of the first azimuth angle data with the minimum third variance as a mean value of the first azimuth angle data.
The first azimuth data is divided, and the number of divisions can be determined according to the first azimuth data. As an example, the first azimuth data may be divided into 5 shares.
And calculating a third variance of each piece of first azimuth data according to each piece of divided first azimuth data. The variation of each first azimuth data can be judged by the third variance of each first azimuth data. And selecting a piece of first azimuth angle data with the minimum third party difference, calculating the mean value of the first azimuth angle data, and taking the calculated mean value of the first azimuth angle data as the mean value of the first azimuth angle data.
S502: and calculating a second azimuth data mean value corresponding to the second azimuth data.
And calculating a second azimuth data mean value corresponding to the second azimuth data, and judging the driving direction of the vehicle after the lane change driving by using the second azimuth data mean value.
In the second azimuth data, there may be a case where the fluctuation of part of the data is large. Similarly, calculating a second azimuth data mean corresponding to the second azimuth data includes:
dividing the second azimuth data into a plurality of parts, calculating a fourth variance of each part of the second azimuth data, and calculating a mean value of the part of the second azimuth data with the smallest fourth variance as a mean value of the second azimuth data.
It should be noted that the number of the second azimuth data partitions may be independent of the number of the first azimuth data partitions, and the number of the second azimuth data partitions may be the same as or different from the number of the first azimuth data partitions. As an example, the second azimuth data may be divided into 5.
And calculating a fourth variance of each divided second azimuth data, and determining a second azimuth data with small fluctuation according to the fourth variance. And calculating the average value of the second azimuth data with the smallest fourth difference as the second azimuth data average value.
S503: and calculating the absolute value of the difference between the first azimuth data mean value and the second azimuth data mean value to obtain a target value.
After the first azimuth data mean value and the second azimuth data mean value are obtained, the absolute value of the difference between the first azimuth data mean value and the second azimuth data mean value is calculated. And taking the absolute value of the difference of the mean values of the azimuth data obtained by calculation as a target numerical value. The target value can represent the change condition of the azimuth angle of the vehicle before and after the lane change.
After the target value is obtained, the target value may be compared with a preset threshold value to determine the type of lane change driving behavior.
S504: and if the target value is greater than or equal to 0 and less than or equal to the first threshold value, determining that the type of the lane change driving behavior is lane change driving behavior.
For lane-change driving behavior, the difference in the directional angle of the vehicle before and after the lane change is small. And comparing the target value with a first threshold value, and determining the type of the lane change driving behavior as the lane change driving behavior if the target value is greater than or equal to 0 and less than or equal to the first threshold value. The first threshold may be a maximum threshold of the corresponding lane-change driving behavior, and a numerical value corresponding to the first threshold is smaller and may be 5. The first threshold may be determined based on an average difference in azimuth before and after the lane change of the vehicle during normal lane change driving.
S505: and if the target value is greater than or equal to the second threshold value, determining that the type of the lane change driving behavior is the U-turn driving behavior.
When the vehicle performs a u-turn driving behavior, the difference between the direction angles of the vehicle before and after the lane change is large, and even a difference of 180 degrees may occur. Therefore, when the target value is greater than or equal to the second threshold value, the type of the lane change driving behavior may be determined as the u-turn driving behavior. The second threshold is a minimum threshold corresponding to the u-turn driving behavior, and the second threshold may be determined according to an average difference of the azimuth angles before and after the u-turn driving behavior of the vehicle. The value of the second threshold may be larger, for example, the second threshold may be 175.
S506: and if the target value is greater than or equal to the third threshold value and less than or equal to the fourth threshold value, determining that the type of the lane change driving behavior is the turning driving behavior.
When the vehicle turns, the difference value of the corresponding direction angles before and after the vehicle turns is within a certain interval. When the target value is greater than or equal to the third threshold value and less than or equal to the fourth threshold value, the type of lane change driving behavior is determined as the turning driving behavior. Wherein the third threshold is the minimum threshold of the turning driving behavior, and the fourth threshold is the maximum threshold of the turning driving behavior. The third threshold and the fourth threshold may be determined according to an average difference of the azimuth angles before and after the vehicle turning driving behavior. In one possible implementation, the third threshold may be 60 and the fourth threshold may be 120.
In the embodiment of the present application, a first azimuth data mean value corresponding to the first azimuth data is calculated, and a second azimuth data mean value corresponding to the second azimuth data is calculated. By calculating the target value and comparing the target value with the corresponding threshold value, the driving behavior of the corresponding lane can be determined to be changed. By comparing the size relationship between the target value and the threshold value, the specific type of the behavior of the lane corresponding to the target value can be accurately determined.
Based on the method for identifying the vehicle driving behavior provided by the above method embodiment, the embodiment of the present application further provides an apparatus for identifying the vehicle driving behavior, which will be described below with reference to the accompanying drawings.
Referring to fig. 6, the drawing is a schematic structural diagram of an identification device for vehicle driving behavior according to an embodiment of the present application. As shown in fig. 6, the vehicle travel behavior recognition device includes:
the acquisition unit 601 is used for acquiring driving azimuth data in the driving process of the vehicle;
a first calculating unit 602, configured to calculate, when an acquisition duration reaches a preset duration, a data feature of driving azimuth data within the preset duration;
the input unit 603 is configured to input data characteristics of the driving azimuth data within the preset duration into a vehicle driving behavior classifier, and obtain a classification result output by the vehicle driving behavior classifier;
a first classification unit 604, configured to, when the classification result is normal driving, store driving azimuth data within the preset duration, and re-execute the driving azimuth data in the vehicle driving process and subsequent steps;
a second classification unit 605, configured to determine, when the classification result is that a lane changes, driving azimuth data within a preset time period that is newly stored as first azimuth data, and re-execute the step of collecting driving azimuth data and subsequent steps in the driving process of the vehicle until the classification result is obtained again as normal driving, and store the driving azimuth data within the preset time period as second azimuth data;
a determining unit 606, configured to determine a type of the driving behavior of the lane change according to the first azimuth data and the second azimuth data.
Optionally, the apparatus further comprises:
the system comprises an original data acquisition unit, a data processing unit and a data processing unit, wherein the original data acquisition unit is used for acquiring original driving azimuth data when the driving behavior of the vehicle changes in lanes of different types;
the segmentation unit is used for carrying out data segmentation on each piece of original driving azimuth data according to a sliding window to obtain driving azimuth data to be trained, and the length of the sliding window is preset duration;
the second calculation unit is used for calculating the data characteristics of the driving azimuth data to be trained;
and the training unit is used for training by using the data characteristics of the driving azimuth data to be trained and the classification result label value corresponding to the driving azimuth data to be trained to obtain the vehicle driving behavior classifier, wherein the classification result label value corresponding to the driving azimuth data to be trained comprises normal driving or lane change driving.
Optionally, the feature calculating unit is specifically configured to calculate a first variance, a first range, and a first average absolute deviation of the driving azimuth data to be trained as the data feature of the driving azimuth data to be trained.
Optionally, the vehicle driving behavior classifier is a naive bayes classifier;
the input unit is specifically configured to calculate, by using the naive bayes classifier, a first posterior probability that the data features of the travel azimuth data within the preset duration belong to a normal travel classification result and a second posterior probability that the data features of the travel azimuth data within the preset duration belong to a lane change travel classification result;
and if the first posterior probability is greater than the second posterior probability, determining that the classification result is normal driving, and if the first posterior probability is less than the second posterior probability, determining that the classification result is lane change driving.
Optionally, the first calculating unit is specifically configured to calculate, when the acquisition duration reaches a preset duration, a second variance, a second pole difference, and a second average absolute deviation value of the driving azimuth data within the preset duration as data characteristics of the driving azimuth data within the preset duration.
Optionally, the determining unit includes:
the first calculation subunit is used for calculating a first azimuth data mean value corresponding to the first azimuth data;
the second calculating subunit is configured to calculate a second azimuth data mean value corresponding to the second azimuth data;
the third calculation subunit is used for calculating the absolute value of the difference between the first azimuth data mean value and the second azimuth data mean value to obtain a target numerical value;
the first classification subunit is used for determining that the type of the lane change driving behavior is lane change driving behavior if the target value is greater than or equal to 0 and less than or equal to a first threshold value;
the second classification subunit is used for determining that the type of the driving behavior of the lane change is the U-turn driving behavior if the target value is greater than or equal to a second threshold value;
and the third classification subunit is used for determining that the type of the lane change driving behavior is the turning driving behavior if the target value is greater than or equal to a third threshold value and less than or equal to a fourth threshold value.
Optionally, the first calculating subunit is specifically configured to divide the first azimuth data into multiple parts, calculate a third variance of each part of the first azimuth data, and calculate a mean value of the part of the first azimuth data with the smallest third variance as a mean value of the first azimuth data;
the second calculating subunit is specifically configured to divide the second azimuth data into multiple parts, calculate a fourth variance of each part of the second azimuth data, and calculate a mean value of one part of the second azimuth data with the smallest fourth variance as a mean value of the second azimuth data.
In addition, the embodiment of the application also provides a vehicle driving behavior recognition device, which comprises: the vehicle driving behavior recognition method comprises a memory, a processor and a computer program which is stored on the memory and can run on the processor, wherein when the processor executes the computer program, the recognition method of the vehicle driving behavior is realized.
In addition, the embodiment of the present application further provides a computer-readable storage medium, where instructions are stored, and when the instructions are executed on a terminal device, the instructions cause the terminal device to execute the method for identifying the vehicle driving behavior according to the embodiment.
Therefore, the vehicle driving behaviors can be classified according to the data characteristics by acquiring the driving azimuth data of the vehicle driving and correspondingly calculating the data characteristics of the driving azimuth data within the preset duration. And when the classification result is that the lane is changed to drive, respectively taking the latest stored driving azimuth data within the preset duration and driving azimuth data within the preset duration of normal driving, which are acquired again later, as the first azimuth data and the second azimuth data. The type of the driving behavior of the lane change can be accurately obtained through the first azimuth data before the lane change and the second azimuth data after the lane change. And the driving azimuth data is collected, stored and analyzed in real time, so that the type of the driving behavior of the lane change can be identified and determined in time according to the driving azimuth data.
It should be noted that, in the present specification, the embodiments are described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments may be referred to each other. For the system or the device disclosed by the embodiment, the description is simple because the system or the device corresponds to the method disclosed by the embodiment, and the relevant points can be referred to the method part for description.
It should be understood that in the present application, "at least one" means one or more, "a plurality" means two or more. "and/or" for describing an association relationship of associated objects, indicating that there may be three relationships, e.g., "a and/or B" may indicate: only A, only B and both A and B are present, wherein A and B may be singular or plural. The character "/" generally indicates that the former and latter associated objects are in an "or" relationship. "at least one of the following" or similar expressions refer to any combination of these items, including any combination of single item(s) or plural items. For example, at least one (one) of a, b, or c, may represent: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", wherein a, b, c may be single or plural.
It is further noted that, herein, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other identical elements in a process, method, article, or apparatus that comprises the element.
The steps of a method or algorithm described in connection with the embodiments disclosed herein may be embodied directly in hardware, in a software module executed by a processor, or in a combination of the two. A software module may reside in Random Access Memory (RAM), memory, Read Only Memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the application. Thus, the present application is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (10)
1. A method for identifying a driving behavior of a vehicle, the method comprising:
collecting driving azimuth data in the driving process of a vehicle;
when the acquisition duration reaches a preset duration, calculating the data characteristics of the driving azimuth data in the preset duration;
inputting the data characteristics of the driving azimuth data within the preset duration into a vehicle driving behavior classifier, and obtaining a classification result output by the vehicle driving behavior classifier;
when the classification result is normal driving, storing the driving azimuth data within the preset duration, and re-executing the driving azimuth data in the vehicle driving process and the subsequent steps;
when the classification result is that the lane changes to drive, determining the newly stored driving azimuth data within the preset duration as first azimuth data, and re-executing the acquisition of the driving azimuth data in the driving process of the vehicle and the subsequent steps until the classification result is obtained again as normal driving, and storing the driving azimuth data within the preset duration as second azimuth data;
and determining the type of the driving behavior of the lane according to the first azimuth data and the second azimuth data.
2. The method of claim 1, further comprising:
collecting original driving azimuth data of a vehicle when driving behaviors of the vehicle change in lanes of different types;
performing data segmentation on each piece of original driving azimuth data according to a sliding window to obtain driving azimuth data to be trained, wherein the length of the sliding window is preset duration;
calculating the data characteristics of the driving azimuth data to be trained;
and training by using the data characteristics of the driving azimuth data to be trained and the classification result label value corresponding to the driving azimuth data to be trained to obtain a vehicle driving behavior classifier, wherein the classification result label value corresponding to the driving azimuth data to be trained comprises normal driving or lane change driving.
3. The method of claim 2, wherein the calculating the data characteristic of the driving azimuth data to be trained comprises:
and calculating a first variance, a first range and a first average absolute deviation value of the driving azimuth data to be trained as data characteristics of the driving azimuth data to be trained.
4. The method according to claim 1 or 2, characterized in that the vehicle driving behavior classifier is a naive bayes classifier;
the step of inputting the data characteristics of the driving azimuth data within the preset duration into a vehicle driving behavior classifier to obtain the classification result output by the vehicle driving behavior classifier comprises the following steps:
calculating a first posterior probability that the data characteristics of the driving azimuth data in the preset time length belong to a normal driving classification result and a second posterior probability that the data characteristics of the driving azimuth data in the preset time length belong to a lane change driving classification result by using the naive Bayes classifier;
and if the first posterior probability is greater than the second posterior probability, determining that the classification result is normal driving, and if the first posterior probability is less than the second posterior probability, determining that the classification result is lane change driving.
5. The method of claim 1, wherein calculating the data characteristic of the azimuth of travel data within a preset time period when the acquisition time period reaches the preset time period comprises:
and when the acquisition time length reaches a preset time length, calculating a second variance, a second polar difference and a second average absolute deviation value of the driving azimuth data in the preset time length as data characteristics of the driving azimuth data in the preset time length.
6. The method of claim 1, wherein determining the type of lane change driving behavior based on the first azimuth data and the second azimuth data comprises:
calculating a first azimuth data mean value corresponding to the first azimuth data;
calculating a second azimuth data mean value corresponding to the second azimuth data;
calculating the absolute value of the difference between the first azimuth data mean value and the second azimuth data mean value to obtain a target numerical value;
if the target value is greater than or equal to 0 and less than or equal to a first threshold value, determining that the type of the lane change driving behavior is lane change driving behavior;
if the target value is larger than or equal to a second threshold value, determining that the type of the lane change driving behavior is a U-turn driving behavior;
and if the target value is greater than or equal to a third threshold value and less than or equal to a fourth threshold value, determining that the type of the lane change driving behavior is the turning driving behavior.
7. The method of claim 6, wherein the calculating a mean value of the first azimuth data corresponding to the first azimuth data comprises:
dividing the first azimuth angle data into a plurality of parts, calculating a third variance of each part of the first azimuth angle data, and calculating a mean value of the part of the first azimuth angle data with the minimum third variance as a mean value of the first azimuth angle data;
the calculating a second azimuth data mean value corresponding to the second azimuth data includes:
dividing the second azimuth data into a plurality of parts, calculating a fourth variance of each part of second azimuth data, and calculating a mean value of the part of second azimuth data with the smallest fourth variance as a mean value of the second azimuth data.
8. An apparatus for recognizing a driving behavior of a vehicle, the apparatus comprising:
the acquisition unit is used for acquiring driving azimuth data in the driving process of the vehicle;
the first calculation unit is used for calculating the data characteristics of the driving azimuth data in the preset duration when the acquisition duration reaches the preset duration;
the input unit is used for inputting the data characteristics of the driving azimuth data within the preset duration into a vehicle driving behavior classifier and obtaining a classification result output by the vehicle driving behavior classifier;
the first classification unit is used for storing the driving azimuth data within the preset duration when the classification result is normal driving, and re-executing the driving azimuth data in the driving process of the collected vehicle and the subsequent steps;
the second classification unit is used for determining the latest stored driving azimuth data within the preset duration as the first azimuth data when the classification result indicates that the lane changes and the driving, and re-executing the acquisition of the driving azimuth data in the driving process of the vehicle and the subsequent steps until the classification result indicates that the vehicle normally drives again, and storing the driving azimuth data within the preset duration as the second azimuth data;
and the determining unit is used for determining the type of the driving behavior of the lane change according to the first azimuth angle data and the second azimuth angle data.
9. An apparatus for recognizing a running behavior of a vehicle, characterized by comprising: a memory, a processor, and a computer program stored on the memory and executable on the processor, when executing the computer program, implementing the method of identifying a driving behavior of a vehicle according to any one of claims 1 to 7.
10. A computer-readable storage medium, characterized in that instructions are stored therein, which, when run on a terminal device, cause the terminal device to execute the method of identifying a vehicle travel behavior according to any one of claims 1 to 7.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202011049638.6A CN112163521A (en) | 2020-09-29 | 2020-09-29 | Vehicle driving behavior identification method, device and equipment |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202011049638.6A CN112163521A (en) | 2020-09-29 | 2020-09-29 | Vehicle driving behavior identification method, device and equipment |
Publications (1)
Publication Number | Publication Date |
---|---|
CN112163521A true CN112163521A (en) | 2021-01-01 |
Family
ID=73861170
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202011049638.6A Pending CN112163521A (en) | 2020-09-29 | 2020-09-29 | Vehicle driving behavior identification method, device and equipment |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN112163521A (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112711054A (en) * | 2021-03-24 | 2021-04-27 | 腾讯科技(深圳)有限公司 | Object behavior detection method and device, electronic equipment and storage medium |
CN113997954A (en) * | 2021-11-29 | 2022-02-01 | 广州文远知行科技有限公司 | Vehicle driving intention prediction method, device and equipment and readable storage medium |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2019056498A1 (en) * | 2017-09-19 | 2019-03-28 | 平安科技(深圳)有限公司 | Driving model training method, driver recognition method, device, apparatus and medium |
CN109615855A (en) * | 2018-12-10 | 2019-04-12 | 北京新能源汽车股份有限公司 | Recognition method for changing driving behavior, storage medium and vehicle-mounted terminal |
WO2019205020A1 (en) * | 2018-04-25 | 2019-10-31 | 深圳市锐明技术股份有限公司 | Road condition recognition method, apparatus and device |
CN111209838A (en) * | 2019-12-31 | 2020-05-29 | 清华大学 | Driving intention-based dynamic identification method for lane change behavior of surrounding vehicles |
-
2020
- 2020-09-29 CN CN202011049638.6A patent/CN112163521A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2019056498A1 (en) * | 2017-09-19 | 2019-03-28 | 平安科技(深圳)有限公司 | Driving model training method, driver recognition method, device, apparatus and medium |
WO2019205020A1 (en) * | 2018-04-25 | 2019-10-31 | 深圳市锐明技术股份有限公司 | Road condition recognition method, apparatus and device |
CN109615855A (en) * | 2018-12-10 | 2019-04-12 | 北京新能源汽车股份有限公司 | Recognition method for changing driving behavior, storage medium and vehicle-mounted terminal |
CN111209838A (en) * | 2019-12-31 | 2020-05-29 | 清华大学 | Driving intention-based dynamic identification method for lane change behavior of surrounding vehicles |
Non-Patent Citations (2)
Title |
---|
何杰;杨娇;杭文;陈志伟;吴德华;: "考虑驾驶员行为特性的行车安全仿真试验", 解放军理工大学学报(自然科学版), no. 06, 15 December 2013 (2013-12-15) * |
范菁;吴青青;叶阳;董天阳;: "基于智能手机的车辆行为实时判别与渐进矫正方法研究", 计算机科学, no. 03, 15 March 2017 (2017-03-15) * |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112711054A (en) * | 2021-03-24 | 2021-04-27 | 腾讯科技(深圳)有限公司 | Object behavior detection method and device, electronic equipment and storage medium |
CN113997954A (en) * | 2021-11-29 | 2022-02-01 | 广州文远知行科技有限公司 | Vehicle driving intention prediction method, device and equipment and readable storage medium |
CN113997954B (en) * | 2021-11-29 | 2023-11-21 | 广州文远知行科技有限公司 | Method, device and equipment for predicting vehicle driving intention and readable storage medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Chen et al. | Pointlanenet: Efficient end-to-end cnns for accurate real-time lane detection | |
CN108090456B (en) | Training method for recognizing lane line model, and lane line recognition method and device | |
CN110197502B (en) | Multi-target tracking method and system based on identity re-identification | |
CN108460427B (en) | Classification model training method and device and classification method and device | |
CN110288627B (en) | Online multi-target tracking method based on deep learning and data association | |
CN110738690A (en) | unmanned aerial vehicle video middle vehicle speed correction method based on multi-target tracking framework | |
CN113076922B (en) | Object detection method and device | |
CN115620518B (en) | Intersection traffic conflict judging method based on deep learning | |
CN110781785A (en) | Traffic scene pedestrian detection method improved based on fast RCNN algorithm | |
CN117292338B (en) | Vehicle accident identification and analysis method based on video stream analysis | |
KR20180070258A (en) | Method for detecting and learning of objects simultaneous during vehicle driving | |
CN112163521A (en) | Vehicle driving behavior identification method, device and equipment | |
CN109784155B (en) | Visual target tracking method based on verification and error correction mechanism and intelligent robot | |
CN112101114B (en) | Video target detection method, device, equipment and storage medium | |
Qing et al. | A novel particle filter implementation for a multiple-vehicle detection and tracking system using tail light segmentation | |
CN112669277B (en) | Vehicle association method, computer equipment and device | |
CN109615007B (en) | Deep learning network target detection method based on particle filtering | |
CN117237911A (en) | Image-based dynamic obstacle rapid detection method and system | |
CN110765898A (en) | Method and device for determining object and key point thereof in image | |
CN113706580B (en) | Target tracking method, system, equipment and medium based on relevant filtering tracker | |
CN112258602B (en) | Stop line generation method and device, electronic equipment and storage medium | |
CN117456407B (en) | Multi-target image tracking method and device | |
CN116563769B (en) | Video target identification tracking method, system, computer and storage medium | |
CN113963027B (en) | Uncertainty detection model training method and device, and uncertainty detection method and device | |
CN110907928B (en) | Target object position determining 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 |