A data security warning system and application method for intelligent connected vehicles

Abstract

The present invention relates to the field of intelligent driving technology, and specifically to a data safety warning system and application method for an intelligent networked vehicle, including: obtaining a driving path curve and a predicted driving curve of a target driver; screening the driving path curve to obtain a matching driving curve; classifying the matching driving curve to obtain a conventional behavior class and a special behavior class; obtaining a reaction time coefficient of the target driver according to the amount of data contained in the conventional behavior class and the special behavior class and the length of the data distribution; obtaining an adjustment coefficient of the target driver according to the amount of data contained in the conventional behavior class and the shape distribution, combined with the difference distribution between the predicted driving curve and the matching driving curve in the conventional behavior class; adjusting the predicted driving curve, comparing the adjusted result with the driving path of a normal situation, and providing a safety warning to the target driver. The present invention can effectively characterize the predicted driving situation of the driver, and has a better warning effect on the driver.

Description

Technical Field

[0001] The present invention relates to the field of intelligent driving technology, and in particular to an intelligent networked vehicle data security warning system and an application method. Background Art

[0002] In today's era, intelligent connected vehicles have become a significant trend in the automotive industry's pursuit of intelligent and smart development. Through close integration with the Internet, intelligent connected vehicles not only achieve interoperability between vehicles, but also improve the interaction efficiency between vehicles, roads, and pedestrians, thereby greatly optimizing the driving experience and enhancing driving safety.

[0003] Existing intelligent cars based on intelligent networking mainly analyze driving videos collected by different cars on different road sections, and then extract information such as the car's location and moving speed, build a driving prediction model to obtain the driver's driving path, and warn the driver in advance of possible safety problems through the predicted driving path planning. However, different drivers have different driving behavior habits, and their response time to crisis situations is also different. It is difficult to accurately represent the driver's predicted driving situation by only obtaining the driver's predicted driving route through the prediction model, and the warning effect on the driver is poor. Summary of the invention

[0004] In order to solve the technical problem that the existing methods are difficult to accurately represent the driver's predicted driving situation and the warning effect on the driver is poor, the purpose of the present invention is to provide a smart connected vehicle data security warning system and application method, and the technical solutions adopted are as follows:

[0005] In a first aspect, the present invention provides an application method of a data security warning system for an intelligent networked vehicle, comprising:

[0006] Obtaining the driving path curve of each road in the vehicle driving record of the target driver, predicting the driving state of the target driver, and obtaining the predicted driving curve of the target driver;

[0007] According to the inclination of the predicted driving curve and the driving path curve, the driving path curve is screened to obtain a matching driving curve; according to the difference between the predicted driving curve and each matching driving curve, combined with the data distribution of the predicted driving curve and the data distribution of each matching driving curve, the matching driving curve is classified to obtain a normal behavior class and a special behavior class;

[0008] According to the amount of data contained in the regular behavior class and the special behavior class, and the length of the data distribution of the matching driving curves in the regular behavior class and the special behavior class, the reaction time coefficient of the target driver is obtained;

[0009] According to the amount of data contained in the conventional behavior class and the shape distribution of the matching driving curve therein, combined with the difference distribution between the predicted driving curve and the matching driving curve in the conventional behavior class, the adjustment coefficient of the target driver is obtained;

[0010] The predicted driving curve is adjusted according to the reaction time coefficient and the adjustment coefficient, and the adjusted result is compared with the driving path under normal conditions to issue a safety warning to the target driver.

[0011] Preferably, the screening of the driving path curve to obtain a matching driving curve according to the inclination of the predicted driving curve and the driving path curve specifically includes:

[0012] Respectively obtain a first slope corresponding to the position of the starting data point of the predicted driving curve and a second slope corresponding to the position of the ending data point; obtain a first slope corresponding to the position of the starting data point of each driving path curve and a second slope corresponding to the position of the ending data point;

[0013] The difference between the predicted driving curve and the first slope corresponding to each driving path curve is recorded as a first difference, and the difference between the predicted driving curve and the second slope corresponding to each driving path curve is recorded as a second difference;

[0014] When both the first difference and the second difference are smaller than a preset difference threshold, the corresponding driving path curve is recorded as a matching driving curve.

[0015] Preferably, the matching driving curves are classified according to the difference between the predicted driving curve and each matching driving curve, in combination with the data distribution of the predicted driving curve and the data distribution of each matching driving curve, to obtain a conventional behavior class and a special behavior class, specifically including:

[0016] Acquire a first data point on the predicted driving curve where the difference between the slope of the data point and the first slope is greater than a preset difference threshold, and determine an initial reaction segment of the predicted driving curve based on a data distribution between the initial data point and the first data point;

[0017] According to the difference and the curve length between the predicted driving curve and each matching driving curve, combined with the difference between the initial reaction segment of the predicted driving curve and the initial reaction segment of each matching driving curve, a classification metric distance between the predicted driving curve and each matching driving curve is obtained;

[0018] According to the classification metric distance, the matching driving curves are classified to obtain a conventional behavior class and a special behavior class.

[0019] Preferably, the method for obtaining the classification metric distance between the predicted driving curve and each matching driving curve is specifically as follows:

[0020] determining a difference coefficient based on a difference between a curve length of the predicted driving curve and a curve length of each matching driving curve;

[0021] Determine a first distance coefficient based on a DTW distance between the predicted driving curve and each matching driving curve; determine a second distance coefficient based on a DTW distance between an initial reaction segment of the predicted driving curve and an initial reaction segment of each matching driving curve;

[0022] According to the difference coefficient, the first distance coefficient and the second distance coefficient, a classification metric distance between the predicted driving curve and each matching driving curve is obtained, and the difference coefficient, the first distance coefficient and the second distance coefficient are all positively correlated with the classification metric distance.

[0023] Preferably, the matching driving curves are classified according to the classification metric distance to obtain conventional behavior classes and special behavior classes, specifically including:

[0024] Based on the classification metric distance, a clustering algorithm is used to classify the matching driving curves to obtain cluster clusters, and the cluster cluster corresponding to the maximum number of matching driving curves contained in each cluster cluster is taken as a regular behavior class, and all cluster clusters except the regular behavior class in all cluster clusters are taken as special behavior classes.

[0025] Preferably, the reaction time coefficient of the target driver is obtained according to the amount of data included in the conventional behavior class and the special behavior class, and the length of the data distribution of the matching driving curves in the conventional behavior class and the special behavior class, and specifically includes:

[0026] Recording the average of the curve lengths of the initial reaction segments of all matching driving curves in the regular behavior class as a first average, and calculating the product of the ratio of the total number of all matching driving curves in the regular behavior class and the first average to obtain a first time coefficient of the regular behavior class;

[0027] Recording the average of the curve lengths of the initial reaction segments of all matching driving curves in all special behavior classes as a second average, and calculating the product of the ratio of the total number of all matching driving curves in all special behavior classes and the second average to obtain a second time coefficient of the special behavior class;

[0028] The maximum value of the first time coefficient and the second time coefficient is used as the reaction time coefficient of the target driver.

[0029] Preferably, the adjustment coefficient of the target driver is obtained based on the amount of data contained in the conventional behavior class and the shape distribution of the matching driving curve therein, combined with the difference distribution between the predicted driving curve and the matching driving curve in the conventional behavior class, and specifically includes:

[0030] The mean of the kurtosis of all matching driving curves in the conventional behavior class is calculated and recorded as the shape characteristic factor;

[0031] Perform DTW matching on the predicted driving curve and each matching driving curve respectively, and obtain the matching data point corresponding to each data point on the predicted driving curve and each matching driving curve;

[0032] For any matching driving curve, the cumulative sum of the differences between the data points on the predicted driving curve and the corresponding matching data points on the matching driving curve is taken as the feature difference cumulative value;

[0033] The adjustment coefficient of the target driver is determined based on the product of the average of the cumulative values ​​of the characteristic differences corresponding to the predicted driving curve and all the matching driving curves, the shape characteristic factor, and the proportion of the total number of all matching driving curves in the conventional behavior class, and the value of the adjustment coefficient is [-1,1].

[0034] Preferably, the adjusting the predicted driving curve according to the reaction time coefficient and the adjustment coefficient specifically includes:

[0035] Taking the starting data point of the predicted driving curve as the starting point, obtaining a fitting behavior line segment based on the first slope of the predicted driving curve, wherein the length of the fitting behavior line segment is equal to the value of the reaction time coefficient;

[0036] The sum of the value 1 and the adjustment coefficient is used as the first proportional coefficient, and the difference between the value 1 and the adjustment coefficient is used as the second proportional coefficient;

[0037] Starting from the horizontal coordinate corresponding to the end point of the fitting behavior segment on the predicted driving curve, the vertical coordinate of the predicted driving curve is proportionally adjusted according to the first proportional coefficient, and the horizontal coordinate of the predicted driving curve is proportionally adjusted according to the second proportional coefficient.

[0038] Preferably, the step of comparing the adjusted result with the normal driving path and providing a safety warning to the target driver specifically includes:

[0039] When there is an intersection between the adjusted result curve and the straight line of the normal driving path, the predicted driving time of the target driver is determined based on the curve length from the adjusted result curve to the intersection and the predicted speed of the target driver, and the normal driving time of the target driver to the intersection under normal circumstances is obtained. If the predicted driving time is equal to the normal driving time, a safety warning is issued to the target driver.

[0040] In a second aspect, the present invention provides a smart connected vehicle data security warning system, which is used to implement the steps of an application method of the smart connected vehicle data security warning system, and the system includes:

[0041] A data acquisition module is used to obtain the driving path curve of each road in the vehicle driving record of the target driver, predict the driving state of the target driver, and obtain the predicted driving curve of the target driver;

[0042] A feature analysis module, for screening the driving path curves to obtain matching driving curves according to the inclination of the predicted driving curve and the driving path curve; and for classifying the matching driving curves to obtain conventional behavior classes and special behavior classes according to the difference between the predicted driving curve and each matching driving curve, combined with the data distribution of the predicted driving curve and the data distribution of each matching driving curve;

[0043] A time analysis module, used to obtain the reaction time coefficient of the target driver according to the amount of data contained in the regular behavior class and the special behavior class, and the length of the data distribution of the matching driving curve in the regular behavior class and the special behavior class;

[0044] An adjustment analysis module, for obtaining an adjustment coefficient of a target driver according to the amount of data contained in the conventional behavior class and the shape distribution of the matching driving curve therein, combined with the difference distribution between the predicted driving curve and the matching driving curve in the conventional behavior class;

[0045] The safety warning module is used to adjust the predicted driving curve according to the reaction time coefficient and the adjustment coefficient, and to compare the adjusted result with the driving path under normal conditions to issue a safety warning to the target driver.

[0046] The embodiments of the present invention have at least the following beneficial effects:

[0047] The present invention first obtains the driving path curve in the historical record of the target driver and the predicted driving curve of the prediction result, and provides a data basis for the subsequent analysis of the characteristic differences between the prediction result and the historical record. Then, it is necessary to screen out the historical records that are relatively similar to the road conditions and the driving behavior of the driver in the prediction result, that is, the matching driving curve, and combine the differences between the relatively similar historical records and the prediction results to classify the matching driving curves, and quantitatively characterize the driving behavior characteristics of the target driver in the historical records, that is, including conventional behavior classes and special behavior classes. Further, two aspects are considered for the prediction results. On the one hand, the data volume and data distribution length of different behavior characteristics are analyzed, and the reaction time coefficient is quantitatively characterized to reflect the degree of reaction of the target driver to the emergency. On the other hand, the data volume size and shape distribution of the conventional behavior characteristics of the target driver are analyzed, and the adjustment coefficient is quantified to reflect the size of the predicted driving curve to be adjusted to the conventional behavior characteristics in combination with the differences between the conventional behavior characteristics and the predicted driving curve. Finally, the prediction results are adjusted in combination with the characteristic analysis results of the two aspects. This embodiment fully combines the posture behavior habit characteristics of the driver to adjust the prediction results, which can effectively characterize the driver's predicted driving situation and has a better warning effect on the driver. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0049] Figure 1 It is a flowchart of the steps of an application method of a data security warning system for an intelligent networked vehicle provided by the present invention;

[0050] Figure 2 is a flow chart of the steps of the method for classifying matching driving curves provided by the present invention;

[0051] Figure 3 It is a flow chart of the steps of the method for obtaining the adjustment coefficient of the target driver provided by the present invention;

[0052] Figure 4 It is a system block diagram of a data security warning system for intelligent networked vehicles provided by the present invention;

[0053] Figure 5 The present invention provides a schematic diagram of the structure of a computer device. DETAILED DESCRIPTION

[0054] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the specific implementation method, structure, features and effects of a smart connected vehicle data security warning system and application method proposed by the present invention in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0055] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0056] The specific scheme of a smart connected vehicle data security warning system and application method provided by the present invention is described in detail below with reference to the accompanying drawings.

[0057] See also Figure 1 , which shows a flowchart of a method for applying a data security warning system for a smart connected vehicle provided by an embodiment of the present invention, the method comprising the following steps:

[0058] Step S100, obtaining the driving path curve of each road in the vehicle driving record of the target driver, predicting the driving state of the target driver, and obtaining the predicted driving curve of the target driver.

[0059] At present, the driving model of intelligent connected vehicles mainly uses neural networks to predict the driver's driving behavior, and then obtains the driver's driving route according to the actual environment and displays it on the central control display. Considering that different drivers have different driving habits, the prediction model is constructed by analyzing the historical driving behavior habits of the driver to be analyzed to obtain the prediction results.

[0060] In this embodiment, a deep neural network can be deployed as an end-to-end driving solution, starting from perception and ending with path planning, using a Markov decision process (MDP) for reinforcement learning. Using a deep neural network model to predict the vehicle's driving route is a relatively mature deep learning technology, that is, this method is a well-known technology and will not be introduced in detail here. Based on this, based on the target driver's vehicle driving record, a deep neural network model can be used to predict the target driver's driving route, and finally obtain the target driver's predicted driving path. It should be noted that, considering that the sample of a single driver is single when performing path prediction, the driving record samples of all drivers are selected for training, so that the model can adapt to the driving route prediction of most drivers.

[0061] First, it is necessary to perform data fitting on the driver's historical records, and fit the route of the vehicle into a curve for expression. It can be understood that this embodiment uses the driver to be analyzed as the target for data analysis. Specifically, the vehicle currently driven by the target driver is taken as the origin, and the direction of travel of the target driver is taken as the Y-axis direction, that is, the Y-axis is parallel to the road markings, and the X-axis is perpendicular to the road markings. A plane rectangular coordinate system is constructed, and the route of each travel of the target driver on the road is mapped into the coordinate system, and the route is smoothed to obtain a driving path curve.

[0062] It should be noted that the driving path curve is the performance of the target driver's driving route in the historical vehicle driving record, and a driving path curve corresponds to a straight road. In the same way, the target driver's path prediction result output by the model is mapped to the coordinate system to obtain the target driver's predicted driving curve.

[0063] Step S200, based on the inclination of the predicted driving curve and the driving path curve, the driving path curve is screened to obtain a matching driving curve; based on the difference between the predicted driving curve and each matching driving curve, combined with the data distribution of the predicted driving curve and the data distribution of each matching driving curve, the matching driving curve is classified to obtain a normal behavior class and a special behavior class.

[0064] In the predicted result of the target driver's driving route, if the target driver drives at a certain speed according to the predicted result, if there is a situation where other normally driving vehicles interfere with each other, it is necessary to warn the driver of the vehicle in advance. In the actual driving process, the posture and behavior habits of different drivers may cause certain differences between the actual driving situation and the predicted results, which in turn leads to poor warning effects. Therefore, this embodiment jointly analyzes the driver's driving behavior habits and response habits to crisis situations through the predicted results and historical records, and then adjusts the predicted results. Based on this, it is first necessary to screen out historical records that are more similar to the road conditions and driver driving behaviors in the predicted results to characterize the driver's driving behavior characteristics.

[0065] Firstly, the predicted driving curve corresponding to the prediction result of the target driver is compared with the inclination of all driving path curves in the historical records, and the parts with similar route change trends to the predicted form are preliminarily screened out to provide a data basis for subsequent characteristic behavior classification.

[0066] It is understandable that changes in the vehicle's driving direction are all caused by the steering guidance of the front wheels. When the target driver is driving the vehicle, when the front wheels adjust their directions, the body of the vehicle will change following the direction of the front wheels, which is reflected in the driving route as an angular deviation in the vehicle's driving path. The difference in the same deviation behavior lies in the influence of the operation time and road conditions. Therefore, it is necessary to preliminarily screen out historical records that are relatively similar to the predicted results of the target driver.

[0067] Specifically, the first slope corresponding to the position of the starting data point of the predicted driving curve and the second slope corresponding to the position of the ending data point are obtained respectively; the first slope corresponding to the position of the starting data point and the second slope corresponding to the position of the ending data point of each driving path curve are obtained. Among them, the starting data point can be regarded as the first data point on the corresponding curve, and the method for obtaining the slope of the data point is a well-known technology and will not be introduced in detail here. The first slope reflects the path deviation and steering of the corresponding curve at the initial position, and the second slope reflects the path deviation and steering of the corresponding curve at the ending position.

[0068] Then, by comparing the difference between the predicted result and the historical record in the path deviation and steering at the corresponding same position, the historical records with smaller differences and more similarity can be screened out. Specifically, the difference between the predicted driving curve and the first slope corresponding to each driving path curve is recorded as the first difference, and the difference between the predicted driving curve and the second slope corresponding to each driving path curve is recorded as the second difference. In this embodiment, the first difference is obtained by calculating the absolute value of the difference between the predicted driving curve and the first slope corresponding to each driving path curve, and the second difference is obtained by calculating the absolute value of the difference between the predicted driving curve and the second slope corresponding to each driving path curve.

[0069] The first difference reflects the difference between the predicted driving curve corresponding to the prediction result of the target driver and the path deviation direction of any driving path curve in the historical record at the initial position, and the second difference reflects the difference between the predicted driving curve corresponding to the prediction result of the target driver and the path deviation direction of any driving path curve in the historical record at the terminal position.

[0070] When both the first difference and the second difference are smaller than the preset difference threshold, the corresponding driving path curve is recorded as the matching driving curve. In this embodiment, the difference threshold is 0.1. When both the first difference and the second difference meet the condition of small difference, it means that the predicted driving curve corresponding to the predicted result of the target driver is similar to any driving path curve in the historical record.

[0071] Furthermore, although the path turning conditions of the initially screened matching driving curves at the starting and ending positions are similar to the predicted results of the target driver, there may be certain differences in factors such as the length of the driving path, intermediate turns, etc. These differences may be caused by the target driver's corresponding driving habits or reaction habits to emergencies. Therefore, the screened matching driving curves can be feature classified.

[0072] In this embodiment, if Figure 2 As shown, the method steps of classifying the matching driving curve can be implemented by steps S201 to S203.

[0073] Step S201, obtaining a first data point on the predicted driving curve whose slope differs from a first slope by more than a preset difference threshold, and determining an initial reaction segment of the predicted driving curve based on a data distribution between the initial data point and the first data point.

[0074] Assuming that the target driver's driving route at the initial position on the predicted driving curve has a certain degree of deviation, when the deviation of a data point on the predicted driving curve from the initial position changes significantly, it can be roughly considered that the target driver has made a fine-tuning operation to the deviation to correct the driving route. The period from the initial position to the fine-tuning operation can be roughly regarded as the driver's reaction stage to the actual driving situation. This initial reaction stage can well characterize the driver's driving behavior habits.

[0075] Based on this, starting from the starting data point on the predicted driving curve, each data point is traversed in turn, and the absolute value of the difference between the slope of each data point and the first slope is compared to see whether it is greater than the difference threshold. The data point corresponding to the first appearance of a slope greater than the difference threshold is regarded as the beginning of the fine-tuning operation, and then the curve portion between the initial starting point on the predicted driving curve and the data point corresponding to the start of the fine-tuning operation can be regarded as the initial reaction segment of the target driver, which characterizes the target driver's reaction to driving deviation.

[0076] Step S202, according to the difference between the predicted driving curve and each matching driving curve and the curve length, combined with the difference between the initial reaction segment of the predicted driving curve and the initial reaction segment of each matching driving curve, obtain the classification metric distance between the predicted driving curve and each matching driving curve.

[0077] In this embodiment, the differences between the predicted driving curve and the matched driving curve in three aspects are analyzed, and the difference distances between different curves are quantified, including the difference in data distribution between different curves, the difference in curve length between different curves, and the difference in initial reaction segments between different curves, thereby avoiding the problem of inaccurate classification caused by only considering the difference in data distribution between different curves.

[0078] Specifically, a difference coefficient is determined based on the difference between the curve length of the predicted driving curve and the curve length of each matching driving curve; a first distance coefficient is determined based on the DTW distance between the predicted driving curve and each matching driving curve; a second distance coefficient is determined based on the DTW distance between the initial reaction segment of the predicted driving curve and the initial reaction segment of each matching driving curve; and a classification metric distance between the predicted driving curve and each matching driving curve is obtained based on the difference coefficient, the first distance coefficient, and the second distance coefficient, wherein the difference coefficient, the first distance coefficient, and the second distance coefficient are all positively correlated with the classification metric distance.

[0079] In this embodiment, taking any matching driving curve as an example, the classification metric distance between the predicted driving curve of the target driver and the nth matching driving curve can be expressed as:

[0080]

[0081] in, represents the classification metric distance between the predicted driving curve of the target driver and the nth matching driving curve, represents the predicted driving curve of the target driver, represents the nth matching driving curve, Indicates the length of the nth matching driving curve, represents the curve length of the predicted driving curve, represents the DTW distance between the predicted driving curve of the target driver and the nth matching driving curve, represents the DTW distance between the initial reaction segment of the predicted driving curve and the initial reaction segment of the nth matching driving curve, represents the initial reaction segment of the predicted driving curve, Represents the initial reaction segment of the nth matching driving curve.

[0082] It should be noted that the method for obtaining the curve length is a well-known technology and will not be introduced here. It can be understood that, whether it is a predicted driving curve or a matched driving curve, the curve length represents the actual path length of the driver on the road.

[0083] is the difference coefficient. The larger its value is, the greater the difference between the length of the predicted driving curve and the matching driving curve is. is the first distance coefficient, is the second distance coefficient. The greater the difference between the two, the greater the metric distance between the corresponding predicted driving curve and the matching driving curve.

[0084] Step S203: classify the matching driving curves according to the classification metric distance to obtain a normal behavior class and a special behavior class.

[0085] Based on the classification metric distance between the predicted driving curve and each matching driving curve, the matching driving curves are classified into clusters using a clustering algorithm. The clustering algorithm can be processed using the DBSCAN algorithm, which can be selected by the implementer according to the specific implementation scenario. The cluster corresponding to the maximum number of matching driving curves contained in each cluster is taken as a regular behavior class, and all clusters except the regular behavior class in all clusters are taken as special behavior classes.

[0086] Each cluster can represent the data set corresponding to each behavior habit. The number of matching driving curves in the cluster can reflect the amount of data corresponding to different behavior habits. When the amount of data contained in the cluster is more, it means that the behavior habit corresponding to the cluster belongs to a more conventional behavior, while the behavior habits corresponding to other clusters can be roughly regarded as more special behavior reactions. Therefore, the matching driving curves in the conventional behavior class reflect the more conventional driving behavior habits, and the matching driving curves in the special behavior class reflect the more special driving behavior habits.

[0087] Step S300, obtaining the reaction time coefficient of the target driver according to the amount of data included in the regular behavior class and the special behavior class, and the length of the data distribution of the matching driving curves in the regular behavior class and the special behavior class.

[0088] Adjusting the predicted result of the target driver's driving path needs to be considered from two aspects. The target driver has a certain reaction stage to the initial position deviation. The length of this reaction time can better represent the driver's reaction habits to emergencies. Then the predicted driving curve can be divided into two parts for adjustment, one for the initial reaction stage and the other for the remaining driving route except the initial reaction stage. Based on this, the reaction time of the initial reaction stage is first analyzed to quantify the coefficient of adjustment of the curve of the target driver's initial reaction stage.

[0089] Specifically, the mean of the curve lengths of the initial reaction segments of all matching driving curves in the normal behavior class is recorded as the first mean, and the product of the proportion of the total number of all matching driving curves in the normal behavior class and the first mean is calculated to obtain the first time coefficient of the normal behavior class. The mean of the curve lengths of the initial reaction segments of all matching driving curves in all special behavior classes is recorded as the second mean, and the product of the proportion of the total number of all matching driving curves in all special behavior classes and the second mean is calculated to obtain the second time coefficient of the special behavior class. The maximum value of the first time coefficient and the second time coefficient is used as the reaction time coefficient of the target driver.

[0090] In this embodiment, as a specific example, the calculation formula of the target driver's reaction time coefficient can be expressed as:

[0091]

[0092] in, represents the target driver's reaction time coefficient, t represents time, Represents the total number of all matching driving curves in the regular behavior class, Represents the total number of all matching driving curves in all special behavior classes, Represents the total number of all matching driving curves, The first mean represents the mean of the curve lengths of the initial reaction segments of all matching driving curves in the conventional behavior class; It represents the mean of the curve lengths of the initial reaction segments of all matching driving curves in all special behavior classes, that is, the second mean; max represents a function for finding the maximum value.

[0093] Indicates the proportion of the total number of matching driving curves in the conventional behavior class, Indicates the proportion of the total number of all matching driving curves in all special behavior classes. is the first time coefficient, is the second time coefficient. The first time coefficient reflects the overall situation of the curve length of the matching driving curve in the regular behavior class, and the second time coefficient reflects the overall situation of the curve length of the matching driving curve in the special behavior class. Then the maximum value between the two is selected to finally represent the initial reaction distance of the target driver in the historical records. That is, the reaction time coefficient represents the possible reaction distance of the target driver to the emergency situation.

[0094] Step S400, obtaining the adjustment coefficient of the target driver according to the amount of data contained in the conventional behavior class and the shape distribution of the matching driving curve therein, combined with the difference distribution between the predicted driving curve and the matching driving curve in the conventional behavior class.

[0095] The larger the amount of data contained in the conventional behavior class, the more cases the target driver's historical records show conventional driving behavior habits. Therefore, when adjusting the prediction results of the target driver, it is more necessary to adjust the predicted driving curve to the characteristics of the matching driving curve of the conventional behavior class. The amount of data contained in the special behavior class is used to characterize the unstable operation of the target driver. The larger the amount of data contained in the special behavior class, the more difficult it is to predict the behavior of the target driver. Therefore, when adjusting the prediction results of the target driver, the adjustment ratio should be minimized. Based on this, by analyzing the amount of data contained in the conventional behavior class and the shape distribution of the matching driving curve in the conventional behavior class, combined with the difference between the predicted driving curve and the matching driving curve in the conventional behavior class, the coefficient expression that needs to be adjusted to characterize the predicted driving curve of the target driver can be quantified.

[0096] In this embodiment, if Figure 3 As shown, the method for obtaining the adjustment coefficient of the target driver can be implemented by steps S401 to S404.

[0097] Step S401, calculating the mean of the kurtosis of all matching driving curves in the conventional behavior class and recording it as the shape feature factor.

[0098] Kurtosis is a statistic that measures the sharpness of the data distribution shape. Calculating the kurtosis of the matching driving curve can reflect the shape distribution of the matching driving curve. The matching driving curve studied in this embodiment is the driving route of the target driver on a straight road. Under normal circumstances, the driving route of the vehicle should be approximately a straight line. When the target driver has an offset direction, a deflected curve will appear in the driving route, and then the target driver will fine-tune this offset phenomenon. At this time, a regression curve will appear in the driving route. In this way, the kurtosis of the matching driving curve can reflect the degree of fine-tuning in this case. The larger the kurtosis value, the greater the sharpness of the corresponding matching driving curve, and then it indicates that the target driver has a greater degree of fine-tuning in driving.

[0099] It should be noted that the calculation method of kurtosis is a well-known technology and will not be introduced in detail here. However, it can be understood that in this embodiment, the positive direction of the Y axis is the direction of travel of the vehicle, that is, the matching driving curve is a curved shape that appears along the direction of travel of the vehicle, that is, the positive direction of the Y axis, and then when calculating the kurtosis, it is necessary to calculate the horizontal coordinate value corresponding to each vertical coordinate.

[0100] Each matching behavior curve in the conventional behavior class corresponds to a shape characteristic factor, which characterizes the degree of curvature of the curve in the matching behavior curve and reflects the degree of fine-tuning of the target driver during the corresponding driving.

[0101] Step S402: Perform DTW matching on the predicted driving curve and each matching driving curve respectively to obtain a matching data point corresponding to each data point on the predicted driving curve and each matching driving curve.

[0102] The dynamic time warping algorithm can be used to obtain the matching of the predicted driving curve and the corresponding data points on each matching driving curve, which is convenient for subsequent analysis of the differences between the corresponding matching data points. Each data point on the predicted driving curve has a matching data point on each matching driving curve.

[0103] Step S403: for any matching driving curve, the accumulated sum of the differences between the data points on the predicted driving curve and the corresponding matching data points on the matching driving curve is taken as the feature difference accumulated value.

[0104] It should be noted that the traditional method of calculating the DTW distance is to align and match the data of two different sequences, and then calculate the absolute value of the difference between the matching data to obtain the DTW distance between the two different sequences. In this embodiment, based on the idea of ​​the DTW algorithm, the data difference between the predicted driving curve and each matching driving curve is calculated. The larger the difference value, the more aggressive the behavior habits represented by the prediction result of the target driver are compared with the historical records, and the smaller the difference value, the more inert the behavior habits represented by the prediction result of the target driver are compared with the historical records.

[0105] Furthermore, the characteristic difference accumulation value reflects the behavior difference between a matching driving curve and a predicted driving curve.

[0106] Step S404, determining the adjustment coefficient of the target driver based on the product of the average of the cumulative values ​​of the characteristic differences corresponding to the predicted driving curve and all the matching driving curves, the shape characteristic factor, and the proportion of the total number of all the matching driving curves in the conventional behavior class.

[0107] The final adjustment coefficient is quantified by combining the characteristic analysis results of the matching behavior curve in two aspects in the conventional behavior class and the data difference distribution corresponding to the predicted driving curve and the matching driving curve. In this embodiment, as a specific example, the calculation formula of the adjustment coefficient of the target driver can be expressed as:

[0108]

[0109] in, represents the adjustment coefficient of the target driver, represents the mean of the kurtosis of all matching driving curves in the regular behavior class, Represents the total number of all matching driving curves in the regular behavior class, Represents the total number of all matching driving curves, It represents the mean of the accumulated characteristic differences between the predicted driving curve and all matching driving curves. is a normalization function used to adjust the coefficient value range to [-1,1].

[0110] Indicates the proportion of the total number of all matching driving curves in the conventional behavior class. The more data contained in the conventional behavior class, the greater the kurtosis of the matching driving curve in the conventional behavior class, indicating that the proportion of conventional behavior habits is greater, and the greater the degree of correction of the target driver to the offset situation, the greater the degree of adjustment of the corresponding predicted driving curve, that is, the greater the value of the adjustment coefficient. The larger the mean value of the cumulative value of the characteristic difference, the greater the difference between the prediction result and the historical record, and the more it is necessary to adjust the predicted driving curve in the direction of the path expression in the conventional behavior class, that is, the greater the degree of adjustment, the greater the value of the adjustment coefficient.

[0111] Step S500, adjusting the predicted driving curve according to the reaction time coefficient and the adjustment coefficient, comparing the adjusted result with the normal driving path, and issuing a safety warning to the target driver.

[0112] In this embodiment, the adjustment operation is divided into two aspects. First, the initial reaction stage of the target driver is adjusted according to the reaction degree of the target driver to the initial deviation, so as to simulate the driving path corresponding to the actual reaction of the target driver to the deviation. Second, the prediction form curve is corrected and adjusted according to the fine-tuning operation of the target driver to the deviation, combined with the behavior habits in the historical records.

[0113] Specifically, first, taking the starting data point of the predicted driving curve as the starting point, a fitting behavior segment is obtained based on the first slope of the predicted driving curve, and the length of the fitting behavior segment is equal to the value of the reaction time coefficient. In this embodiment, the driving path of the target driver in the initial reaction stage is approximately simulated as straight-line driving, that is, before the fine-tuning operation, it is assumed that the target driver does not change the driving path.

[0114] Then, the sum of the value 1 and the adjustment coefficient is used as the first proportional coefficient, which can be expressed as , the difference between the value 1 and the adjustment coefficient is used as the second proportional coefficient, which can be expressed as ;in is the first proportionality coefficient, is the second proportionality coefficient, is the adjustment factor.

[0115] Starting from the horizontal coordinate corresponding to the end point of the fitting behavior segment on the predicted driving curve, the vertical coordinate of the predicted driving curve is proportionally adjusted according to the first proportional coefficient, and the horizontal coordinate of the predicted driving curve is proportionally adjusted according to the second proportional coefficient.

[0116] The larger the value of the adjustment coefficient, the greater the degree of adjustment to the normal behavior habits, and the larger the value of the adjustment coefficient, the more aggressive the target driver's normal behavior habits are. When the value of the adjustment coefficient is greater than 0, the proportion of the driver's path along the driving direction is enlarged. The smaller the value of the adjustment coefficient, the more inert the target driver's normal behavior habits are. When the value of the adjustment coefficient is less than 0, the proportion of the driver's path along the driving direction is reduced.

[0117] Finally, the adjusted curve is compared with the normal driving path to issue a safety warning to the target driver.

[0118] Specifically, when there is an intersection between the adjusted result curve and the straight line of the normal driving path, the predicted driving time of the target driver is determined based on the curve length from the adjusted result curve to the intersection and the predicted speed of the target driver, and the normal driving time of the target driver to the intersection under normal circumstances is obtained. If the predicted driving time is equal to the normal driving time, a safety warning is issued to the target driver.

[0119] It should be noted that the predicted speed of the target driver can be obtained through the prediction model, and the driving conditions under normal circumstances can be directly obtained from the historical driving records of the target driver.

[0120] In other embodiments, the adjusted result can more accurately reflect the predicted driving situation of the target driver, and a safety warning can be given to the target driver by judging whether the driver has crossed the line. It is understandable that based on the predicted driving situation of the driver, giving a warning of whether the driver will have dangerous driving behavior in the future is a well-known technology and will not be described in detail here.

[0121] like Figure 4 As shown, an embodiment of the present invention provides a data security warning system for an intelligent networked vehicle, which is used to implement the steps of an application method of the data security warning system for an intelligent networked vehicle, and the system includes:

[0122] A data acquisition module is used to obtain the driving path curve of each road in the vehicle driving record of the target driver, predict the driving state of the target driver, and obtain the predicted driving curve of the target driver;

[0123] A feature analysis module, for screening the driving path curves to obtain matching driving curves according to the inclination of the predicted driving curve and the driving path curve; and for classifying the matching driving curves to obtain conventional behavior classes and special behavior classes according to the difference between the predicted driving curve and each matching driving curve, combined with the data distribution of the predicted driving curve and the data distribution of each matching driving curve;

[0124] A time analysis module, used to obtain the reaction time coefficient of the target driver according to the amount of data contained in the regular behavior class and the special behavior class, and the length of the data distribution of the matching driving curve in the regular behavior class and the special behavior class;

[0125] An adjustment analysis module, for obtaining an adjustment coefficient of a target driver according to the amount of data contained in the conventional behavior class and the shape distribution of the matching driving curve therein, combined with the difference distribution between the predicted driving curve and the matching driving curve in the conventional behavior class;

[0126] The safety warning module is used to adjust the predicted driving curve according to the reaction time coefficient and the adjustment coefficient, and to compare the adjusted result with the driving path under normal conditions to issue a safety warning to the target driver.

[0127] It should be noted that the intelligent connected vehicle data security warning system and the application method embodiment of the intelligent connected vehicle data security warning system provided in the above embodiment belong to the same concept. The specific implementation process is detailed in the method embodiment and will not be repeated here.

[0128] The present application also provides a computer device. Figure 5 , which shows a schematic diagram of the structure of a computer device provided by an embodiment of the present invention, the computer device includes a memory 501, a processor 502, and a computer program 503 stored in the memory 501 and running on the processor 502, wherein when the processor 502 executes the computer program 503, the computer device can execute any application method of the intelligent connected vehicle data security warning system introduced above.

[0129] An embodiment of the present application also provides a computer program product. When the computer program product is run on a computer device, the computer device can execute any one of the application methods of the intelligent connected vehicle data security warning system introduced above.

[0130] An embodiment of the present application also provides a computer-readable storage medium, in which a computer program code is stored. When the computer program code runs on a computer device, the computer device can execute any application method of the intelligent connected vehicle data security warning system introduced above.

[0131] In the embodiments provided in the present application, it should be understood that the provided computer device, computer program product and computer-readable storage medium are all used to execute the corresponding methods provided above. Therefore, the beneficial effects that can be achieved can refer to the beneficial effects in the methods provided above and will not be repeated here.

[0132] The embodiments described above are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. An application method of a data security warning system for an intelligent networked vehicle, characterized in that: The method comprises the following steps: Obtaining the driving path curve of each road in the vehicle driving record of the target driver, predicting the driving state of the target driver, and obtaining the predicted driving curve of the target driver; According to the inclination of the predicted driving curve and the driving path curve, the driving path curve is screened to obtain a matching driving curve; according to the difference between the predicted driving curve and each matching driving curve, combined with the data distribution of the predicted driving curve and the data distribution of each matching driving curve, the matching driving curve is classified to obtain a normal behavior class and a special behavior class; According to the amount of data contained in the regular behavior class and the special behavior class, and the length of the data distribution of the matching driving curves in the regular behavior class and the special behavior class, the reaction time coefficient of the target driver is obtained; According to the amount of data contained in the conventional behavior class and the shape distribution of the matching driving curve therein, combined with the difference distribution between the predicted driving curve and the matching driving curve in the conventional behavior class, the adjustment coefficient of the target driver is obtained; The predicted driving curve is adjusted according to the reaction time coefficient and the adjustment coefficient, and the adjusted result is compared with the driving path under normal conditions to issue a safety warning to the target driver.

2. The application method of the intelligent networked vehicle data security warning system according to claim 1 is characterized in that: The step of screening the driving path curve to obtain a matching driving curve according to the predicted driving curve and the inclination of the driving path curve specifically includes: Respectively obtain a first slope corresponding to the position of the starting data point of the predicted driving curve and a second slope corresponding to the position of the ending data point; obtain a first slope corresponding to the position of the starting data point of each driving path curve and a second slope corresponding to the position of the ending data point; The difference between the predicted driving curve and the first slope corresponding to each driving path curve is recorded as a first difference, and the difference between the predicted driving curve and the second slope corresponding to each driving path curve is recorded as a second difference; When both the first difference and the second difference are smaller than a preset difference threshold, the corresponding driving path curve is recorded as a matching driving curve.

3. The application method of the intelligent networked vehicle data security warning system according to claim 2 is characterized in that: According to the difference between the predicted driving curve and each matching driving curve, combined with the data distribution of the predicted driving curve and the data distribution of each matching driving curve, the matching driving curve is classified to obtain a conventional behavior class and a special behavior class, specifically including: Acquire a first data point on the predicted driving curve where the difference between the slope of the data point and the first slope is greater than a preset difference threshold, and determine an initial reaction segment of the predicted driving curve based on a data distribution between the initial data point and the first data point; According to the difference and the curve length between the predicted driving curve and each matching driving curve, combined with the difference between the initial reaction segment of the predicted driving curve and the initial reaction segment of each matching driving curve, a classification metric distance between the predicted driving curve and each matching driving curve is obtained; According to the classification metric distance, the matching driving curves are classified to obtain a conventional behavior class and a special behavior class.

4. The application method of the intelligent networked vehicle data security warning system according to claim 3 is characterized in that: The method for obtaining the classification metric distance between the predicted driving curve and each matching driving curve is specifically as follows: determining a difference coefficient based on a difference between a curve length of the predicted driving curve and a curve length of each matching driving curve; Determine a first distance coefficient based on a DTW distance between the predicted driving curve and each matching driving curve; determine a second distance coefficient based on a DTW distance between an initial reaction segment of the predicted driving curve and an initial reaction segment of each matching driving curve; According to the difference coefficient, the first distance coefficient and the second distance coefficient, a classification metric distance between the predicted driving curve and each matching driving curve is obtained, and the difference coefficient, the first distance coefficient and the second distance coefficient are all positively correlated with the classification metric distance.

5. The application method of the intelligent networked vehicle data security warning system according to claim 3 is characterized in that: The matching driving curves are classified according to the classification metric distance to obtain conventional behavior classes and special behavior classes, specifically including: Based on the classification metric distance, a clustering algorithm is used to classify the matching driving curves to obtain cluster clusters, and the cluster cluster corresponding to the maximum number of matching driving curves contained in each cluster cluster is taken as a regular behavior class, and all cluster clusters except the regular behavior class in all cluster clusters are taken as special behavior classes.

6. The application method of the intelligent networked vehicle data security warning system according to claim 3 is characterized in that: The reaction time coefficient of the target driver is obtained according to the amount of data included in the conventional behavior class and the special behavior class, and the length of the data distribution of the matching driving curve in the conventional behavior class and the special behavior class, specifically including: Recording the average of the curve lengths of the initial reaction segments of all matching driving curves in the regular behavior class as a first average, and calculating the product of the ratio of the total number of all matching driving curves in the regular behavior class and the first average to obtain a first time coefficient of the regular behavior class; Recording the average of the curve lengths of the initial reaction segments of all matching driving curves in all special behavior classes as a second average, and calculating the product of the ratio of the total number of all matching driving curves in all special behavior classes and the second average to obtain a second time coefficient of the special behavior class; The maximum value of the first time coefficient and the second time coefficient is used as the reaction time coefficient of the target driver.

7. The application method of the intelligent networked vehicle data security warning system according to claim 6 is characterized in that: The adjustment coefficient of the target driver is obtained according to the amount of data contained in the conventional behavior class and the shape distribution of the matching driving curve therein, combined with the difference distribution between the predicted driving curve and the matching driving curve in the conventional behavior class, specifically including: The mean of the kurtosis of all matching driving curves in the conventional behavior class is calculated and recorded as the shape characteristic factor; Perform DTW matching on the predicted driving curve and each matching driving curve respectively, and obtain the matching data point corresponding to each data point on the predicted driving curve and each matching driving curve; For any matching driving curve, the cumulative sum of the differences between the data points on the predicted driving curve and the corresponding matching data points on the matching driving curve is taken as the feature difference cumulative value; The adjustment coefficient of the target driver is determined based on the product of the average of the cumulative values ​​of the characteristic differences corresponding to the predicted driving curve and all the matching driving curves, the shape characteristic factor, and the proportion of the total number of all matching driving curves in the conventional behavior class, and the value of the adjustment coefficient is [-1,1].

8. The application method of the intelligent networked vehicle data security warning system according to claim 7 is characterized in that: The adjusting the predicted driving curve according to the reaction time coefficient and the adjustment coefficient specifically includes: Taking the starting data point of the predicted driving curve as the starting point, obtaining a fitting behavior line segment based on the first slope of the predicted driving curve, wherein the length of the fitting behavior line segment is equal to the value of the reaction time coefficient; The sum of the value 1 and the adjustment coefficient is used as the first proportional coefficient, and the difference between the value 1 and the adjustment coefficient is used as the second proportional coefficient; Starting from the horizontal coordinate corresponding to the end point of the fitting behavior segment on the predicted driving curve, the vertical coordinate of the predicted driving curve is proportionally adjusted according to the first proportional coefficient, and the horizontal coordinate of the predicted driving curve is proportionally adjusted according to the second proportional coefficient.

9. The application method of the intelligent networked vehicle data security warning system according to claim 1 is characterized in that: The step of comparing the adjusted result with the normal driving path and providing a safety warning to the target driver specifically includes: When there is an intersection between the adjusted result curve and the straight line of the normal driving path, the predicted driving time of the target driver is determined based on the curve length from the adjusted result curve to the intersection and the predicted speed of the target driver, and the normal driving time of the target driver to the intersection under normal circumstances is obtained. If the predicted driving time is equal to the normal driving time, a safety warning is issued to the target driver.

10. A data security warning system for intelligent networked vehicles, characterized in that: The system is used to implement the steps of an application method of a smart connected vehicle data security warning system as described in any one of claims 1 to 9, and the system includes: A data acquisition module is used to obtain the driving path curve of each road in the vehicle driving record of the target driver, predict the driving state of the target driver, and obtain the predicted driving curve of the target driver; A feature analysis module, for screening the driving path curves to obtain matching driving curves according to the inclination of the predicted driving curve and the driving path curve; and for classifying the matching driving curves to obtain conventional behavior classes and special behavior classes according to the difference between the predicted driving curve and each matching driving curve, combined with the data distribution of the predicted driving curve and the data distribution of each matching driving curve; A time analysis module, used to obtain the reaction time coefficient of the target driver according to the amount of data contained in the regular behavior class and the special behavior class, and the length of the data distribution of the matching driving curve in the regular behavior class and the special behavior class; An adjustment analysis module, for obtaining an adjustment coefficient of a target driver according to the amount of data contained in the conventional behavior class and the shape distribution of the matching driving curve therein, combined with the difference distribution between the predicted driving curve and the matching driving curve in the conventional behavior class; The safety warning module is used to adjust the predicted driving curve according to the reaction time coefficient and the adjustment coefficient, and to compare the adjusted result with the driving path under normal conditions to issue a safety warning to the target driver.

Images