Method, device and equipment for identifying vehicle driving behavior

Abstract

The embodiment of the present application discloses a method, device and equipment for identifying vehicle driving behavior, which first collects driving azimuth angle data during vehicle driving; then calculates the data features of the driving azimuth angle data within a preset time length and inputs them into the vehicle driving behavior classifier to obtain the output classification result. When the classification result is normal driving, the driving azimuth angle data is stored and the above steps are executed again. When the classification result is that the lane changes, the driving azimuth angle data within the preset time length that has been stored is determined as the first azimuth angle data, and the above steps of collection and classification are executed again until the classification result is normal driving; then the driving azimuth angle data within the preset time length that has been stored is used as the second azimuth angle data; finally, the type of lane changing driving behavior is determined based on the first azimuth angle data and the second azimuth angle data. In this way, the type of lane changing driving behavior can be accurately and timely identified.

Description

Technical Field

[0001] The present application relates to the field of vehicle control, and specifically to a method, device and equipment for identifying vehicle driving behavior. Background Art

[0002] During the driving process, the driving behavior of the vehicle changes in real time. For example, when the vehicle turns, overtakes, or makes a U-turn, it is necessary to change lanes to drive the vehicle to the correct position.

[0003] For smart cars, the ability to accurately and quickly identify vehicle driving behavior is crucial. By determining a vehicle's driving behavior, its path can be predicted and its safety ensured. However, current methods for identifying lane changes cannot quickly and accurately identify these driving behaviors. Summary of the Invention

[0004] In view of this, embodiments of the present application provide a method, apparatus, and device for identifying vehicle driving behavior, which can quickly and accurately identify the type of vehicle lane change driving behavior.

[0005] To solve the above problems, the technical solutions provided in the embodiments of the present application are as follows:

[0006] A method for identifying vehicle driving behavior, the method comprising:

[0007] Collect driving azimuth angle data of the vehicle during driving;

[0008] When the collection time reaches a preset time, the data features of the driving azimuth angle data within the preset time are calculated;

[0009] Inputting the data features of the driving azimuth angle data within the preset time period into a vehicle driving behavior classifier to obtain a classification result output by the vehicle driving behavior classifier;

[0010] When the classification result is normal driving, storing the driving azimuth angle data within the preset time period, and re-performing the step of collecting the driving azimuth angle data during the vehicle driving process and subsequent steps;

[0011] When the classification result is that the lane change occurs, the most recently stored driving azimuth angle data within the preset time period is determined as the first azimuth angle data, and the steps of collecting the driving azimuth angle data during the vehicle's driving process and subsequent steps are re-executed until the classification result is again that the vehicle is driving normally, and the driving azimuth angle data within the preset time period is stored as the second azimuth angle data;

[0012] The type of lane-changing driving behavior is determined according to the first azimuth data and the second azimuth data.

[0013] In a possible implementation, the method further includes:

[0014] Collecting original driving azimuth angle data when a vehicle changes driving behavior in different types of lanes;

[0015] Each piece of the original driving azimuth angle data is segmented according to a sliding window to obtain driving azimuth angle data to be trained, where the length of the sliding window is a preset time length;

[0016] Calculating data features of the driving azimuth data to be trained;

[0017] A vehicle driving behavior classifier is trained using the data features of the driving azimuth angle data to be trained and the classification result label values ​​corresponding to the driving azimuth angle data to be trained. The classification result label values ​​corresponding to the driving azimuth angle data to be trained include normal driving or lane change driving.

[0018] In a possible implementation, the calculating of data features of the to-be-trained driving azimuth data includes:

[0019] A first variance, a first range, and a first mean absolute deviation of the to-be-trained driving azimuth data are calculated as data features of the to-be-trained driving azimuth data.

[0020] In a possible implementation, the vehicle driving behavior classifier is a naive Bayes classifier;

[0021] The step of inputting the data features of the driving azimuth angle data within the preset time period into a vehicle driving behavior classifier to obtain a classification result output by the vehicle driving behavior classifier includes:

[0022] Utilizing the naive Bayes classifier, calculating a first posterior probability that the data feature of the driving azimuth angle data within the preset duration belongs to a classification result of normal driving, and a second posterior probability that the data feature of the driving azimuth angle data within the preset duration belongs to a classification result of lane change driving;

[0023] If the first a posteriori probability is greater than the second a posteriori probability, the classification result is determined to be normal driving; if the first a posteriori probability is less than the second a posteriori probability, the classification result is determined to be driving with a lane change.

[0024] In a possible implementation, when the collection duration reaches a preset duration, calculating the data features of the driving azimuth angle data within the preset duration includes:

[0025] When the collection time reaches the preset time, the second variance, the second range and the second mean absolute deviation of the driving azimuth angle data within the preset time are calculated as data features of the driving azimuth angle data within the preset time.

[0026] In a possible implementation, determining the type of lane-changing driving behavior based on the first azimuth data and the second azimuth data includes:

[0027] Calculating a mean of the first azimuth angle data corresponding to the first azimuth angle data;

[0028] Calculating a mean value of the second azimuth angle data corresponding to the second azimuth angle data;

[0029] Calculating the absolute value of the difference between the mean of the first azimuth angle data and the mean of the second azimuth angle data to obtain a target value;

[0030] If the target value is greater than or equal to 0 and less than or equal to a first threshold, determining that the type of the lane change driving behavior is a lane change driving behavior;

[0031] If the target value is greater than or equal to a second threshold, determining that the type of the lane change driving behavior is a U-turn driving behavior;

[0032] If the target value is greater than or equal to the third threshold and less than or equal to the fourth threshold, it is determined that the type of the lane changing driving behavior is a turning driving behavior.

[0033] In a possible implementation, calculating a mean of the first azimuth data corresponding to the first azimuth data includes:

[0034] Divide the first azimuth angle data into multiple parts, calculate the third deviation of each part of the first azimuth angle data, and calculate the mean of the first azimuth angle data with the smallest third deviation as the mean of the first azimuth angle data;

[0035] The calculating the second azimuth data mean corresponding to the second azimuth data includes:

[0036] The second azimuth angle data is divided into multiple parts, the fourth variance of each part of the second azimuth angle data is calculated, and the mean of the part of the second azimuth angle data with the smallest fourth variance is calculated as the mean of the second azimuth angle data.

[0037] A vehicle driving behavior recognition device, comprising:

[0038] A collection unit, used for collecting driving azimuth angle data of the vehicle during driving;

[0039] A first calculation unit is configured to calculate data features of the driving azimuth angle data within a preset time period when the collection time period reaches a preset time period;

[0040] An input unit, configured to input data features of the driving azimuth angle data within the preset time period into a vehicle driving behavior classifier, and obtain a classification result output by the vehicle driving behavior classifier;

[0041] A first classification unit is configured to, when the classification result is normal driving, store the driving azimuth angle data within the preset time period, and re-execute the steps of collecting the driving azimuth angle data during the vehicle driving process and subsequent steps;

[0042] a second classification unit configured to, when the classification result is lane change, determine the most recently stored driving azimuth angle data within a preset time period as first azimuth angle data, and re-execute the steps of collecting driving azimuth angle data during vehicle driving and subsequent steps until the classification result is normal driving again, and store the driving azimuth angle data within the preset time period as second azimuth angle data;

[0043] A determination unit is used to determine the type of lane changing driving behavior based on the first azimuth data and the second azimuth data.

[0044] A vehicle driving behavior recognition device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the above-mentioned vehicle driving behavior recognition method is implemented.

[0045] A computer-readable storage medium stores instructions, which, when executed on a terminal device, enable the terminal device to execute the above-mentioned method for identifying vehicle driving behavior.

[0046] It can be seen that the embodiments of the present application have the following beneficial effects:

[0047] An embodiment of the present application provides a method for identifying vehicle driving behavior, which collects driving azimuth angle data during vehicle driving; when the collection time reaches a preset time, calculates the data features of the driving azimuth angle data within the preset time; by inputting the data features of the driving azimuth angle data within the preset time into a vehicle driving behavior classifier, a classification result output by the vehicle driving behavior classifier can be obtained; when the classification result is normal driving, the driving azimuth angle data within the preset time is stored, and the above steps are re-executed; when the classification result is lane change driving, the most recently stored driving azimuth angle data within the preset time is determined as the first azimuth angle data, and the collection of driving azimuth angle data and subsequent steps are re-executed until the classification result is normal driving; the driving azimuth angle data within the preset time is stored as the second azimuth angle data; finally, the type of lane change driving behavior is determined based on the first azimuth angle data and the second azimuth angle data.

[0048] Thus, by collecting the driving azimuth angle data of the vehicle and calculating the data features of the driving azimuth angle data within a preset time period, the vehicle driving behavior can be classified according to the data features. When the classification result is that the lane change occurs, the most recently stored driving azimuth angle data within the preset time period and the driving azimuth angle data of normal driving stored within the preset time period that are subsequently collected are used as the first azimuth angle data and the second azimuth angle data, respectively. The type of lane change driving behavior can be accurately obtained by using the first azimuth angle data before the lane change and the second azimuth angle data after the lane change. Moreover, by collecting, storing and analyzing the driving azimuth angle data in real time, the type of lane change driving behavior can be identified and determined in a timely manner based on the driving azimuth angle data. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 A schematic diagram of a framework of an exemplary application scenario provided in an embodiment of the present application;

[0050] Figure 2 A flowchart of a method for identifying vehicle driving behavior provided in an embodiment of the present application;

[0051] Figure 3 Schematic diagram of three vehicle lane changes provided in the embodiments of the present application;

[0052] Figure 4 A flowchart of a method for training a vehicle driving behavior classifier provided in an embodiment of the present application;

[0053] Figure 5 A flowchart of a method for determining the type of lane-changing driving behavior provided in an embodiment of the present application;

[0054] Figure 6 A schematic diagram of the structure of a vehicle driving behavior recognition device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0055] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, the embodiments of the present application are further described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0056] In order to facilitate understanding and explanation of the technical solutions provided by the embodiments of the present application, the background technology of the present application will be described below.

[0057] After researching traditional methods for identifying vehicle driving behavior, the inventors discovered that existing methods rely on analyzing camera data or radar-based vehicle driving data. These methods place high demands on storage and computing equipment, making them inadequate for intelligent vehicle control. Furthermore, these methods are unable to identify specific types of vehicle driving behavior. Based on this, an embodiment of the present application provides a method for identifying vehicle driving behavior, by collecting driving azimuth angle data during vehicle driving, when the collection time reaches a preset time, calculating the data features of the driving azimuth angle data within the preset time; by inputting the data features of the driving azimuth angle data within the preset time into a vehicle driving behavior classifier, the classification result output by the vehicle driving behavior classifier can be obtained; when the classification result is normal driving, the driving azimuth angle data within the preset time is stored, and the above steps are re-executed; when the classification result is lane change driving, the latest stored driving azimuth angle data within the preset time is determined as the first azimuth angle data, and the collection of driving azimuth angle data and subsequent steps are re-executed until the classification result is normal driving; the driving azimuth angle data within the preset time is stored as the second azimuth angle data; finally, the type of lane change driving behavior is determined based on the first azimuth angle data and the second azimuth angle data.

[0058] In order to facilitate understanding of the vehicle driving behavior recognition method provided in the embodiment of the present application, the following Figure 1 See the example scenario shown. Figure 1 , which is a framework diagram of an exemplary application scenario provided in an embodiment of the present application.

[0059] In practical applications, the vehicle's driving azimuth angle data is first collected during driving. When the collection duration reaches a preset duration, the data features of the driving azimuth angle data collected during the preset duration are calculated and input into the vehicle driving behavior classifier to obtain a classification result. If the classification result is normal driving, it means that the vehicle lane has not changed at this time. The driving azimuth angle data within the current preset duration is stored, and the collection, data feature extraction, and classification of the driving azimuth angle data continue. If the classification result continues to be normal driving, the driving azimuth angle data within the current preset duration is stored again, and the collection, data feature extraction, and classification of the driving azimuth angle data continue, and so on.

[0060] If the classification result indicates a lane change, the vehicle's lane has changed. The most recently stored driving azimuth angle data within a preset duration (i.e., the driving azimuth angle data stored within a preset duration when the previous classification result was normal driving) is obtained as the first azimuth angle data. The driving azimuth angle data is then collected, and data features are extracted and classified. This continues until the classification result changes back to normal driving, indicating that the lane change has been completed and the vehicle returns to normal driving. The driving azimuth angle data within the current preset duration is stored as the second azimuth angle data. Finally, the type of lane change behavior is determined based on the first and second azimuth angle data.

[0061] Those skilled in the art will understand that Figure 1 The framework diagram shown is only an example in which the embodiments of the present application can be implemented. The scope of application of the embodiments of the present application is not limited by any aspect of the framework.

[0062] To facilitate understanding of the present application, a method for identifying vehicle driving behavior provided in an embodiment of the present application is described below with reference to the accompanying drawings.

[0063] See also Figure 2 , which is a flow chart of a method for identifying vehicle driving behavior provided by an embodiment of the present application, such as Figure 2 As shown, the method may include S201-S206:

[0064] S201: Collecting driving azimuth angle data of the vehicle during driving.

[0065] When a vehicle is driving, it will change its route by adjusting its driving azimuth angle. When it needs to change lanes, turn around or make a turn, it is necessary to control the vehicle's driving azimuth angle to adjust the vehicle's driving.

[0066] The driving bearing angle can be defined as the angle between the vehicle's direction of travel and true north. A clockwise direction can be defined as a positive direction, while a counterclockwise direction can be defined as a negative direction. The driving bearing angle data can range from positive 180 degrees to negative 180 degrees.

[0067] Driving azimuth angle data can be collected by equipment installed on the vehicle, such as a global positioning system device or a magnetic needle device. During the vehicle's driving process, the driving azimuth angle data of the vehicle is collected so that the vehicle's driving behavior can be identified based on the driving azimuth angle data.

[0068] S202: When the collection time reaches a preset time, the data features of the driving azimuth angle data within the preset time are calculated.

[0069] It is understandable that the driving behavior of a vehicle needs to be identified through driving azimuth angle data over a period of time, and the driving behavior of the vehicle cannot be accurately determined by only the driving azimuth angle data at a single moment.

[0070] When collecting driving azimuth angle data, a preset duration can be set. When the collection duration reaches the preset duration, the driving azimuth angle data within the preset duration can be processed. The preset duration can be determined based on the average operation duration of driving behaviors such as lane changes and turns during vehicle operation. In one possible implementation, the preset duration can be set to 3 seconds.

[0071] After obtaining the driving azimuth angle data within a preset duration, data features of the driving azimuth angle data within the preset duration can be calculated. The data features of the driving azimuth angle data within the preset duration can be time-domain features of the driving azimuth angle data, including but not limited to variance, range, and average. The data features of the driving azimuth angle data within the preset duration can represent the driving behavior of the vehicle within the preset duration.

[0072] In one possible implementation, the fluctuation of the driving azimuth angle data within a preset time period can be calculated as a data feature of the driving azimuth angle data. This embodiment of the application also provides a method for calculating data features, which will be described in detail in the following specific implementations.

[0073] S203: Inputting data features of the driving azimuth angle data within a preset time period into a vehicle driving behavior classifier to obtain a classification result output by the vehicle driving behavior classifier.

[0074] The obtained data features are input into the vehicle driving behavior classifier. According to the classification results output by the vehicle driving behavior classifier, the driving behavior of the vehicle within a preset time period can be determined, and different driving behaviors can be processed accordingly.

[0075] The vehicle driving behavior classifier outputs a classification result of normal driving or lane-changing driving. Normal driving means that the data characteristics of the driving azimuth angle data within a preset duration represent normal vehicle driving, with no other driving behaviors. Lane-changing driving means that the data characteristics of the driving azimuth angle data within a preset duration represent lane-changing driving. Based on the classification result output by the vehicle driving behavior classifier, a preliminary judgment of the vehicle's driving behavior can be made, and the driving azimuth angle data can be processed accordingly based on the vehicle's driving behavior.

[0076] The vehicle driving behavior classifier is a classification model that can realize data feature classification. The vehicle driving behavior classifier can be pre-trained based on data features that represent different vehicle driving behaviors.

[0077] The present application also provides a method for training a vehicle driving behavior classifier. Furthermore, when the vehicle driving behavior classifier is a naive Bayesian classifier, a method for classifying vehicle driving behaviors based on the data features of driving azimuth angle data is provided. For details, please refer to the specific implementation method below.

[0078] S204: When the classification result is normal driving, the driving azimuth angle data within a preset time period is stored, and the collection of the driving azimuth angle data during the vehicle driving process and subsequent steps are re-executed.

[0079] When the classification result is normal driving, the vehicle lane has not changed, and there is no need to determine the type of driving behavior of the vehicle lane change through the driving azimuth angle data.

[0080] The driving azimuth angle data within a preset time period is stored so that after determining that the vehicle lane changes, the stored driving azimuth angle data can be used to identify the type of lane change driving behavior.

[0081] In a possible implementation, the driving azimuth angle data within a preset time period may be stored in a storage device, such as a data server.

[0082] After storing the driving azimuth angle data within a preset duration, in order to be able to identify the vehicle's driving behavior in real time, S201 is re-executed to collect the driving azimuth angle data during the vehicle's driving process, and the subsequent steps. Then, the classification result corresponding to the data characteristics of the driving azimuth angle data for the next preset duration is obtained, and the driving azimuth angle data is processed accordingly based on the classification result. If the classification result corresponding to the data characteristics of the driving azimuth angle data for the next preset duration is still normal driving, the driving azimuth angle data for the preset duration is saved and the collection and subsequent steps of S201 are continued. If the classification result corresponding to the data characteristics of the driving azimuth angle data for the next preset duration is lane change, the subsequent steps corresponding to lane change in S205 are executed.

[0083] S205: When the classification result is that the lane has changed, the driving azimuth angle data within the preset time length that has been stored most recently is determined as the first azimuth angle data, and the driving azimuth angle data collected during the vehicle's driving process and subsequent steps are re-executed until the classification result is normal driving again, and the driving azimuth angle data within the preset time length is stored as the second azimuth angle data.

[0084] When the classification result is that the lane is changed, it is necessary to further determine the type of the lane-changing driving behavior.

[0085] When a vehicle changes lanes, the vehicle's driving azimuth angle data will change accordingly. And after the lane change, the driving azimuth angle data corresponding to different types of lane change behaviors are also different. For example, see Figure 3 As shown, the embodiment of the present application provides three schematic diagrams of vehicles changing lanes. Figure 3 As shown in a, when a vehicle performs a normal lane change, the direction of the vehicle before and after the normal lane change is basically the same, and the corresponding driving azimuth angle data of the vehicle before and after the normal lane change are basically the same. Figure 3 As shown in b, the driving azimuth angle data of the vehicle before and after the vehicle turns should be opposite. Figure 3 As shown in c, the driving azimuth angle data of the vehicle before and after the vehicle turns are different, and there is a certain angle difference.

[0086] Therefore, after determining that the vehicle lane changes, the type of the vehicle lane change driving behavior can be determined based on the driving azimuth angle data before and after the vehicle lane changes.

[0087] The driving azimuth angle data before the vehicle changes lanes may be the most recently stored driving azimuth angle data within a preset time period. The most recently stored driving azimuth angle data within the preset time period is determined as the first azimuth angle data.

[0088] For the driving azimuth angle data after the vehicle changes lanes, after determining that the classification result of the driving azimuth angle data within the current preset duration is lane change driving, S201 of collecting the driving azimuth angle data during vehicle driving and subsequent steps may be re-executed until the classification result again is normal driving. The driving azimuth angle data within the preset duration corresponding to this time is the driving azimuth angle data after the vehicle changes lanes. The driving azimuth angle data within the preset duration for which the classification result is normal driving, obtained after re-execution of the driving azimuth angle data collection and subsequent steps, is stored and used as the second azimuth angle data.

[0089] S206: Determine the type of lane-changing driving behavior based on the first azimuth data and the second azimuth data.

[0090] According to the obtained first azimuth angle data and the second azimuth angle data, the change in the azimuth angle 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.

[0091] Lane changing driving behaviors include normal lane changing, U-turn and turning, etc.

[0092] An embodiment of the present application also provides a method for determining the type of lane-changing driving behavior based on first azimuth angle data and second azimuth angle data. Please see below for details.

[0093] In an embodiment of the present application, by collecting the driving azimuth angle data of the vehicle and calculating the data features of the driving azimuth angle data within a preset time period, the classification result corresponding to the driving azimuth angle data can be obtained by the vehicle driving behavior classifier. By timely collecting the driving azimuth angle data of the vehicle and classifying it according to the data features, the driving behavior of the vehicle can be classified in a timely manner, which facilitates the determination of the type of lane-changing driving behavior. When the lane is changed, the type of lane-changing driving behavior can be determined by the driving azimuth angle data before the lane change and the driving azimuth angle data after the lane change, that is, the first azimuth angle data and the second azimuth angle data. The type of lane-changing driving behavior of the vehicle can be more accurately identified by the driving azimuth angle data before and after the lane change.

[0094] Before using the vehicle driving behavior classifier to classify the data features of the driving azimuth angle data input within a preset time period, the vehicle driving behavior classifier needs to be trained first.

[0095] See also Figure 4, which is a flow chart of a method for training a vehicle driving behavior classifier provided in an embodiment of the present application. In the embodiment of the present application, in addition to the above-mentioned method for identifying vehicle driving behavior, it also includes S401-S404:

[0096] S401: Collecting original driving azimuth angle data when a vehicle changes driving behavior in lanes belonging to different types.

[0097] For different types of lane change behaviors, raw driving heading angle data is collected. This raw driving heading angle data can include all driving heading angle data during the lane change process. The collection duration of this raw driving heading angle data can be the length of time it takes for the vehicle to complete a lane change. As an example, the collection duration of this raw driving heading angle data can be 15 seconds.

[0098] By collecting the original driving azimuth angle data belonging to different types of lane-changing driving behaviors, the driving azimuth angle data corresponding to different types of lane-changing driving behaviors can be obtained, so that the corresponding data features can be extracted subsequently using the collected original driving azimuth angle data.

[0099] To improve the classification accuracy of the trained vehicle driving behavior classifier, multiple sets of raw driving azimuth angle data corresponding to different types of lane change driving behaviors may be collected. In one possible implementation, at least 50 sets of raw driving azimuth angle data corresponding to different types of lane change driving behaviors may be collected.

[0100] S402: Each piece of original driving azimuth angle data is segmented according to a sliding window to obtain driving azimuth angle data to be trained. The length of the sliding window is a preset time length.

[0101] Each piece of raw driving heading angle data is segmented using a sliding window approach. The sliding window can have a preset duration. Within the preset duration, data from the vehicle's lane change behavior can be captured. The preset duration can be shorter than the duration of the raw driving heading angle data collection. For example, the preset duration can be 3 seconds.

[0102] In addition, to prevent the original driving azimuth angle data from being lost during data segmentation, when the sliding windows are moved, there may be partially overlapping data between adjacent sliding windows. In one possible implementation, there may be 50% overlapping data between adjacent sliding windows.

[0103] For example, if the raw driving azimuth angle data collection time is 15 seconds, the preset time is 3 seconds, and there is 50% overlap between adjacent sliding windows, then after data segmentation for each set of raw driving azimuth angle data, 9 training driving azimuth angle data sets can be obtained. The corresponding calculation formula is 15 / (3*50%)-1=9.

[0104] By sliding the window, the driving azimuth angle data to be trained corresponding to each sliding window is obtained, thereby realizing data segmentation of the original driving azimuth angle data.

[0105] S403: Calculate data features of the driving azimuth angle data to be trained.

[0106] After the data is segmented to obtain the driving azimuth angle data to be trained, the data features of each driving azimuth angle data to be trained are calculated.

[0107] The data features of the driving azimuth angle data to be trained can reflect the characteristics of the driving azimuth angle data to be trained under driving behaviors, and can reflect the characteristics of the driving azimuth angle data to be trained under different driving behaviors. The data features are used to train the vehicle driving behavior classifier.

[0108] The data feature may be a time domain feature of the driving azimuth angle data to be trained. The embodiment of the present application further provides a method for calculating the time domain feature of the driving azimuth angle data to be trained, as described below.

[0109] S404: A vehicle driving behavior classifier is trained using the data features of the driving azimuth angle data to be trained and the classification result label values ​​corresponding to the driving azimuth angle data to be trained. The classification result label values ​​corresponding to the driving azimuth angle data to be trained include normal driving or lane change driving.

[0110] The obtained data features of the driving azimuth angle data to be trained and the classification result label values ​​corresponding to each driving azimuth angle data to be trained are used to train and generate a vehicle driving behavior classifier.

[0111] The classification result label values ​​corresponding to the training driving azimuth angle data include normal driving and lane-changing driving. The classification result label values ​​can be determined based on the driving behavior of the vehicle corresponding to the training driving azimuth angle data. The training driving azimuth angle data with a classification result label value of normal driving can be the training driving azimuth angle data before or after the vehicle has completed a lane change. The training driving azimuth angle data with a lane-changing driving azimuth angle data can be the training driving azimuth angle data during the vehicle's lane change.

[0112] Based on the above content, it can be seen that in the embodiment of the present application, by collecting the original driving azimuth angle data of the vehicle when the driving behavior changes in different types of lanes, and segmenting the original driving azimuth angle data, the data features of the segmented driving azimuth angle data to be trained can be obtained. Using the data features of the driving azimuth angle data to be trained and the classification result label values ​​corresponding to the driving azimuth angle data to be trained, the corresponding vehicle driving behavior classifier can be trained and generated. By segmenting the original driving azimuth angle data in a sliding window manner, the driving azimuth angle data to be trained in different driving behaviors during the vehicle lane change process can be obtained. In this way, the driving azimuth angle data to be trained for different driving behaviors can be used to train the vehicle driving behavior classifier to obtain a more accurate vehicle driving behavior classifier.

[0113] Furthermore, specifically, the data features of the driving azimuth angle data to be trained may include variance, range, and mean absolute deviation.

[0114] Calculate the data features of the driving azimuth data to be trained, including:

[0115] The first variance, the first range, and the first mean absolute deviation of the driving azimuth angle data to be trained are calculated as data features of the driving azimuth angle data to be trained.

[0116] The fluctuations in the values ​​of the training driving azimuth angle data vary under different driving behaviors. During normal driving, since the vehicle maintains a certain direction, the corresponding training driving azimuth angle data will fluctuate slightly around a certain value. However, when changing lanes, the vehicle will continuously adjust its direction, and the corresponding training driving azimuth angle data will fluctuate more significantly according to a certain development trend.

[0117] In order to characterize the fluctuation of the driving azimuth angle data to be trained, the data features of the driving azimuth angle data to be trained can be calculated. The data features are three-dimensional feature vectors consisting of the first variance, first range, and first mean absolute deviation of the driving azimuth angle data to be trained.

[0118] The first variance is the average of the squared differences between the values ​​of each training driving azimuth angle data and the mean of the training driving azimuth angle data. The first variance can characterize the degree of dispersion of the training driving azimuth angle data. The first range is the difference between the maximum and minimum values ​​in the training driving azimuth angle data, which can also characterize the degree of dispersion of the training driving azimuth angle data. The first mean absolute deviation is the average of the absolute values ​​of the deviations between individual training driving azimuth angle data and the arithmetic mean of the training driving azimuth angle data. The first mean absolute deviation can prevent errors from canceling each other out and accurately reflect the magnitude of data errors.

[0119] By calculating the first variance, first range, and first mean absolute deviation of the training driving azimuth data, the fluctuation of the training driving azimuth data can be accurately reflected from multiple perspectives. Using the data characteristics of the training driving azimuth data to train the vehicle driving behavior classifier makes the resulting vehicle driving behavior classifier more accurate, thereby improving the accuracy of vehicle driving behavior classification and enabling more accurate recognition of lane change behavior.

[0120] When classifying the data features of the vehicle's driving azimuth angle data, it is necessary to ensure the real-time classification so as to quickly determine the classification results corresponding to the data features of the vehicle's driving azimuth angle data and identify the vehicle's driving behavior.

[0121] In a possible implementation, the vehicle driving behavior classifier may use a naive Bayes classifier with a shorter execution time, so as to facilitate online and timely classification of data features of the vehicle driving azimuth data.

[0122] Based on the above-mentioned training method for the vehicle driving behavior classifier, when the vehicle driving behavior classifier is a naive Bayes classifier, the embodiment of the present application also provides a method for obtaining classification results using the naive Bayes classifier.

[0123] Inputting the data features of the driving azimuth angle data within a preset time period into the vehicle driving behavior classifier to obtain the classification results output by the vehicle driving behavior classifier includes the following two steps:

[0124] A1: Use a naive Bayes classifier to calculate the first a posteriori probability that the data features of the driving azimuth angle data within a preset time period belong to the normal driving classification result, and the second a posteriori probability that the data features of the driving azimuth angle data within a preset time period belong to the lane change driving classification result.

[0125] The data features of the driving azimuth angle data within the preset time length are input into the naive Bayes classifier to calculate the first posterior probability that the data features of the driving azimuth angle data within the preset time length belong to the normal driving classification result, and the second posterior probability that the data features of the driving azimuth angle data within the preset time length belong to the lane change driving classification result.

[0126] By using a naive Bayesian classifier to calculate the posterior probabilities of the driving heading angle data features corresponding to the normal driving classification results and the lane change driving classification results, the probabilities of the driving heading angle data features corresponding to the two different driving classification results can be obtained. The classification result can then be determined based on the relationship between the first posterior probability and the second posterior probability.

[0127] A2: If the first posterior probability is greater than the second posterior probability, the classification result is determined to be normal driving. If the first posterior probability is less than the second posterior probability, the classification result is determined to be lane change driving.

[0128] When the first a posteriori probability is greater than the second a posteriori probability, the probability that the data feature of the driving azimuth angle data corresponds to the normal driving classification result is greater, and the data feature of the driving azimuth angle data corresponds to the normal driving classification result. Conversely, when the second a posteriori probability is greater than the first a posteriori probability, the probability that the data feature of the driving azimuth angle data corresponds to the lane change driving classification result is greater, and the data feature of the driving azimuth angle data corresponds to the lane change driving classification result.

[0129] Based on the above, it can be seen that using a naive Bayesian classifier as a vehicle driving behavior classifier can improve the efficiency of classifying the data features of driving azimuth angle data. When determining the classification result, by first calculating the first and second posterior probabilities corresponding to the data features of the driving azimuth angle data within a preset duration and then comparing them, the classification result to which the data features of the driving azimuth angle data within the preset duration belong can be quickly determined. This allows for timely classification of driving azimuth angle data, improving the efficiency of determining driving behavior.

[0130] In a possible implementation, the data features of the driving azimuth angle data within a preset time period may include variance, range, and mean absolute deviation.

[0131] In a method for calculating data features of driving azimuth angle data provided in an embodiment of the present application, when the collection time reaches a preset time, calculating the data features of the driving azimuth angle data within the preset time includes:

[0132] When the collection time reaches the preset time, the second variance, the second range and the second mean absolute deviation of the driving azimuth angle data within the preset time are calculated as data features of the driving azimuth angle data within the preset time.

[0133] When the collection time reaches the preset time, the second variance, second range and second mean absolute deviation of the driving azimuth angle data within the preset time are calculated as data features to form a three-dimensional feature vector corresponding to the driving azimuth angle data within the preset time, so that the obtained data features can be input into the vehicle driving behavior classifier to obtain the corresponding classification results.

[0134] The method for calculating the second variance, second range and second mean absolute deviation of the driving azimuth angle data within a preset time length is similar to the above-mentioned method for calculating the first variance, first range and first mean absolute deviation of the driving azimuth angle data to be trained, and will not be repeated here.

[0135] When determining the lane change driving behavior type based on the first azimuth angle data and the second azimuth angle data, the determination may be made using predetermined thresholds corresponding to different lane change driving behavior types.

[0136] For details, see Figure 5 , which is a flow chart of a method for determining the type of lane-changing driving behavior provided by an embodiment of the present application. Determining the type of lane-changing driving behavior based on the first azimuth data and the second azimuth data includes S501-S506:

[0137] S501: Calculate a first azimuth angle data mean corresponding to the first azimuth angle data.

[0138] After obtaining the first azimuth angle data, a first azimuth angle mean value corresponding to the first azimuth angle data is calculated. The first azimuth angle mean value is an average value of the vehicle's driving azimuth angle data. The first azimuth angle mean value can reflect the average azimuth angle of the first azimuth angle data, making it easier to determine the azimuth angle of the corresponding vehicle before the lane change based on the first azimuth angle mean value.

[0139] The first azimuth angle data may have a portion with large fluctuations. In this case, the first azimuth angle data mean value determined based on the first azimuth angle data may be affected by some of the data, and the obtained first azimuth angle data mean value may not accurately represent the vehicle's driving direction.

[0140] To address the above problem, before calculating the first azimuth data mean corresponding to the first azimuth data, the first azimuth data may be divided first, and the first azimuth data mean may be calculated using the data with the smallest variation.

[0141] Calculating the first azimuth angle data mean corresponding to the first azimuth angle data, including:

[0142] The first azimuth angle data is divided into multiple parts, the third deviation of each part of the first azimuth angle data is calculated, and the mean of the part of the first azimuth angle data with the smallest third deviation is calculated as the mean of the first azimuth angle data.

[0143] First, the first azimuth angle data is divided, and the number of divisions can be determined according to the first azimuth angle data. As an example, the first azimuth angle data can be divided into 5 parts.

[0144] Then, based on each of the divided first azimuth angle data, a third deviation is calculated for each of the first azimuth angle data. The third deviation of each of the first azimuth angle data can be used to determine the change in each of the first azimuth angle data. The first azimuth angle data with the smallest third deviation is selected, and the mean of its first azimuth angle data is calculated. This calculated mean of the first azimuth angle data is used as the mean of the first azimuth angle data.

[0145] S502: Calculate a second azimuth data mean corresponding to the second azimuth data.

[0146] The second azimuth angle data mean corresponding to the second azimuth angle data is calculated, and the driving direction of the vehicle after the lane change can also be determined using the second azimuth angle data mean.

[0147] In the second azimuth angle data, some data may fluctuate greatly. Similarly, calculating the second azimuth angle data mean corresponding to the second azimuth angle data includes:

[0148] The second azimuth angle data is divided into multiple parts, the fourth variance of each part of the second azimuth angle data is calculated, and the mean of the part of the second azimuth angle data with the smallest fourth variance is calculated as the mean of the second azimuth angle data.

[0149] It should be noted that the number of shares into which the second azimuth data is divided may be independent of the number of shares into which the first azimuth data is divided, and the number of shares into which the second azimuth data is divided may be the same as or different from the number of shares into which the first azimuth data is divided. As an example, the second azimuth data may be divided into five shares.

[0150] Calculate the fourth variance of each divided portion of the second azimuth angle data, determine the portion of the second azimuth angle data with the smallest fluctuation based on the fourth variance, and calculate the mean of the portion of the second azimuth angle data with the smallest fourth variance as the mean of the second azimuth angle data.

[0151] S503: Calculate the absolute value of the difference between the mean value of the first azimuth angle data and the mean value of the second azimuth angle data to obtain a target value.

[0152] After obtaining the first azimuth angle data mean and the second azimuth angle data mean, the absolute value of the difference between the first azimuth angle data mean and the second azimuth angle data mean is calculated. The calculated absolute value of the difference between the azimuth angle data mean is used as a target value. The target value can represent the change in the azimuth angle of the vehicle before and after the lane change.

[0153] After the target value is obtained, it can be compared with a preset threshold to determine the type of lane changing driving behavior.

[0154] S504: If the target value is greater than or equal to 0 and less than or equal to the first threshold, determine that the type of the lane change driving behavior is the lane change driving behavior.

[0155] For lane change driving behavior, the difference in the vehicle's heading angle before and after the lane change is small. The target value is compared with a first threshold. If the target value is greater than or equal to 0 and less than or equal to the first threshold, the lane change driving behavior is determined to be a lane change driving behavior. The first threshold can be the maximum threshold for the corresponding lane change driving behavior. The value corresponding to the first threshold is relatively small, such as 5. The first threshold can be determined based on the average difference in the vehicle's heading angle before and after the lane change during normal lane change driving.

[0156] S505: If the target value is greater than or equal to the second threshold, determine that the type of the lane change driving behavior is a U-turn driving behavior.

[0157] When a vehicle performs a U-turn, the difference in the vehicle's heading angle before and after the lane change is significant, potentially even reaching 180 degrees. Therefore, when the target value is greater than or equal to the second threshold, the lane change behavior can be determined as a U-turn. The second threshold is the minimum threshold corresponding to a U-turn and can be determined based on the average difference in the vehicle's heading angle before and after the U-turn. The second threshold can be relatively large, for example, 175 degrees.

[0158] S506: If the target value is greater than or equal to the third threshold and less than or equal to the fourth threshold, determine that the type of the lane change driving behavior is a turning driving behavior.

[0159] When a vehicle turns, the difference in the corresponding azimuth angles before and after the turn is within a certain range. When the target value is greater than or equal to the third threshold and less than or equal to the fourth threshold, the lane change driving behavior is determined to be a turning driving behavior. The third threshold is the minimum threshold for turning driving behavior, and the fourth threshold is the maximum threshold for turning driving behavior. Both the third and fourth thresholds can be determined based on the average difference in the azimuth angles before and after the vehicle turns. In one possible implementation, the third threshold can be 60 and the fourth threshold can be 120.

[0160] In this embodiment of the present application, the mean of the first azimuth angle data corresponding to the first azimuth angle data is first calculated, and the mean of the second azimuth angle data corresponding to the second azimuth angle data is first calculated. By calculating a target value and comparing it with a corresponding threshold value, the corresponding lane change driving behavior can be determined. By comparing the magnitude relationship between the target value and the threshold value, the specific type of lane change behavior corresponding to the target value can be more accurately determined.

[0161] Based on the vehicle driving behavior identification method provided by the above method embodiment, the embodiment of the present application also provides a vehicle driving behavior identification device, and the vehicle driving behavior identification device will be described below with reference to the accompanying drawings.

[0162] See also Figure 6 , which is a structural diagram of a vehicle driving behavior recognition device provided by an embodiment of the present application. Figure 6 As shown, the vehicle driving behavior recognition device includes:

[0163] The acquisition unit 601 is used to acquire driving azimuth angle data of the vehicle during driving;

[0164] The first calculation unit 602 is configured to calculate the data features of the driving azimuth angle data within a preset time period when the collection time period reaches a preset time period;

[0165] An input unit 603 is configured to input the data features of the driving azimuth angle data within the preset time period into a vehicle driving behavior classifier to obtain a classification result output by the vehicle driving behavior classifier;

[0166] The first classification unit 604 is configured to, when the classification result is normal driving, store the driving azimuth angle data within the preset time period, and re-execute the steps of collecting the driving azimuth angle data during the vehicle driving process and subsequent steps;

[0167] The second classification unit 605 is configured to, when the classification result is lane change, determine the most recently stored driving azimuth angle data within a preset time period as the first azimuth angle data, and re-execute the steps of collecting the driving azimuth angle data during vehicle driving and subsequent steps until the classification result is normal driving again, and store the driving azimuth angle data within the preset time period as the second azimuth angle data;

[0168] The determining unit 606 is configured to determine a type of lane-changing driving behavior based on the first azimuth angle data and the second azimuth angle data.

[0169] Optionally, the device further includes:

[0170] A raw data acquisition unit, used to collect raw driving azimuth angle data when a vehicle changes driving behavior in different types of lanes;

[0171] a segmentation unit, configured to segment each piece of the original driving azimuth angle data according to a sliding window to obtain driving azimuth angle data to be trained, wherein the length of the sliding window is a preset time length;

[0172] A second calculation unit is used to calculate the data features of the driving azimuth data to be trained;

[0173] A training unit is used to train a vehicle driving behavior classifier using the data features of the driving azimuth angle data to be trained and the classification result label values ​​corresponding to the driving azimuth angle data to be trained, wherein the classification result label values ​​corresponding to the driving azimuth angle data to be trained include normal driving or lane change driving.

[0174] Optionally, the feature calculation unit is specifically used to calculate a first variance, a first range and a first mean absolute deviation value of the driving azimuth angle data to be trained as data features of the driving azimuth angle data to be trained.

[0175] Optionally, the vehicle driving behavior classifier is a naive Bayes classifier;

[0176] The input unit is specifically configured to calculate, using the naive Bayes classifier, a first posterior probability that the data feature of the driving azimuth angle data within the preset time period belongs to a normal driving classification result, and a second posterior probability that the data feature of the driving azimuth angle data within the preset time period belongs to a lane change driving classification result;

[0177] If the first a posteriori probability is greater than the second a posteriori probability, the classification result is determined to be normal driving; if the first a posteriori probability is less than the second a posteriori probability, the classification result is determined to be driving with a lane change.

[0178] Optionally, the first calculation unit is specifically used to calculate the second variance, second range and second mean absolute deviation value of the driving azimuth angle data within the preset time length as data features of the driving azimuth angle data within the preset time length when the collection time length reaches the preset time length.

[0179] Optionally, the determining unit includes:

[0180] A first calculation subunit, configured to calculate a mean value of first azimuth angle data corresponding to the first azimuth angle data;

[0181] A second calculation subunit, configured to calculate a mean of the second azimuth data corresponding to the second azimuth data;

[0182] a third calculation subunit, configured to calculate an absolute value of a difference between a mean value of the first azimuth angle data and a mean value of the second azimuth angle data to obtain a target value;

[0183] a first classification subunit, configured to determine that the lane change driving behavior is a lane-changing driving behavior if the target value is greater than or equal to 0 and less than or equal to a first threshold;

[0184] a second classification subunit, configured to determine that the lane change driving behavior is a U-turn driving behavior if the target value is greater than or equal to a second threshold;

[0185] The third classification subunit is configured to determine that the type of the lane changing driving behavior is a turning driving behavior if the target value is greater than or equal to a third threshold and less than or equal to a fourth threshold.

[0186] Optionally, the first calculation subunit is specifically configured to divide the first azimuth angle data into multiple parts, calculate a third deviation of each part of the first azimuth angle data, and calculate a mean of the first azimuth angle data with the smallest third deviation as the mean of the first azimuth angle data;

[0187] The second calculation subunit is specifically used to divide the second azimuth data into multiple parts, calculate the fourth variance of each part of the second azimuth data, and calculate the mean of the part of the second azimuth data with the smallest fourth variance as the mean of the second azimuth data.

[0188] In addition, an embodiment of the present application also provides a device for identifying vehicle driving behavior, including: a memory, a processor, and a computer program stored on the memory and runnable on the processor. When the processor executes the computer program, it implements the vehicle driving behavior identification method described in the above embodiment.

[0189] In addition, an embodiment of the present application also provides a computer-readable storage medium, which stores instructions. When the instructions are executed on a terminal device, the terminal device executes the vehicle driving behavior identification method described in the above embodiment.

[0190] Thus, by collecting the driving azimuth angle data of the vehicle and calculating the data features of the driving azimuth angle data within a preset time period, the vehicle driving behavior can be classified according to the data features. When the classification result is that the lane change occurs, the most recently stored driving azimuth angle data within the preset time period and the driving azimuth angle data of normal driving stored within the preset time period that are subsequently collected are used as the first azimuth angle data and the second azimuth angle data, respectively. The type of lane change driving behavior can be accurately obtained by using the first azimuth angle data before the lane change and the second azimuth angle data after the lane change. Moreover, by collecting, storing and analyzing the driving azimuth angle data in real time, the type of lane change driving behavior can be identified and determined in a timely manner based on the driving azimuth angle data.

[0191] It should be noted that the various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. Reference can be made to the common and similar parts between the various embodiments. For the systems or devices disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple, and the relevant parts can be referred to the method description.

[0192] It should be understood that in this application, "at least one (item)" means one or more, and "plurality" means two or more. "And / or" is used to describe the association relationship of associated objects, indicating that three relationships may exist. For example, "A and / or B" can mean: only A exists, only B exists, and A and B exist at the same time, where A and B can be singular or plural. The character " / " generally indicates that the previous and next associated objects are in an "or" relationship. "At least one of the following items" or similar expressions refers to any combination of these items, including any combination of single items or plural items. For example, at least one of a, b or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.

[0193] It should also be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of additional identical elements in the process, method, article, or device comprising the element.

[0194] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein may be implemented directly using hardware, a software module executed by a processor, or a combination of the two. The software module may be placed in a random access memory (RAM), internal memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.

[0195] The above description of the disclosed embodiments is intended to enable one skilled in the art to implement or use the present application. Various modifications to these embodiments will be readily apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application is not limited to the embodiments shown herein, but is intended to conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for identifying vehicle driving behavior, characterized in that: The method comprises: Collect driving azimuth angle data of the vehicle during driving; When the collection time reaches a preset time, calculating the second variance, the second range, and the second mean absolute deviation of the driving azimuth angle data within the preset time as data features of the driving azimuth angle data within the preset time; Using a naive Bayesian classifier, a first posterior probability is calculated that the data features of the driving azimuth angle data within the preset time period belong to a classification result of normal driving, and a second posterior probability is calculated that the data features of the driving azimuth angle data within the preset time period belong to a classification result of lane change driving; if the first posterior probability is greater than the second posterior probability, the classification result is determined to be normal driving; if the first posterior probability is less than the second posterior probability, the classification result is determined to be lane change driving; the naive Bayesian classifier is a vehicle driving behavior classifier; When the classification result is normal driving, storing the driving azimuth angle data within the preset time period, and re-performing the step of collecting the driving azimuth angle data during the vehicle driving process and subsequent steps; When the classification result is that the lane change occurs, the most recently stored driving azimuth angle data within the preset time period is determined as the first azimuth angle data, and the steps of collecting the driving azimuth angle data during the vehicle's driving process and subsequent steps are re-executed until the classification result is again that the vehicle is driving normally, and the driving azimuth angle data within the preset time period is stored as the second azimuth angle data; Calculating a mean of the first azimuth angle data corresponding to the first azimuth angle data; Calculating a mean value of the second azimuth angle data corresponding to the second azimuth angle data; Calculating the absolute value of the difference between the mean of the first azimuth angle data and the mean of the second azimuth angle data to obtain a target value; If the target value is greater than or equal to 0 and less than or equal to a first threshold, determining that the type of the lane change driving behavior is a lane change driving behavior; If the target value is greater than or equal to a second threshold, determining that the type of the lane change driving behavior is a U-turn driving behavior; If the target value is greater than or equal to the third threshold and less than or equal to the fourth threshold, it is determined that the type of the lane changing driving behavior is a turning driving behavior.

2. The method according to claim 1, characterized in that The method further comprises: Collecting original driving azimuth angle data when a vehicle changes driving behavior in different types of lanes; Each piece of the original driving azimuth angle data is segmented according to a sliding window to obtain driving azimuth angle data to be trained, where the length of the sliding window is a preset time length; Calculating data features of the driving azimuth data to be trained; A vehicle driving behavior classifier is trained using the data features of the driving azimuth angle data to be trained and the classification result label values ​​corresponding to the driving azimuth angle data to be trained. The classification result label values ​​corresponding to the driving azimuth angle data to be trained include normal driving or lane change driving.

3. The method according to claim 2, characterized in that The data features of the driving azimuth data to be trained are calculated, including: A first variance, a first range, and a first mean absolute deviation of the to-be-trained driving azimuth data are calculated as data features of the to-be-trained driving azimuth data.

4. The method according to claim 1, wherein The calculating the first azimuth angle data mean corresponding to the first azimuth angle data includes: Divide the first azimuth angle data into multiple parts, calculate the third deviation of each part of the first azimuth angle data, and calculate the mean of the first azimuth angle data with the smallest third deviation as the mean of the first azimuth angle data; The calculating the second azimuth data mean corresponding to the second azimuth data includes: The second azimuth angle data is divided into multiple parts, the fourth variance of each part of the second azimuth angle data is calculated, and the mean of the part of the second azimuth angle data with the smallest fourth variance is calculated as the mean of the second azimuth angle data.

5. A vehicle driving behavior recognition device, characterized in that: The device comprises: A collection unit, used for collecting driving azimuth angle data of the vehicle during driving; a first calculation unit, configured to calculate, when the collection time reaches a preset time, a second variance, a second range, and a second mean absolute deviation of the driving azimuth angle data within the preset time as data features of the driving azimuth angle data within the preset time; an input unit configured to calculate, using a naive Bayesian classifier, a first a posteriori probability that the data features of the driving azimuth angle data within the preset duration belong to a classification result of normal driving and a second a posteriori probability that the data features of the driving azimuth angle data within the preset duration belong to a classification result of lane change; if the first a posteriori probability is greater than the second a posteriori probability, the classification result is determined to be normal driving; and if the first a posteriori probability is less than the second a posteriori probability, the classification result is determined to be lane change driving; the naive Bayesian classifier is a vehicle driving behavior classifier; A first classification unit is configured to, when the classification result is normal driving, store the driving azimuth angle data within the preset time period, and re-execute the steps of collecting the driving azimuth angle data during the vehicle driving process and subsequent steps; a second classification unit configured to, when the classification result is lane change, determine the most recently stored driving azimuth angle data within a preset time period as first azimuth angle data, and re-execute the steps of collecting driving azimuth angle data during vehicle driving and subsequent steps until the classification result is normal driving again, and store the driving azimuth angle data within the preset time period as second azimuth angle data; A determination unit is used to calculate the mean of the first azimuth data corresponding to the first azimuth data; calculate the mean of the second azimuth data corresponding to the second azimuth data; calculate the absolute value of the difference between the mean of the first azimuth data and the mean of the second azimuth data to obtain a target value; if the target value is greater than or equal to 0 and less than or equal to a first threshold value, determine that the type of lane change driving behavior is lane changing driving behavior; if the target value is greater than or equal to the second threshold value, determine that the type of lane change driving behavior is U-turn 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, determine that the type of lane change driving behavior is turning driving behavior.

6. A vehicle driving behavior recognition device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the method for identifying vehicle driving behavior according to any one of claims 1 to 4 is implemented.

7. A computer-readable storage medium, characterized in that The computer-readable storage medium stores instructions, and when the instructions are executed on the terminal device, the terminal device executes the vehicle driving behavior recognition method according to any one of claims 1 to 4.

Images