CN108268678B - Driving behavior analysis method, device and system - Google Patents
Driving behavior analysis method, device and system Download PDFInfo
- Publication number
- CN108268678B CN108268678B CN201611262696.0A CN201611262696A CN108268678B CN 108268678 B CN108268678 B CN 108268678B CN 201611262696 A CN201611262696 A CN 201611262696A CN 108268678 B CN108268678 B CN 108268678B
- Authority
- CN
- China
- Prior art keywords
- driving behavior
- behavior analysis
- vehicle
- analysis model
- driving
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/30—Circuit design
- G06F30/36—Circuit design at the analogue level
- G06F30/367—Design verification, e.g. using simulation, simulation program with integrated circuit emphasis [SPICE], direct methods or relaxation methods
Landscapes
- Engineering & Computer Science (AREA)
- Computer Hardware Design (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Microelectronics & Electronic Packaging (AREA)
- Evolutionary Computation (AREA)
- Geometry (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Traffic Control Systems (AREA)
Abstract
A driving behavior analysis method, a device and a system are provided. The method comprises the following steps: the method comprises the steps of obtaining driving behavior data of a vehicle within a preset first time period and driving mileage of the vehicle within the first time period; processing the driving behavior data within the first duration to obtain the occurrence frequency of each driving behavior within a preset mileage; inputting the obtained times of the occurrence of each driving behavior within a preset mileage into a trained driving behavior analysis model, calculating and outputting the collision occurrence probability of the vehicle by using the driving behavior analysis model, wherein the input parameters of the driving behavior analysis model comprise the driving behavior, a first mapping relation exists between the driving behavior and the collision occurrence probability of the vehicle, and the corresponding weight of each driving behavior in the driving behavior analysis model in the first mapping relation is obtained through training. By adopting the scheme, the accuracy of driving behavior analysis and driving safety can be improved.
Description
Technical Field
The invention relates to the technical field of vehicle networking, in particular to a driving behavior analysis method, a driving behavior analysis device and a driving behavior analysis system.
Background
With the development of society, automobiles have become indispensable vehicles for people in daily traffic. Currently, driving behavior analysis of vehicles is used in a variety of situations. For example, the insurance company analyzes the driving behavior of the vehicle, and determines the vehicle premium based on the driving behavior analysis result. As another example, some businesses determine the driver's driving experience based on the driver's driving behavior when recruiting the driver.
However, the accuracy of the driving behavior analysis result obtained by the conventional driving behavior analysis method is low.
Disclosure of Invention
The invention solves the technical problem of how to improve the accuracy of driving behavior analysis and driving safety.
In order to solve the technical problem, an embodiment of the present invention provides a driving behavior analysis method, including: the method comprises the steps of obtaining driving behavior data of a vehicle within a preset first time period and driving mileage of the vehicle within the first time period; processing the driving behavior data within the first duration to obtain the occurrence frequency of each driving behavior within a preset mileage; inputting the obtained times of the occurrence of each driving behavior within a preset mileage into a trained driving behavior analysis model, calculating and outputting the collision occurrence probability of the vehicle by using the driving behavior analysis model, wherein the input parameters of the driving behavior analysis model comprise the driving behavior, a first mapping relation exists between the driving behavior and the collision occurrence probability of the vehicle, and the corresponding weight of each driving behavior in the driving behavior analysis model in the first mapping relation is obtained through training.
Optionally, after calculating the collision occurrence probability of the vehicle, the method further includes: according to the collision occurrence probability of the vehicle, the influence of each driving behavior on the collision occurrence probability of the vehicle is analyzed by combining the driving behavior data in the first duration and adopting the driving behavior analysis model; and generating and outputting driving behavior improvement suggestions aiming at the influence of each driving behavior on the vehicle collision occurrence probability.
Optionally, the driving behavior comprises: rapid acceleration, rapid deceleration, idling, fatigue driving and overspeed.
Optionally, the driving behavior data in the first duration are respectively processed according to a time interval, so that the frequency of each driving behavior in a preset mileage in the corresponding time interval is obtained; inputting the times of the driving behaviors within the preset mileage within the obtained corresponding time interval into a trained driving behavior analysis model; and calculating and outputting the collision occurrence probability of the vehicle by adopting the driving behavior analysis model, wherein the input parameters of the driving behavior analysis model have corresponding weights in different time intervals.
Optionally, the input parameters of the driving behavior analysis model further include collision event related factors, the collision occurrence probability of the vehicle and the driving behavior and collision event related factors have a second mapping relationship, and weights corresponding to the driving behavior and collision event related factors in the driving behavior analysis model in the second mapping relationship are obtained through training respectively.
Optionally, the method further comprises: acquiring collision event associated factor data corresponding to the vehicle within the first duration; processing the data of the factors related to the collision event in the first time length; and inputting the processed data of the factors related to the collision event into a trained driving behavior analysis model.
Optionally, the crash event related factors include at least one of: sharp turns, places with multiple accidents, road conditions, climate, driving age, personality, region and violation records.
Optionally, the driving behavior analysis model is trained in the following manner: acquiring a training sample, wherein the training sample comprises driving behavior data, collision event associated factor data and a collision degree value; marking the training samples with the collision degree value larger than or equal to a preset threshold value as positive samples, and marking the training samples with the collision degree value smaller than the threshold value as negative samples; and performing logistic regression training on the positive sample and the negative sample by adopting a logistic regression algorithm according to the driving behavior data and the collision event associated factor data to obtain the driving behavior analysis model.
Optionally, the driving behavior analysis model is trained in the following manner: acquiring a training sample, wherein the training sample comprises driving behavior data and a collision degree value; marking the training samples with the collision degree value larger than or equal to a threshold value as positive samples, and marking the training samples with the collision degree value smaller than the threshold value as negative samples; and performing logistic regression training on the positive sample and the negative sample according to the driving behavior data by adopting a logistic regression algorithm to obtain the driving behavior analysis model.
Optionally, when training the driving behavior analysis model, the method includes: randomly selecting a preset number of training samples from the training samples as test samples; and verifying the accuracy of the driving behavior analysis model by adopting the test sample verification.
Optionally, the performing accuracy verification on the driving behavior analysis model by using the test sample verification includes: randomly dividing the test sample into N groups of test sub-samples, wherein N is a natural number and is more than or equal to 3; and carrying out accuracy verification on a group of test subsamples except the N-1 group by adopting any N-1 groups of test subsamples until the N groups of test subsamples are verified.
Optionally, the calculating the collision occurrence probability of the vehicle includes: acquiring a threshold adjustment coefficient; acquiring weights respectively corresponding to the driving behaviors in the first mapping relation from the trained driving behavior analysis model according to the threshold adjustment coefficients, wherein the driving behaviors respectively correspond to different weights under different threshold adjustment coefficients in the driving behavior analysis model; and calculating the collision occurrence probability of the vehicle according to the acquired weights of the driving behaviors respectively corresponding to the driving behaviors in the first mapping relation.
Optionally, the method further comprises: and when a preset model updating triggering event is monitored, acquiring the driving behavior data of the vehicle as an updating sample, adopting the updating sample to update and train the driving behavior analysis model, and using the driving behavior analysis model obtained by updating and training as a trained driving behavior analysis model.
Optionally, the first duration is a time period taking the current time as an end time.
An embodiment of the present invention further provides a driving behavior analysis device, including: the device comprises an acquisition unit, a processing unit, an input unit, a driving behavior analysis model, a calculation unit and a first output unit, wherein: the acquisition unit is suitable for acquiring driving behavior data of a vehicle within a preset first time period and driving mileage of the vehicle within the first time period; the processing unit is suitable for processing the driving behavior data within the first duration to obtain the occurrence frequency of each driving behavior within a preset mileage; the input unit is suitable for inputting the obtained times of the driving behaviors within the preset mileage into the trained driving behavior analysis model; the input parameters of the driving behavior analysis model comprise driving behaviors, a first mapping relation exists between the driving behaviors and the collision occurrence probability of the vehicle, and the corresponding weight of each driving behavior in the driving behavior analysis model in the first mapping relation is obtained through training; the calculation unit is suitable for calculating the collision occurrence probability of the vehicle by adopting the driving behavior analysis model; the first output unit is adapted to output the collision occurrence probability of the vehicle calculated by the calculation unit.
Optionally, the apparatus further comprises: analysis unit and second output unit, wherein: the analysis unit is suitable for analyzing the influence of each driving behavior on the collision occurrence probability of the vehicle by adopting the driving behavior analysis model according to the collision occurrence probability of the vehicle and the data corresponding to all the driving behaviors in the first time period after calculating the collision occurrence probability of the vehicle, and generating a driving behavior improvement suggestion aiming at the influence of each driving behavior on the collision occurrence probability of the vehicle; the second output unit is adapted to output the driving behavior improvement advice generated by the analysis unit.
Optionally, the processing unit is adapted to process the driving behavior data within the first duration according to a time interval, so as to obtain the number of times that each driving behavior within a corresponding time interval occurs within a preset mileage; the input unit is suitable for inputting the frequency of the driving behaviors within the preset mileage within the obtained corresponding time interval into the trained driving behavior analysis model; the calculating unit is suitable for calculating the collision occurrence probability of the vehicle by adopting the driving behavior analysis model, and the input parameters of the driving behavior analysis model have corresponding weights in different time intervals.
Optionally, the input parameters of the driving behavior analysis model further include collision event related factors, the collision occurrence probability of the vehicle and the driving behavior and collision event related factors have a second mapping relationship, and weights corresponding to the driving behavior and collision event related factors in the driving behavior analysis model in the second mapping relationship are obtained through training respectively.
Optionally, the obtaining unit is further adapted to obtain a collision event related factor corresponding to the vehicle within the first duration; the processing unit is further suitable for processing the data of the factors related to the collision event in the first time length; the input unit is further adapted to input the processed collision event related factor data to the trained driving behavior analysis model.
Optionally, the apparatus further comprises: a first training unit adapted to train the driving behavior analysis model in the following way: acquiring a training sample, wherein the training sample comprises driving behavior data, collision event associated factor data and a collision degree value; marking the training samples with the collision degree value larger than or equal to a preset threshold value as positive samples, and marking the training samples with the collision degree value smaller than the threshold value as negative samples; and performing logistic regression training on the positive sample and the negative sample by adopting a logistic regression algorithm according to the driving behavior data and the collision event associated factor data to obtain the driving behavior analysis model.
Optionally, the apparatus further comprises: a second training unit adapted to train the driving behavior analysis model in the following manner: acquiring a training sample, wherein the training sample comprises driving behavior data and a collision degree value; marking the training samples with the collision degree value larger than or equal to a threshold value as positive samples, and marking the training samples with the collision degree value smaller than the threshold value as negative samples; and performing logistic regression training on the positive sample and the negative sample according to the driving behavior data by adopting a logistic regression algorithm to obtain the driving behavior analysis model.
Optionally, the apparatus further comprises a testing unit adapted to randomly select a preset number of training samples from the training samples as testing samples; and verifying the accuracy of the driving behavior analysis model by adopting the test sample verification.
Optionally, the test unit is adapted to randomly divide the test sample into N groups of test subsamples, where N is a natural number and N is greater than or equal to 3; and carrying out accuracy verification on a group of test subsamples except the N-1 group by adopting any N-1 groups of test subsamples until the N groups of test subsamples are verified.
Optionally, the calculating unit is adapted to obtain a threshold adjustment coefficient; acquiring weights corresponding to the driving behaviors from the trained driving behavior analysis model according to the threshold adjustment coefficients, wherein the driving behaviors correspond to different weights in the first mapping relation under different threshold adjustment coefficients in the driving behavior analysis model; and calculating the collision occurrence probability of the vehicle according to the acquired weights of the driving behaviors respectively corresponding to the driving behaviors in the first mapping relation.
Optionally, the apparatus further comprises: and the updating unit is suitable for acquiring the driving behavior data of the vehicle as an updating sample when a preset model updating triggering event is monitored, updating and training the driving behavior analysis model by adopting the updating sample, and taking the driving behavior analysis model obtained by updating and training as a trained driving behavior analysis model.
An embodiment of the present invention further provides a driving behavior analysis system, including: data acquisition device and above arbitrary driving behavior analytical equipment, wherein: the data acquisition device is suitable for acquiring driving behavior data of a vehicle within a preset first time period and driving mileage of the vehicle within the first time period.
Optionally, the data acquisition device is adapted to acquire driving behavior data of the vehicle within a preset first time period through a driving state sensor mounted on the vehicle, or acquire the driving behavior data of the vehicle within the preset first time period from a vehicle memory.
Compared with the prior art, the technical scheme of the embodiment of the invention has the following beneficial effects:
and calculating the collision occurrence probability of the vehicle according to the corresponding weight of each driving behavior in the driving behavior analysis model in the first mapping relation by using the trained driving behavior analysis model and taking the occurrence frequency of each driving behavior in a preset mileage as an input parameter. And averaging the occurrence times of each driving behavior in the first duration to a preset mileage, so that the influence of the single driving behavior on the calculated collision probability can be reduced, the accuracy of driving behavior analysis can be improved, and the driving safety can be improved.
Furthermore, by means of the collision occurrence probability of the vehicle and the combination of the driving behavior data in the first duration, the influence of each driving behavior on the collision occurrence probability of the vehicle is analyzed to obtain bad driving behaviors, and a driving behavior improvement suggestion is given, so that the method can help to remind and improve the bad driving behaviors of the user, reduce the collision occurrence probability of the vehicle and improve driving safety.
Furthermore, the driving behavior data in the first duration are respectively processed according to the time intervals to obtain the times of the driving behaviors in the corresponding time intervals within the preset mileage, so that different impact degrees of the driving behaviors in different time intervals on the collision can be considered, and the accuracy of driving behavior analysis is further improved.
Furthermore, the input parameters in the driving behavior analysis model may further include collision event related factor data corresponding to the vehicle within the first duration, and when the collision probability of the vehicle is calculated, the accuracy of the driving behavior analysis may be further improved by integrating the collision event related factors.
Furthermore, the weights corresponding to the driving behaviors in the driving behavior analysis model are selected by selecting a threshold adjustment coefficient, so that the requirements of different users on the precision in different application scenes can be met.
Furthermore, when a preset model updating triggering event is monitored, the driving behavior data of the vehicle are obtained as an updating sample, the driving behavior analysis model is updated and trained, the updated driving behavior analysis model is used as a trained driving behavior analysis model, and the driving behavior analysis model is updated regularly, so that the matching degree of the driving behavior analysis model and the vehicle is higher, and the accuracy of driving behavior analysis can be further improved.
Drawings
FIG. 1 is a flow chart of a driving behavior analysis method in an embodiment of the invention;
FIG. 2 is a schematic structural diagram of a driving behavior analysis apparatus according to an embodiment of the present invention;
fig. 3 is a schematic structural diagram of another driving behavior analysis device according to an embodiment of the present invention.
Detailed Description
The existing driving behavior analysis method usually collects vehicle data in real time and analyzes the driving behavior of a user according to the vehicle data collected in real time, and because the influence factors of the driving behavior of the user are different in different scenes, the obtained driving behavior cannot reflect the actual driving situation of the user, and the accuracy is low.
In order to solve the above problem, in the embodiment of the present invention, a trained driving behavior analysis model is adopted, the number of times that each driving behavior occurs within a preset mileage is taken as an input parameter, and the collision occurrence probability of the vehicle is calculated according to the corresponding weight of each driving behavior in the driving behavior analysis model in the first mapping relationship. And averaging the occurrence times of each driving behavior in the first duration to a preset mileage, so that the influence of the single driving behavior on the calculated collision probability can be reduced, the accuracy of driving behavior analysis can be improved, and the driving safety can be improved.
In order to make the aforementioned objects, features and advantages of the present invention comprehensible, embodiments accompanied with figures are described in detail below.
Referring to fig. 1, a flowchart of a driving behavior analysis method according to an embodiment of the present invention is shown, and details are described below with reference to specific steps.
In specific implementation, the driving behavior data and the driving mileage of the vehicle in the first time period can be acquired through an On-Board Diagnostic (OBD) system of the vehicle, or the driving behavior data and the driving mileage of the vehicle in the first time period can be acquired through communication between an On-Board TBOX and a corresponding controller in the vehicle, or the driving behavior data can be acquired in real time through various sensors mounted On the vehicle and stored. It is understood that the driving behavior data and the driving mileage of the vehicle in the first time period can be obtained through other intelligent hardware installed on the vehicle.
In an embodiment of the present invention, the driving behavior may include: rapid acceleration, rapid deceleration, idling, fatigue driving, overspeed, and the like. In another embodiment of the present invention, the driving behavior comprises: neutral sliding, over long engine preheating time, no flameout after long-time parking, over-high running speed in the running-in period or brake stepping on by keeping the D gear during parking, and the like. In a specific implementation, the influence of various driving behaviors on the collision occurrence probability of the vehicle can be counted according to the big data, and the driving behavior with higher relevance is used as an input parameter of a driving behavior analysis model adopted in the driving behavior analysis process, and is not limited to the above listed driving behaviors.
In specific implementation, the first duration may be a natural week, a natural month, a natural quarter, or a specific duration may be set as needed. For example, 45 days, 52 days, etc., and specific values of the first time period are not limited in the embodiment of the present invention.
In a specific implementation, the starting point and the ending point of the first time length may also be set according to actual needs. When the driving behavior within a certain time period needs to be checked, the starting point and the ending point of the first time length can be set according to the time period to be checked. In an embodiment of the present invention, the first duration is a time period taking a current time as an end time. For example, the first duration is one month, and the current time is 2016, 12, 22, 11, 25, 36, then the first duration is 11, 25, 00, 23, 2016, 12, 22, 11, 24, 59.
And 12, processing the driving behavior data within the first duration to obtain the frequency of each driving behavior within a preset mileage.
In specific implementation, the acquired driving behavior data in the first time period may be processed, the total number of times of occurrence of each driving behavior in the first time period is calculated, and then the number of times of occurrence of each driving behavior in a preset mileage is calculated according to the driving mileage in the first time period.
For example, a sudden acceleration occurs m times, a sudden deceleration n times, in one within a natural monthThe driving distance in a natural month is P kilometers, and then the rapid acceleration occurs in a preset distance Q kilometers
Secondly, a sudden deceleration occurs
Wherein m, n, P and Q are positive integers, and P is more than or equal to Q.
And step 13, inputting the times of the driving behaviors within the preset mileage into a trained driving behavior analysis model, and calculating and outputting the collision occurrence probability of the vehicle by using the driving behavior analysis model.
In a specific implementation, the input parameters of the driving behavior analysis model may include driving behavior, and the driving behavior has a first mapping relation with the collision occurrence probability of the vehicle. The corresponding weight of each driving behavior in the driving behavior analysis model in the first mapping relation can be obtained through training, and the weight is used for representing the correlation degree between each driving behavior and the collision occurrence probability of the vehicle.
And inputting the obtained times of the driving behaviors within the preset mileage as input parameters into a trained driving behavior analysis model, and calculating the collision occurrence probability of the vehicle according to the corresponding weight of the driving behaviors in the driving behavior analysis model in the first mapping relation. And when the collision occurrence probability of the vehicle is obtained through calculation, outputting a calculation result.
According to the above, the trained driving behavior analysis model is adopted, the frequency of occurrence of each driving behavior in the preset mileage is taken as an input parameter, and the collision occurrence probability of the vehicle is calculated according to the corresponding weight of each driving behavior in the driving behavior analysis model in the first mapping relation. And averaging the occurrence times of each driving behavior in the first duration to a preset mileage, so that the influence of the single driving behavior on the calculated collision probability can be reduced, the accuracy of driving behavior analysis can be improved, and the driving safety can be improved.
In particular, in order to more intuitively display the analysis result of the driving behavior to the user, in an embodiment of the present invention, a correspondence relationship between the probability of collision of the vehicle and the driving behavior evaluation value may be set. After the collision occurrence probability of the vehicle is calculated, the collision occurrence probability of the vehicle may be converted into a corresponding evaluation value according to a preset correspondence relationship.
In one embodiment of the present invention, the driving behavior evaluation value may be represented by a value of 0 to 100, and the driving behavior evaluation value is inversely related to the collision occurrence probability of the vehicle. The larger the driving behavior evaluation value is, the better the driving behavior is, and the smaller the collision occurrence probability of the vehicle is, whereas the smaller the driving behavior evaluation value is, the more the bad driving behavior is, and the larger the collision occurrence probability of the vehicle is.
In a specific implementation, in order to improve the bad driving behaviors of the user, reduce the collision occurrence probability of the vehicle, and improve driving safety, in an embodiment of the present invention, after the collision occurrence probability of the vehicle is obtained, the influence of each driving behavior on the collision occurrence probability of the vehicle may be analyzed by using the driving behavior analysis model in combination with the driving behavior data in the first time period according to the collision occurrence probability of the vehicle, and a driving behavior improvement suggestion is generated and output for the influence of each driving behavior on the collision occurrence probability of the vehicle.
For example, in the driving behavior data in the first period of time, it is shown that the number of times of rapid deceleration is 20, and in the collision occurrence probability of the vehicle, the impact of the rapid deceleration on the probability of collision occurrence is large, so that the driving behavior improvement advice can be given: the number of rapid decelerations is reduced as much as possible. In specific implementation, the driving behavior improvement suggestion can be output through an associated vehicle-mounted entertainment device, can be output through a mobile terminal such as a mobile phone, can be output through other terminal equipment, or can be output in an audio mode through an audio device.
In practical applications, the impact of the same driving behavior on the collision occurrence probability of the vehicle is different in different time periods. For example, fatigue driving has a greater influence on the collision occurrence probability of a vehicle at night than fatigue driving in the daytime. For another example, the collision probability of the vehicle by the rapid acceleration occurring at the time of the peak of work is larger than that by the rapid acceleration occurring at night. Therefore, if the weights corresponding to the same driving behavior are the same at any time when calculating the collision occurrence probability of the vehicle, the accuracy of the calculated collision occurrence probability of the vehicle, that is, the accuracy of the analysis of the driving behavior of the vehicle, may be affected.
In order to improve the accuracy of the resulting probability of a collision of the vehicle and to improve the accuracy of the analysis of the driving behaviour of the vehicle. In an embodiment of the invention, the driving behavior data in the first duration are respectively processed according to time intervals to obtain the times of occurrence of each driving behavior in a preset mileage in the corresponding time interval, and the obtained times of occurrence of each driving behavior in the preset mileage in the corresponding time interval are input into a trained driving behavior analysis model; and calculating and outputting the collision occurrence probability of the vehicle by adopting the driving behavior analysis model. The input parameters of the driving behavior analysis model have corresponding weights in different time intervals.
In specific implementation, the set duration may be divided according to a preset time division manner to obtain different time intervals, and each driving behavior is subdivided into the number of times of occurrence in each time interval. And obtaining the frequency of each driving behavior occurring within the preset mileage in the corresponding time interval according to the frequency of each driving behavior occurring within each time interval and the driving mileage of the vehicle within the first time length. In the process of training the driving behavior analysis model, corresponding weights of the same driving behavior in different time intervals may be different.
For example, each natural day is divided in the time dimension into morning, afternoon and evening. Wherein the time interval corresponding to the morning is 05:00: 00-11: 59:59, the time interval corresponding to the afternoon is 12:00: 00-18: 59:59, and the time interval corresponding to the evening is 17:00: 00-24: 00:00 and 00:00: 00-04: 59. The weight of fatigue driving in the morning, the weight of fatigue driving in the afternoon and the weight of fatigue driving in the evening can be different, and the weight of fatigue driving in the evening is respectively greater than the weight of fatigue driving in the morning and the weight of fatigue driving in the afternoon. The driving behavior data comprises time data of the occurrence of each driving behavior, and the frequency of the occurrence of each driving behavior in the corresponding time interval can be obtained according to the time data of the occurrence of each driving behavior.
In the specific implementation, because the environment during the running of the vehicle is relatively complex, the collision occurrence probability of the vehicle is not only influenced by the driving behavior, but also influenced by other various factors. In order to enable the analysis of the driving behavior to be more accurate, the obtained collision probability of the vehicle is more accurate. In an embodiment of the present invention, the input parameters of the driving behavior analysis model may further include a collision event related factor, the collision occurrence probability of the vehicle and the driving behavior and collision event related factor have a second mapping relationship, and weights corresponding to the driving behavior and collision event related factors in the driving behavior analysis model in the second mapping relationship are obtained through training.
In specific implementation, the data of the relevant factors of the collision event corresponding to the vehicle in the first duration may be acquired, and the data of the relevant factors of the collision event in the first duration may be processed; and inputting the processed data of the factors related to the collision event into a trained driving behavior analysis model.
In particular implementations, crash event related factor data may be obtained for the vehicle over the first duration. The relationship between each collision event related factor and the collision probability of the vehicle can be determined through big data statistical analysis, the relationship between each collision event related factor and the collision probability of the vehicle can be determined through a correlation analysis method, the factors which may influence the collision probability of the vehicle can be determined, and the influence degree of each collision time related factor on the collision probability of the vehicle can be determined.
In particular implementations, the crash event related factors may include one or more of sharp turns, multiple locations of accidents, road conditions, climate, age of driver, personality, area, and violation records. For example, whether the road conditions are congested and smooth, whether a driver is a new driver with certain driving experience, whether the driver is a northern person or a southern person, is influenced by the character, and the southern person is more cautious to drive than the northern person, and records of violations such as red light running and converse running are carried out.
In specific implementation, a large number of training samples can be used to train the driving behavior analysis model, so as to obtain the weights corresponding to the driving behaviors and the collision event related factors in the second mapping relationship.
In particular implementations, the driving behavior analysis model may be trained in a variety of ways.
In an embodiment of the present invention, a training sample is obtained, where the training sample includes driving behavior data, collision event associated factor data, and a degree of collision value. Marking the training samples with the collision degree value larger than or equal to a preset threshold value as positive samples, and marking the training samples with the collision degree value smaller than the threshold value as negative samples. And performing logistic regression training on the positive sample and the negative sample by adopting a logistic regression algorithm according to the driving behavior data and the collision event associated factor data to obtain the driving behavior analysis model.
In a specific implementation, the value of the degree of collision may be represented in a numerical form of 0 to 100, or may be represented in a numerical form of 0 to 1, and it is understood that the value of the degree of collision may also have other representation forms, and may be specifically set according to needs, and is not limited herein. The preset threshold value may be set according to the expression form of the degree of collision value. For example, when the degree of collision is represented by a numerical value of 0 to 100, the predetermined threshold may be 26. It can be understood that the specific value of the preset threshold may be set according to the actual application scenario and the need, and may also be other values such as 30 and 40, and the specific value is not limited herein.
In another embodiment of the present invention, training samples are obtained, the training samples include driving behavior data and a degree of collision value, the training samples with the degree of collision value greater than or equal to a threshold are labeled as positive samples, and the training samples with the degree of collision value less than the threshold are labeled as negative samples. And performing logistic regression training on the positive sample and the negative sample according to the driving behavior data by adopting a logistic regression algorithm to obtain the driving behavior analysis model.
In an embodiment of the present invention, the driving behavior analysis model is obtained by training using the following logistic regression formula (1).
Wherein g (x) represents a collision occurrence probability of the vehicle; x is the number ofiIs an input parameter; thetaiFor inputting a parameter xiA corresponding weight; theta0Is a constant; i is more than or equal to 1 and less than or equal to n, and both i and n are positive integers.
In a specific implementation, in order to improve the accuracy of the logistic regression type and improve the accuracy of the driving behavior analysis during the training process of the driving behavior analysis model, in an embodiment of the present invention, in the training process of the driving behavior analysis model, a preset number of training samples may be randomly selected from the training samples as test samples, and the accuracy of the driving behavior analysis model is verified by using the test samples.
In a specific implementation, a certain proportion of training samples can be selected from the training samples as test samples. For example, 20% of the training samples are selected as test samples.
In a specific implementation, the driving behavior analysis model may be subjected to accuracy verification in the following manner. And randomly dividing the test sample into N groups of test sub-samples, wherein N is a natural number and is more than or equal to 3. And carrying out accuracy verification on a group of test subsamples except the N-1 group by adopting any N-1 groups of test subsamples until the N groups of test subsamples are verified.
In an embodiment of the present invention, the test samples are randomly divided into three groups of test subsamples, and the accuracy verification is performed on the other group of test subsamples by using two groups of test subsamples until all three groups of test subsamples are verified.
For example, the test samples are divided into A, B and C groups, the C group is verified by A, B two groups, and whether the estimated collision probability of the test samples in the C group by the driving behavior analysis model is compared with the actual collision probability of the C group, so as to meet the accuracy requirement. Similarly, group B was verified using A, C pairs and group A was verified using B, C pairs.
In a specific implementation, accuracy requirements of different application scenarios and different users on driving behavior analysis are different, and in order to meet the accuracy requirements of multiple scenarios, in an embodiment of the present invention, when the collision occurrence probability of the vehicle is calculated, a threshold adjustment coefficient may be obtained, and weights corresponding to driving behaviors in the first mapping relationship are obtained from the trained driving behavior analysis model according to the threshold adjustment coefficient, where the driving behaviors correspond to different weights respectively under different threshold adjustment coefficients in the driving behavior analysis model. And calculating the collision occurrence probability of the vehicle according to the acquired weights of the driving behaviors respectively corresponding to the driving behaviors in the first mapping relation.
In a specific implementation, the corresponding weights of the driving behaviors in the first mapping relation are different under different threshold adjustment coefficients. The corresponding weights of the same driving behavior in different threshold adjustment coefficients can be the same or different. The collision occurrence probability calculated by adopting different threshold adjustment coefficients has different accuracies.
For example, when the threshold adjustment coefficients are Z1, Z2, and Z3, and Z1 ≠ Z2 ≠ Z3, and the threshold adjustment coefficient is Z1, the weight corresponding to rapid acceleration is a1, and the weight corresponding to rapid deceleration is b 1; when the threshold adjustment coefficient is Z2, the weight corresponding to rapid acceleration is a2, and the weight corresponding to rapid deceleration is b 2; when the threshold adjustment coefficient is Z3, the weight corresponding to rapid acceleration is a3, and the weight corresponding to rapid deceleration is b 3. In one embodiment, a1, a2 and a3 may be different, and b1, b2 and b3 may be different. Or any two of a1, a2 and a3 may be the same, and b1, b2 and b3 may be different. Or any two of b1, b2 and b3 may be the same, and a1, a2 and a3 may be different. Alternatively, three of a1, a2 and a3 may be the same, and b1, b2 and b3 may be different.
In practical application, other situations can exist, and the weights corresponding to the driving behaviors are not completely the same only under different threshold adjustment coefficients.
In specific implementation, when a preset model updating triggering event is monitored, the driving behavior data of the vehicle are obtained as an updating sample, the driving behavior analysis model is updated and trained by adopting the updating sample, and the driving behavior analysis model obtained through updating and training is used as a trained driving behavior analysis model. By updating the driving behavior analysis model at regular time, the matching degree of the driving behavior analysis model and the vehicle can be higher, and the accuracy of the driving behavior analysis can be further improved.
In a specific implementation, the preset model updating triggering event may be a preset second time duration, and the driving behavior analysis model is automatically updated whenever the second time duration is reached. The second time period may be two months, one quarter, or one year. The specific value of the second duration may be set according to the value of the first duration or according to an actual application scenario.
The preset model updating triggering condition may also be a change of the area where the vehicle is located, for example, the user lives in the south before and then settles to the northeast, and the driving behavior analysis model may be updated according to the detected main activity area of the vehicle due to a large difference between the environment in the northeast and the south. The preset model updating triggering condition can also be that the user autonomously updates according to the self condition.
In order to enable the technical personnel in the field to understand and realize the invention better, the embodiment of the invention also provides a driving behavior analysis device.
Referring to fig. 2, a schematic structural diagram of a driving behavior analysis apparatus according to an embodiment of the present invention is shown. The driving behavior analysis means 20 may include: an acquisition unit 21, a processing unit 22, an input unit 23, a driving behavior analysis model 24, a calculation unit 25, and a first output unit 26, wherein:
the obtaining unit 21 is adapted to obtain driving behavior data within a preset first time period of a vehicle and driving mileage of the vehicle within the first time period;
the processing unit 22 is adapted to process the driving behavior data within the first duration to obtain the occurrence frequency of each driving behavior within a preset mileage;
the input unit 23 is adapted to input the obtained number of times that each driving behavior occurs within a preset mileage to the trained driving behavior analysis model;
the input parameters of the driving behavior analysis model 24 include driving behaviors, a first mapping relation exists between the driving behaviors and the collision occurrence probability of the vehicle, and the corresponding weight of each driving behavior in the driving behavior analysis model 24 in the first mapping relation is obtained through training;
the calculation unit 25 is adapted to calculate a collision occurrence probability of the vehicle using the driving behavior analysis model 24;
the first output unit 26 is adapted to output the collision occurrence probability of the vehicle calculated by the calculation unit 25.
In particular implementations, the driving behavior may include: rapid acceleration, rapid deceleration, idling, fatigue driving and overspeed.
As can be seen from the above, the trained driving behavior analysis model is adopted, the number of times that each driving behavior occurs within the preset mileage is taken as an input parameter, and the collision occurrence probability of the vehicle is calculated according to the corresponding weight of each driving behavior in the driving behavior analysis model in the first mapping relationship. And averaging the occurrence times of each driving behavior in the first duration to a preset mileage, so that the influence of the single driving behavior on the calculated collision probability can be reduced, the accuracy of driving behavior analysis can be improved, and the driving safety can be improved.
Referring to fig. 3, a schematic structural diagram of another driving behavior analysis apparatus according to an embodiment of the present invention is shown. In a specific implementation, the driving behavior analysis device 20 may further include, on the basis of fig. 2: an analysis unit 31 and a second output unit 32, wherein:
the analysis unit 31 is adapted to, after calculating the collision occurrence probability of the vehicle, analyze, by using the driving behavior analysis model 24, the influence of each driving behavior on the collision occurrence probability of the vehicle in combination with the data corresponding to each of all the driving behaviors in the first time period, and generate a driving behavior improvement suggestion for the influence of each driving behavior on the collision occurrence probability of the vehicle;
the second output unit 32 is adapted to output the driving behavior improvement advice generated by the analysis unit.
In a specific implementation, the processing unit 22 is adapted to respectively process the driving behavior data within the first duration according to a time interval, so as to obtain the number of times that each driving behavior within a corresponding time interval occurs within a preset mileage;
the input unit 23 is adapted to input the obtained times of occurrence of each driving behavior within the preset mileage in the corresponding time interval to the trained driving behavior analysis model 24;
the calculation unit 25 is adapted to calculate the collision occurrence probability of the vehicle using the driving behavior analysis model 24, and the input parameters of the driving behavior analysis model 24 have corresponding weights at different time intervals.
In a specific implementation, the input parameters of the driving behavior analysis model 24 may include collision event related factors, the collision occurrence probability of the vehicle and the driving behavior and collision event related factors have a second mapping relationship, and weights corresponding to the driving behavior and collision event related factors in the driving behavior analysis model 24 in the second mapping relationship are obtained through training.
In a specific implementation, the obtaining unit 21 is further adapted to obtain a collision event related factor corresponding to the vehicle within the first time period;
the input unit 23 is further adapted to input the processed collision event related factor data to a trained driving behavior analysis model 24, and the input parameters of the driving behavior analysis model comprise collision event related factors; the processing unit 22 is further adapted to process the data of the factors related to the collision event within the first time length; the input unit 23 is further adapted to input the processed collision event related factor data to the trained driving behavior analysis model 24.
In particular implementations, the crash event related factors may include one or more of sharp turns, multiple locations of accidents, road conditions, climate, age of driver, personality, region, and violation records.
In an implementation, the driving behavior analysis device 20 may further include a first training unit 33. The first training unit 33 is adapted to train the driving behavior analysis model 24 in the following way: acquiring a training sample, wherein the training sample comprises driving behavior data, collision event associated factor data and a collision degree value; marking the training samples with the collision degree value larger than or equal to a preset threshold value as positive samples, and marking the training samples with the collision degree value smaller than the threshold value as negative samples; and performing logistic regression training on the positive sample and the negative sample by adopting a logistic regression algorithm according to the driving behavior data and the collision event associated factor data to obtain the driving behavior analysis model 24.
In an embodiment of the present invention, the driving behavior analysis model 24 may be obtained by training using the logistic regression formula (1) provided in the above embodiment of the present invention.
In a specific implementation, the driving behavior analysis device 20 may further include a second training unit 34. The second training unit 34 is adapted to train the driving behavior analysis model 24 in the following way: acquiring a training sample, wherein the training sample comprises driving behavior data and a collision degree value; marking the training samples with the collision degree value larger than or equal to a threshold value as positive samples, and marking the training samples with the collision degree value smaller than the threshold value as negative samples; and performing logistic regression training on the positive sample and the negative sample according to the driving behavior data by adopting the logistic regression algorithm to obtain the driving behavior analysis model 24.
In a specific implementation, the driving behavior analysis device 20 may further include a testing unit 35, adapted to randomly select a preset number of training samples from the training samples as testing samples; the accuracy of the driving behavior analysis model 24 is verified using the test sample verification.
In specific implementation, the test unit 35 is adapted to randomly divide the test sample into N groups of test subsamples, where N is a natural number and N is greater than or equal to 3; and carrying out accuracy verification on a group of test subsamples except the N-1 group by adopting any N-1 groups of test subsamples until the N groups of test subsamples are verified.
In a specific implementation, the calculating unit 25 is adapted to obtain a threshold adjustment coefficient; acquiring weights corresponding to the driving behaviors from the trained driving behavior analysis model 24 according to the threshold adjustment coefficients, wherein the driving behaviors correspond to different weights in the first mapping relation under different threshold adjustment coefficients in the driving behavior analysis model 24; and calculating the collision occurrence probability of the vehicle according to the acquired weights of the driving behaviors respectively corresponding to the driving behaviors in the first mapping relation.
In a specific implementation, the driving behavior analysis device 20 may further include an updating unit 36, adapted to, when a preset model update triggering event is monitored, obtain driving behavior data of the vehicle as an update sample, perform update training on the driving behavior analysis model 24 by using the update sample, and use the driving behavior analysis model obtained through the update training as the trained driving behavior analysis model 24.
In a specific implementation, the driving behavior analysis device 20 may be an independent hardware device, may also be a server, and may also be application software or a client installed on a corresponding device. It is understood that the driving behavior analysis device 20 may also exist in other forms.
In a specific implementation, the specific working principle and the working process of the driving behavior analysis device 20 may refer to the description of the driving behavior analysis method provided in the above embodiment of the present invention, and are not described herein again.
An embodiment of the present invention further provides a driving behavior analysis system, where the driving behavior analysis system may include: a data acquisition device and any one of the driving behavior analysis devices as described in the above embodiments of the present invention.
In a specific implementation, the data acquisition device can acquire driving behavior data of a vehicle within a preset first time period and driving mileage of the vehicle within the first time period.
In a specific implementation, the data acquisition device may acquire driving behavior data of the vehicle within a preset first time period through a driving state sensor mounted on the vehicle, or may acquire the driving behavior data of the vehicle within the preset first time period from a vehicle memory.
Those skilled in the art will appreciate that all or part of the steps in the methods of the above embodiments may be implemented by associated hardware instructed by a program, which may be stored in a computer-readable storage medium, and the storage medium may include: ROM, RAM, magnetic or optical disks, and the like.
Although the present invention is disclosed above, the present invention is not limited thereto. Various changes and modifications may be effected therein by one skilled in the art without departing from the spirit and scope of the invention as defined in the appended claims.
Claims (25)
1. A driving behavior analysis method, characterized by comprising:
the method comprises the steps of obtaining driving behavior data of a vehicle within a preset first time period and driving mileage of the vehicle within the first time period;
processing the driving behavior data within the first duration to obtain the occurrence frequency of each driving behavior within a preset mileage;
inputting the obtained times of occurrence of each driving behavior within a preset mileage into a trained driving behavior analysis model, calculating and outputting the collision occurrence probability of the vehicle by using the driving behavior analysis model, wherein the input parameters of the driving behavior analysis model comprise the driving behavior, a first mapping relation exists between the driving behavior and the collision occurrence probability of the vehicle, and the corresponding weight of each driving behavior in the driving behavior analysis model in the first mapping relation is obtained through training;
wherein the driving behavior analysis model is trained in the following manner:
acquiring a training sample, wherein the training sample comprises driving behavior data and a collision degree value;
marking the training samples with the collision degree value larger than or equal to a threshold value as positive samples, and marking the training samples with the collision degree value smaller than the threshold value as negative samples;
and performing logistic regression training on the positive sample and the negative sample according to the driving behavior data by adopting a logistic regression algorithm to obtain the driving behavior analysis model.
2. The driving behavior analysis method according to claim 1, further comprising, after calculating the collision occurrence probability of the vehicle:
according to the collision occurrence probability of the vehicle, the influence of each driving behavior on the collision occurrence probability of the vehicle is analyzed by combining the driving behavior data in the first duration and adopting the driving behavior analysis model;
and generating and outputting driving behavior improvement suggestions aiming at the influence of each driving behavior on the vehicle collision occurrence probability.
3. The driving behavior analysis method according to claim 1, characterized in that the driving behavior includes: rapid acceleration, rapid deceleration, idling, fatigue driving and overspeed.
4. The driving behavior analysis method according to claim 3, wherein the driving behavior data within the first duration are processed according to time intervals, respectively, to obtain the number of times that each driving behavior within the corresponding time interval occurs within a preset mileage;
inputting the times of the driving behaviors within the preset mileage within the obtained corresponding time interval into a trained driving behavior analysis model;
and calculating and outputting the collision occurrence probability of the vehicle by adopting the driving behavior analysis model, wherein the input parameters of the driving behavior analysis model have corresponding weights in different time intervals.
5. The driving behavior analysis method according to claim 1, wherein the input parameters of the driving behavior analysis model further include a collision event related factor, the collision occurrence probability of the vehicle and the driving behavior and collision event related factor have a second mapping relationship, and weights corresponding to the driving behavior and collision event related factors in the driving behavior analysis model in the second mapping relationship are obtained through training respectively.
6. The driving behavior analysis method according to claim 5, characterized by further comprising:
acquiring collision event associated factor data corresponding to the vehicle within the first duration;
processing the data of the factors related to the collision event in the first time length;
and inputting the processed data of the factors related to the collision event into a trained driving behavior analysis model.
7. The driving behavior analysis method according to claim 5, characterized in that the collision event related factor includes at least one of:
sharp turns, places with multiple accidents, road conditions, climate, driving age, personality, region and violation records.
8. The driving behavior analysis method according to claim 5, characterized in that the driving behavior analysis model is trained in the following manner:
acquiring a training sample, wherein the training sample comprises driving behavior data, collision event associated factor data and a collision degree value;
marking the training samples with the collision degree value larger than or equal to a preset threshold value as positive samples, and marking the training samples with the collision degree value smaller than the threshold value as negative samples;
and performing logistic regression training on the positive sample and the negative sample by adopting a logistic regression algorithm according to the driving behavior data and the collision event associated factor data to obtain the driving behavior analysis model.
9. The driving behavior analysis method according to claim 1 or 8, comprising, when training the driving behavior analysis model:
randomly selecting a preset number of training samples from the training samples as test samples;
and verifying the accuracy of the driving behavior analysis model by adopting the test sample verification.
10. The driving behavior analysis method of claim 9, wherein the performing accuracy verification on the driving behavior analysis model using the test sample verification comprises:
randomly dividing the test sample into N groups of test sub-samples, wherein N is a natural number and is more than or equal to 3;
and carrying out accuracy verification on a group of test subsamples except the N-1 group by adopting any N-1 groups of test subsamples until the N groups of test subsamples are verified.
11. The driving behavior analysis method according to claim 1, wherein the calculating of the collision occurrence probability of the vehicle includes:
acquiring a threshold adjustment coefficient;
acquiring weights respectively corresponding to the driving behaviors in the first mapping relation from the trained driving behavior analysis model according to the threshold adjustment coefficients, wherein the driving behaviors respectively correspond to different weights under different threshold adjustment coefficients in the driving behavior analysis model;
and calculating the collision occurrence probability of the vehicle according to the acquired weights of the driving behaviors respectively corresponding to the driving behaviors in the first mapping relation.
12. The driving behavior analysis method according to claim 1, characterized by further comprising:
and when a preset model updating triggering event is monitored, acquiring the driving behavior data of the vehicle as an updating sample, adopting the updating sample to update and train the driving behavior analysis model, and using the driving behavior analysis model obtained by updating and training as a trained driving behavior analysis model.
13. The driving behavior analysis method according to claim 1, characterized in that the first period is a period of time with a current time as an end time.
14. A driving behavior analysis device characterized by comprising: the device comprises an acquisition unit, a processing unit, an input unit, a driving behavior analysis model, a calculation unit and a first output unit, wherein:
the acquisition unit is suitable for acquiring driving behavior data of a vehicle within a preset first time period and driving mileage of the vehicle within the first time period;
the processing unit is suitable for processing the driving behavior data within the first duration to obtain the occurrence frequency of each driving behavior within a preset mileage;
the input unit is suitable for inputting the obtained times of the driving behaviors within the preset mileage into the trained driving behavior analysis model;
the input parameters of the driving behavior analysis model comprise driving behaviors, a first mapping relation exists between the driving behaviors and the collision occurrence probability of the vehicle, and the corresponding weight of each driving behavior in the driving behavior analysis model in the first mapping relation is obtained through training;
the calculation unit is suitable for calculating the collision occurrence probability of the vehicle by adopting the driving behavior analysis model;
the first output unit adapted to output the collision occurrence probability of the vehicle calculated by the calculation unit;
wherein, still include: a second training unit adapted to train the driving behavior analysis model in the following manner:
acquiring a training sample, wherein the training sample comprises driving behavior data and a collision degree value;
marking the training samples with the collision degree value larger than or equal to a threshold value as positive samples, and marking the training samples with the collision degree value smaller than the threshold value as negative samples;
and performing logistic regression training on the positive sample and the negative sample according to the driving behavior data by adopting a logistic regression algorithm to obtain the driving behavior analysis model.
15. The driving behavior analysis device according to claim 14, characterized by further comprising: analysis unit and second output unit, wherein:
the analysis unit is suitable for analyzing the influence of each driving behavior on the collision occurrence probability of the vehicle by adopting the driving behavior analysis model according to the collision occurrence probability of the vehicle and the data corresponding to all the driving behaviors in the first time period after calculating the collision occurrence probability of the vehicle, and generating a driving behavior improvement suggestion aiming at the influence of each driving behavior on the collision occurrence probability of the vehicle;
the second output unit is adapted to output the driving behavior improvement advice generated by the analysis unit.
16. The driving behavior analysis device according to claim 15, wherein the processing unit is adapted to process the driving behavior data in the first duration according to time intervals respectively, so as to obtain the number of times that each driving behavior in the corresponding time interval occurs within a preset mileage;
the input unit is suitable for inputting the frequency of the driving behaviors within the preset mileage within the obtained corresponding time interval into the trained driving behavior analysis model;
the calculating unit is suitable for calculating the collision occurrence probability of the vehicle by adopting the driving behavior analysis model, and the input parameters of the driving behavior analysis model have corresponding weights in different time intervals.
17. The driving behavior analysis device according to claim 14, wherein the input parameters of the driving behavior analysis model further include a collision event related factor, the collision occurrence probability of the vehicle and the driving behavior and collision event related factor have a second mapping relationship, and weights corresponding to the driving behavior and collision event related factors in the driving behavior analysis model in the second mapping relationship are obtained through training, respectively.
18. The driving behavior analysis device according to claim 17, wherein the obtaining unit is further adapted to obtain a collision event related factor corresponding to the vehicle in the first duration;
the processing unit is further suitable for processing the data of the factors related to the collision event in the first time length;
the input unit is further adapted to input the processed collision event related factor data to the trained driving behavior analysis model.
19. The driving behavior analysis device according to claim 17, characterized by further comprising: a first training unit adapted to train the driving behavior analysis model in the following way:
acquiring a training sample, wherein the training sample comprises driving behavior data, collision event associated factor data and a collision degree value;
marking the training samples with the collision degree value larger than or equal to a preset threshold value as positive samples, and marking the training samples with the collision degree value smaller than the threshold value as negative samples;
and performing logistic regression training on the positive sample and the negative sample by adopting a logistic regression algorithm according to the driving behavior data and the collision event associated factor data to obtain the driving behavior analysis model.
20. The driving behavior analysis device according to claim 14 or 19, further comprising a test unit adapted to randomly select a preset number of training samples from the training samples as test samples; and verifying the accuracy of the driving behavior analysis model by adopting the test sample verification.
21. The driving behavior analysis device according to claim 20, wherein the test unit is adapted to randomly divide the test samples into N groups of test subsamples, N being a natural number and N ≧ 3; and carrying out accuracy verification on a group of test subsamples except the N-1 group by adopting any N-1 groups of test subsamples until the N groups of test subsamples are verified.
22. The driving behavior analysis device according to claim 14, characterized in that the calculation unit is adapted to obtain a threshold adjustment coefficient; acquiring weights corresponding to the driving behaviors from the trained driving behavior analysis model according to the threshold adjustment coefficients, wherein the driving behaviors correspond to different weights in the first mapping relation under different threshold adjustment coefficients in the driving behavior analysis model; and calculating the collision occurrence probability of the vehicle according to the acquired weights of the driving behaviors respectively corresponding to the driving behaviors in the first mapping relation.
23. The driving behavior analysis device according to claim 14, characterized by further comprising: and the updating unit is suitable for acquiring the driving behavior data of the vehicle as an updating sample when a preset model updating triggering event is monitored, updating and training the driving behavior analysis model by adopting the updating sample, and taking the driving behavior analysis model obtained by updating and training as a trained driving behavior analysis model.
24. A driving behavior analysis system, comprising: a data acquisition apparatus and a driving behavior analysis apparatus according to any one of claims 14 to 23, wherein:
the data acquisition device is suitable for acquiring driving behavior data of a vehicle within a preset first time period and driving mileage of the vehicle within the first time period.
25. The driving behavior analysis system of claim 24, wherein the data collection device is adapted to collect the driving behavior data of the vehicle within a preset first time period through a driving state sensor mounted on the vehicle, or to obtain the driving behavior data of the vehicle within the preset first time period from a vehicle memory.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611262696.0A CN108268678B (en) | 2016-12-30 | 2016-12-30 | Driving behavior analysis method, device and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611262696.0A CN108268678B (en) | 2016-12-30 | 2016-12-30 | Driving behavior analysis method, device and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108268678A CN108268678A (en) | 2018-07-10 |
| CN108268678B true CN108268678B (en) | 2021-07-23 |
Family
ID=62755268
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201611262696.0A Active CN108268678B (en) | 2016-12-30 | 2016-12-30 | Driving behavior analysis method, device and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108268678B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110143202A (en) * | 2019-04-09 | 2019-08-20 | 南京交通职业技术学院 | A kind of dangerous driving identification and method for early warning and system |
| CN113968234B (en) * | 2021-11-29 | 2023-05-02 | 深圳市科莱德电子有限公司 | Vehicle auxiliary driving control method and device and vehicle-mounted controller |
| CN117422169B (en) * | 2023-10-18 | 2024-06-25 | 酷哇科技有限公司 | Vehicle insurance user driving behavior analysis and prediction method and device based on causal intervention |
Family Cites Families (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TWI447039B (en) * | 2011-11-25 | 2014-08-01 | Driving behavior analysis and warning system and method thereof | |
| CN103310202B (en) * | 2013-06-27 | 2017-04-05 | 西安电子科技大学 | A kind of system and method for ensureing driving safety |
| WO2015061712A1 (en) * | 2013-10-24 | 2015-04-30 | Tourmaline Labs, Inc. | Systems and methods for collecting and transmitting telematics data from a mobile device |
| CN104978777A (en) * | 2014-04-11 | 2015-10-14 | 比亚迪股份有限公司 | System, device and method for analyzing driving behavior |
| CN104050551B (en) * | 2014-05-23 | 2017-03-29 | 北京中交兴路信息科技有限公司 | A kind of intelligent goods dispatch method and system predicted based on vehicle destination |
| CN104063914A (en) * | 2014-07-01 | 2014-09-24 | 深圳市成为智能交通系统有限公司 | Terminal achieving driving risk management and use method thereof |
| CN105488046A (en) * | 2014-09-16 | 2016-04-13 | 钛马信息网络技术有限公司 | Big data analysis system based on vehicle insurance services |
| CN104200267B (en) * | 2014-09-23 | 2016-03-30 | 清华大学 | A kind of vehicle drive Economic Evaluation system and evaluation method |
| CN104494601A (en) * | 2014-12-18 | 2015-04-08 | 清华大学苏州汽车研究院（吴江） | Driving behavior analysis and driving assistance system based on OBD |
| CN105046765B (en) * | 2015-08-19 | 2016-05-04 | 福建省汽车工业集团云度新能源汽车股份有限公司 | Improve the method for driving behavior based on drive recorder |
| CN105679020B (en) * | 2016-01-29 | 2017-12-05 | 深圳市美好幸福生活安全系统有限公司 | A kind of driving behavior analysis apparatus and method |
| CN106095963B (en) * | 2016-06-17 | 2019-12-06 | 上海经达信息科技股份有限公司 | Vehicle driving behavior analysis big data public service platform under internet + era |
| CN106240571B (en) * | 2016-08-03 | 2019-01-25 | 东软集团股份有限公司 | Driving behavior analysis method and apparatus |
-
2016
- 2016-12-30 CN CN201611262696.0A patent/CN108268678B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN108268678A (en) | 2018-07-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11407410B2 (en) | Method and system for estimating an accident risk of an autonomous vehicle | |
| EP3578433B1 (en) | Method for estimating an accident risk of an autonomous vehicle | |
| KR101617349B1 (en) | Diagnostic system and method for the analysis of driving behavior | |
| CN109670970B (en) | Driving behavior scoring method and device and computer readable storage medium | |
| Andre et al. | Driving cycles for emission measurements under European conditions | |
| EP3239686A1 (en) | Method for determining driving characteristics of a vehicle | |
| DE202019005891U1 (en) | Vehicle telematics of vehicle accidents | |
| CN106650157B (en) | Method, device and system for estimating fault occurrence probability of vehicle parts | |
| CN110615001A (en) | Driving safety reminding method, device and medium based on CAN data | |
| US20220340148A1 (en) | Method for estimating an accident risk of an autonomous vehicle | |
| CN109493566B (en) | Fatigue driving prediction method based on GPS data | |
| US20060106515A1 (en) | Motor vehicle operating data collection and analysis | |
| CN109263648B (en) | Driving behavior evaluation method, device and equipment | |
| CN113119985B (en) | Automobile driving data monitoring method, device, equipment and storage medium | |
| Castignani et al. | An evaluation study of driver profiling fuzzy algorithms using smartphones | |
| CN108268678B (en) | Driving behavior analysis method, device and system | |
| CN104590275A (en) | Driving behavior analyzing method | |
| CN106022846A (en) | Automobile insurance pricing method, second-hand automobile pricing method and corresponding devices | |
| CN112937591B (en) | Driving safety monitoring method, device, equipment and computer readable storage medium | |
| CN107918826B (en) | Driver evaluation and scheduling method for driving environment perception | |
| CN112598169A (en) | Traffic operation situation assessment method, system and device | |
| Wu et al. | Clustering of several typical behavioral characteristics of commercial vehicle drivers based on GPS data mining: Case study of highways in China | |
| EP3154839B1 (en) | Driving assistance method for increasing the awareness of drivers of vehicles about fuel consumption and/or other consumption sources of the vehicle | |
| CN113928333B (en) | Energy consumption reduction method and system based on auxiliary driving | |
| CN114511178A (en) | Monitoring method and system for safe driving behaviors of shared trip users |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |