Driving danger early warning method for driver in intelligent network connection environment

Abstract

The invention discloses a driver driving danger early warning method in an intelligent network connection environment, and the method comprises the steps: obtaining the natural driving data of a driver during each trip under different road types and different traffic flows, and equally dividing each trip into travel segments; calculating characteristic parameters of two bad driving behaviors of overspeed and sudden speed change in each travel segment; determining a bad driving behavior characteristic parameter dynamic safety threshold value, and calculating a bad driving behavior score of the driver; dividing risk levels; obtaining a comprehensive score and a risk level of each travel segment of the driver; and obtaining the comprehensive score and the risk level of the next journey segment of the driver, obtaining the risk early warning intensity, and achieving driving danger early warning. According to the method, more accurate risk assessment is obtained, the problem that driving behaviors have different influence degrees on driving safety is solved, driving danger of a driver is early warned in time, active intervention is carried out, and the traffic safety is improved.

Description

Technical Field

[0001] The invention relates to a method for warning a driver's driving danger in an intelligent network environment, and belongs to the technical field of vehicle safe driving. Background Art

[0002] With the economic growth and the acceleration of urbanization, the living standards of Chinese residents have significantly improved, and the number of motor vehicles has continued to rise. Against this background, the complex traffic environment in my country has gradually highlighted traffic safety issues, and road driving safety has become a key challenge in the field of transportation.

[0003] Drivers are an important part of the traffic system. They are information processors, decision makers, regulators and controllers in the traffic system. According to statistics, 70% of traffic accidents are caused by drivers, and dangerous driving is the main cause of traffic accidents. Therefore, based on natural driving data, the driver's driving risk is scored through the driver's dangerous driving behavior, and the driver's driving risk is warned in time and actively intervened, which is conducive to improving traffic safety and has very important practical significance. Research on driving risk warning through driving behavior often ignores the impact of road grade and traffic flow status on drivers and driving risk. The single identification of dangerous driving behavior and the design of dangerous driving behavior warning system through vehicle operating parameters or driving physiological indicators cannot fully reflect the impact of driver's internal factors, vehicle status and external environmental factors on driving behavior. Summary of the invention

[0004] Purpose of the invention: In view of the shortcomings of the prior art, the present invention provides a method for warning of driving hazards for drivers in an intelligent networked environment. The present invention is based on natural driving data, starts from the driver's driving behavior, considers the influence of road grade and traffic flow, sets dynamic thresholds, and uses the entropy weight-matter-element method to set a driving hazard scoring model with variable weights, comprehensively evaluates the driver's driving hazards, and issues dangerous driving warnings.

[0005] Technical solution: A method for warning driver driving danger in an intelligent network environment, comprising the following steps:

[0006] S1. Obtain the natural driving data of drivers during each trip under different road types and traffic flows, and divide each trip into trip segments;

[0007] S2, calculating characteristic parameters of two bad driving behaviors, speeding and sudden speed change, in each travel segment according to the natural driving data obtained in S1;

[0008] S3. Under different road types and traffic flow conditions, the interquartile range method is used to determine the dynamic safety threshold of the characteristic parameters of bad driving behavior, and the bad driving behavior score of the driver is calculated based on the dynamic safety threshold of the characteristic parameters;

[0009] S4. Use k-means clustering method to divide the degree of exceeding the dynamic safety threshold of characteristic parameters into three risk levels: low-level danger, medium-level danger, and high-level danger;

[0010] S5: Based on the bad driving behavior score obtained in S3 and the risk level divided in S4, the entropy weight-matter element method is used to obtain the weights of each indicator under different traffic flow conditions, and the driver's driving risk evaluation model is obtained, thereby obtaining the comprehensive score and risk level of each driver's trip segment;

[0011] S6: By predicting the speed of the next trip, the traffic flow information and speed of the next trip are substituted into S5 to obtain the comprehensive score and risk level of the driver's next trip segment, and the risk warning intensity is obtained to achieve driving hazard warning.

[0012] Preferably, the S1 is specifically:

[0013] Obtain natural driving data of m trip samples of n drivers under different road types and traffic flow conditions, regard the driver's departure from point A to point B as a trip sample, and divide the m trip samples into x a trip segments, where a=1,2,3...,m, and the traffic flow status on each trip segment is obtained through the network connection system;

[0014] Among them, road types include expressways, first-class highways, second-class highways, third-class highways, fourth-class highways, urban expressways, and urban ordinary roads;

[0015] Traffic flow states include free flow, saturated flow, and congested flow;

[0016] Natural driving data includes speed, acceleration and deceleration.

[0017] Preferably, S2 is specifically:

[0018] The characteristic parameters include speed, speed standard deviation, acceleration, acceleration standard deviation, deceleration, and deceleration standard deviation; the acquisition frequency of the characteristic parameters speed and acceleration and deceleration is set to an interval of 1s, and the speed v at the jth second is obtained. j and acceleration a j , which means acceleration when the direction is the same as the vehicle's motion direction In the opposite direction of the vehicle's movement, it indicates deceleration Calculate the speed standard deviation ε for each travel segment v And the standard deviation of acceleration and deceleration Select it as the value of the speed standard deviation and acceleration standard deviation per second of a travel segment. Get the driving behavior characteristic parameters of each travel segment. The i-th characteristic parameter value of each travel segment at the jth second is defined as r ij , for r ij Normalize it, eliminate the limitation of data unit, and transform it into dimensionless form r′ ij , the formula is as follows:

[0019] When i is a positive index:

[0020]

[0021] When i is a negative index:

[0022]

[0023] The specific calculation of driving behavior characteristic parameters is as follows:

[0024]

[0025] Where: is the mean speed of the travel segment; ε v is the speed standard deviation; is the mean acceleration of the travel segment; is the mean deceleration of the travel segment; is the standard deviation of acceleration; is the standard deviation of deceleration.

[0026] Preferably, S3 is specifically:

[0027] Select the interquartile difference method to divide the safety threshold of bad driving behavior, the safety threshold of the i-th bad driving behavior characteristic parameter for:

[0028] r i * =X i +1.5Y i

[0029] Where: X i is the upper quartile of the sample distribution of the i-th bad driving behavior characteristic parameter; Y i is the interquartile difference, i.e., the difference between the upper quartile and the lower quartile of the sample distribution of the i-th bad driving behavior characteristic parameter;

[0030] Get the dynamic safety threshold function of driving behavior:

[0031]

[0032] Where: h1, h2, h3 represent free flow, saturated flow, and congested flow respectively; θ represents different road types, and θ is 1, 2, 3, 4, 5, 6, and 7, representing expressway, first-class highway, second-class highway, third-class highway, fourth-class highway, urban expressway, and urban ordinary road respectively; represents the safety threshold when the traffic flow state is h on the θth road type; They represent the time periods of the three traffic flow states on the θth road type;

[0033] when When , it means that the driver does not have the i-th bad driving behavior; when When , it means that the driver has the i-th bad driving behavior;

[0034] Calculate the number of bad driving behaviors and r′ of drivers at each moment under different road types and traffic flow conditions ij Exceeding safety threshold The degree of risk accumulation is carried out to quantify the driving risk of drivers under different traffic flow conditions, and the unit time behavior risk score q of each bad driving behavior of the driver is calculated. ij :

[0035]

[0036] The behavioral risk score Q of the driver’s i-th bad driving behavior in a trip segment i for:

[0037]

[0038] Where c represents the first moment when bad driving behavior occurs in a trip segment, and d represents the last moment when bad driving behavior occurs in a trip segment.

[0039] Preferably, the S4 is specifically:

[0040] As the magnitude of the characteristic parameter value exceeding the safety threshold increases, the impact of the i-th bad driving behavior on driving safety increases. The k-means clustering method is used to divide the degree of exceeding the safety threshold into three threshold ranges, representing three risk levels: low-level danger Medium risk Advanced Danger i represents the i-th bad driving behavior. Calculate the distance from each point to the corresponding cluster center and find the minimum distance d from each point to the center in each type of data. min and the maximum distance d max , the middle value of the corresponding point with the maximum distance of the previous category and the corresponding point with the minimum distance of the next category is taken as the upper boundary of the previous category and the lower boundary of the next category.

[0041] Preferably, the S5 comprises:

[0042] S501: Based on the bad driving behavior score obtained in S3, the entropy weight method is used to obtain the comprehensive score of each trip segment of the driver;

[0043] S502: Based on the risk level divided in S4, the matter-element method is used to evaluate the risk of driving behavior in each trip segment.

[0044] Preferably, the S501 is specifically:

[0045] The entropy weight method is used to determine the weights of each indicator under different traffic flow states. The natural driving data is divided into three data sets according to the three traffic flow states of free flow, saturated flow, and congested flow. The entropy weight method is used to calculate the weights of indicators under each traffic flow state. Speed, speed standard deviation, acceleration, acceleration standard deviation, deceleration, and deceleration standard deviation are determined as evaluation indicators.

[0046] Assume that there are e indicators to be evaluated and f evaluation samples in each of the three traffic flow states, forming the original evaluation matrix r h , h represents three traffic flow states, 1, 2, 3:

[0047]

[0048] Where r′ iu Represents the value of the i-th indicator in the u-th sample;

[0049] The steps to obtain the index weights of matrix r using the entropy weight method are as follows:

[0050] Calculate the entropy of the i-th indicator under the h-th traffic flow state

[0051]

[0052] in

[0053]

[0054] In the formula, is the proportion of the index value of the i-th index in the u-th sample;

[0055] Calculate the entropy weight of the i-th indicator under the h-th traffic flow state

[0056]

[0057] The driver's composite score for each short trip is:

[0058]

[0059] Preferably, the S502 is specifically:

[0060] Speed, speed standard deviation, acceleration, acceleration standard deviation, deceleration, and deceleration standard deviation are determined as evaluation indicators;

[0061] Construct a matter-element extension model to determine the risk level of the event to be evaluated:

[0062] Determine the classic domain of each indicator

[0063] In the formula, Indicates that the divided classic domain is in the sth risk level under the hth traffic flow state, G i represents the risk assessment index, Indicates the risk assessment level value;

[0064] Determine the section domain of each indicator

[0065] In the formula, For g in P i The value range of

[0066] Determine the object to be evaluated

[0067] In the formula, N0 is the thing to be evaluated, Indicates the actual value of the evaluation index, g i represents the risk assessment index;

[0068] Determine the relevance of the indicator:

[0069] The distance between point k0 and the finite real interval K = (a, b) is:

[0070]

[0071] Determine the comprehensive relevance of indicators:

[0072]

[0073] Determine the risk level of the object to be evaluated:

[0074] K s0 =maxK s (N0)

[0075] Then it is determined that the matter-element to be evaluated belongs to level j0.

[0076] Preferably, the S6 is specifically:

[0077] In the intelligent network environment, the traffic flow information h of the next trip segment is obtained, and a vehicle speed prediction model is established. The vehicle speed of the next trip segment is predicted by the vehicle speed prediction model to obtain the vehicle speed V of the next trip segment; the traffic flow information h and the vehicle speed V of the next trip segment are substituted into S5 to calculate the driving behavior score and risk level of the driver in the next trip segment, and the risk warning intensity is obtained, and the driver is warned according to different risk warning intensities.

[0078] Beneficial effects: The present invention establishes a dynamic safety threshold function for driving behavior by considering the impact of road grade and traffic flow on driving behavior, and adjusts the evaluation criteria for driving behavior in real time according to the road grade and traffic flow status to obtain a more accurate risk assessment; a personalized driver driving risk evaluation model is established based on the entropy weight-matter element method with variable weights, which solves the problem that driving behavior has different degrees of impact on driving safety under different road grades and traffic flow conditions; through the historical data of the driver's dangerous driving behavior, the driving risk of the next short trip is predicted, the driver is warned of the driving risk in time, and active intervention is carried out, which is conducive to improving traffic safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0079] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0080] Figure 1 It is the principle diagram of the method of the present invention;

[0081] Figure 2 A flow chart is established for the driver driving risk assessment model of the present invention. DETAILED DESCRIPTION

[0082] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0083] In the description of the present invention, it is necessary to understand that the terms "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", etc., indicating the orientation or position relationship are based on the orientation or position relationship shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention.

[0084] In the present invention, unless otherwise clearly specified and limited, a first feature being "above" or "below" a second feature may include that the first and second features are in direct contact, or may include that the first and second features are not in direct contact but are in contact through another feature between them. Moreover, a first feature being "above", "above" and "above" a second feature includes that the first feature is directly above and obliquely above the second feature, or simply indicates that the first feature is higher in level than the second feature. A first feature being "below", "below" and "below" a second feature includes that the first feature is directly below and obliquely below the second feature, or simply indicates that the first feature is lower in level than the second feature.

[0085] like Figure 1 As shown, a method for warning driver driving danger in an intelligent network environment includes the following steps:

[0086] S1. Obtain the natural driving data of drivers during each trip under different road types and traffic flows, and divide each trip into trip segments;

[0087] Obtain natural driving data of m trip samples of n drivers under different road types and traffic flow conditions, regard the driver's departure from point A to point B as a trip sample, and divide the m trip samples into x a 1-minute travel segments, where a = 1, 2, 3, ..., m, and the traffic flow status on each travel segment is obtained through the network connection system;

[0088] Among them, road types include expressways, first-class highways, second-class highways, third-class highways, fourth-class highways, urban expressways, and urban ordinary roads;

[0089] Traffic flow states include free flow, saturated flow, and congested flow;

[0090] Natural driving data includes speed, acceleration and deceleration.

[0091] S2, calculating characteristic parameters of two bad driving behaviors, speeding and sudden speed change, in each travel segment according to the natural driving data obtained in S1;

[0092] The characteristic parameters include speed, speed standard deviation, acceleration, acceleration standard deviation, deceleration, and deceleration standard deviation; the acquisition frequency of the characteristic parameters speed and acceleration and deceleration is set to an interval of 1s, and the speed v at the jth second is obtained. j and acceleration a j , which means acceleration when the direction is the same as the vehicle's motion direction In the opposite direction of the vehicle's movement, it indicates deceleration Calculate the speed standard deviation ε for each travel segment v And the standard deviation of acceleration and deceleration Select it as the value of the speed standard deviation and acceleration standard deviation per second of a travel segment. Get the driving behavior characteristic parameters of each travel segment. The i-th characteristic parameter value of each travel segment at the jth second is defined as r ij , for r ij Normalize it, eliminate the limitation of data unit, and transform it into dimensionless form r′ ij , the formula is as follows:

[0093] When i is a positive index:

[0094]

[0095] When i is a negative index:

[0096]

[0097] The specific calculation of driving behavior characteristic parameters is as follows:

[0098]

[0099] Where: is the mean speed of the travel segment; ε v is the speed standard deviation; is the mean acceleration of the travel segment; is the mean deceleration of the travel segment; is the standard deviation of acceleration; is the standard deviation of deceleration.

[0100] S3. Under different road types and traffic flow conditions, the interquartile range method is used to determine the dynamic safety threshold of the characteristic parameters of bad driving behavior, and the bad driving behavior score of the driver is calculated based on the dynamic safety threshold of the characteristic parameters;

[0101] Since the driving behavior characteristic parameters vary under different road types and traffic flow conditions, it is necessary to first determine the safety thresholds under different road types and traffic flow conditions. The quartile difference method is used to divide the safety thresholds of bad driving behaviors. The safety threshold of the i-th bad driving behavior characteristic parameter is for:

[0102] r i * =X i +1.5Y i

[0103] Where: X i is the upper quartile of the sample distribution of the i-th bad driving behavior characteristic parameter; Y i is the interquartile difference, i.e., the difference between the upper quartile and the lower quartile of the sample distribution of the i-th bad driving behavior characteristic parameter;

[0104] Get the dynamic safety threshold function of driving behavior:

[0105]

[0106] Where: h1, h2, h3 represent free flow, saturated flow, and congested flow respectively; θ represents different road types, and θ is 1, 2, 3, 4, 5, 6, and 7, representing expressway, first-class highway, second-class highway, third-class highway, fourth-class highway, urban expressway, and urban ordinary road respectively; represents the safety threshold when the traffic flow state is h on the θth road type; They represent the time periods of the three traffic flow states on the θth road type;

[0107] Different safety thresholds are determined according to different road types and traffic flow conditions to distinguish bad driving behaviors from normal driving behaviors:

[0108] when When , it means that the driver does not have the i-th bad driving behavior; when When , it indicates that the driver has the i-th bad driving behavior; the higher the amplitude of the characteristic parameter exceeding the threshold, the more times it occurs, indicating a higher degree of badness.

[0109] Calculate the number of bad driving behaviors and r′ of drivers at each moment under different road types and traffic flow conditions ij Exceeding safety threshold The degree of risk accumulation is carried out to quantify the driving risk of drivers under different traffic flow conditions, and the unit time behavior risk score q of each bad driving behavior of the driver is calculated. ij :

[0110]

[0111] The behavioral risk score Q of the driver’s i-th bad driving behavior in a trip segment i for:

[0112]

[0113] Where c represents the first moment when bad driving behavior occurs in a trip segment, and d represents the last moment when bad driving behavior occurs in a trip segment.

[0114] S4. Use k-means clustering method to divide the degree of exceeding the dynamic safety threshold of characteristic parameters into three risk levels: low-level danger, medium-level danger, and high-level danger;

[0115] As the magnitude of the characteristic parameter value exceeding the safety threshold increases, the impact of the i-th bad driving behavior on driving safety increases. The k-means clustering method is used to divide the degree of exceeding the safety threshold into three threshold ranges, representing three risk levels: low-level danger Medium risk Advanced Danger i represents the i-th bad driving behavior. Calculate the distance from each point to the corresponding cluster center and find the minimum distance d from each point to the center in each type of data. min and the maximum distance d max , the middle value of the corresponding point with the maximum distance of the previous category and the corresponding point with the minimum distance of the next category is taken as the upper boundary of the previous category and the lower boundary of the next category.

[0116] like Figure 2 As shown, S5: Based on the bad driving behavior score obtained in S3 and the risk level divided by S4, the entropy weight-matter element method is used to obtain the weights of each indicator under different traffic flow conditions, and the driver's driving risk evaluation model is obtained, so as to obtain the comprehensive score and risk level of each driver's trip segment;

[0117] S501: Based on the bad driving behavior score obtained in S3, the entropy weight method is used to obtain the comprehensive score of each trip segment of the driver;

[0118] The entropy weight method is used to determine the weights of each indicator under different traffic flow states. The natural driving data is divided into three data sets according to the three traffic flow states of free flow, saturated flow, and congested flow. The entropy weight method is used to calculate the weights of indicators under each traffic flow state. Speed, speed standard deviation, acceleration, acceleration standard deviation, deceleration, and deceleration standard deviation are determined as evaluation indicators.

[0119] Assume that there are e indicators to be evaluated and f evaluation samples in each of the three traffic flow states, forming the original evaluation matrix r h , h represents three traffic flow states, 1, 2, 3:

[0120]

[0121] Where r′ iu Represents the value of the i-th indicator in the u-th sample;

[0122] The steps to obtain the index weights of matrix r using the entropy weight method are as follows:

[0123] Calculate the entropy of the i-th indicator under the h-th traffic flow state

[0124]

[0125] in

[0126]

[0127] In the formula, is the proportion of the index value of the i-th index in the u-th sample;

[0128] Calculate the entropy weight of the i-th indicator under the h-th traffic flow state

[0129]

[0130] The driver's composite score for each short trip is:

[0131]

[0132] S502: Based on the risk level divided in S4, the matter-element method is used to evaluate the risk of driving behavior in each trip segment.

[0133] Speed, speed standard deviation, acceleration, acceleration standard deviation, deceleration, and deceleration standard deviation are determined as evaluation indicators;

[0134] Construct a matter-element extension model to determine the risk level of the event to be evaluated:

[0135] Determine the classic domain of each indicator

[0136] In the formula, Indicates that the divided classic domain is in the sth risk level under the hth traffic flow state, G i represents the risk assessment index, Indicates the risk assessment level value;

[0137] Determine the section domain of each indicator

[0138] In the formula, For g in P i The value range of

[0139] Determine the object to be evaluated

[0140] In the formula, N0 is the thing to be evaluated, Indicates the actual value of the evaluation index, g i represents the risk assessment index;

[0141] Determine the relevance of the indicator:

[0142] The distance between point k0 and the finite real interval K = (a, b) is:

[0143]

[0144] Determine the comprehensive relevance of indicators:

[0145]

[0146] Determine the risk level of the object to be evaluated:

[0147] K s0 =maxK s (N0)

[0148] Then it is determined that the matter-element to be evaluated belongs to level j0.

[0149] S6: By predicting the speed of the next trip, the traffic flow information and speed of the next trip are substituted into S5 to obtain the comprehensive score and risk level of the driver's next trip segment, and the risk warning intensity is obtained to achieve driving hazard warning.

[0150] In the intelligent network environment, the traffic flow information h of the next trip segment is obtained, and a vehicle speed prediction model is established. The vehicle speed of the next trip segment is predicted by the vehicle speed prediction model to obtain the vehicle speed V of the next trip segment; the traffic flow information h and the vehicle speed V of the next trip segment are substituted into S5 to calculate the driving behavior score and risk level of the driver in the next trip segment, and the risk warning intensity is obtained, and the driver is warned according to different risk warning intensities.

[0151] The risk warning intensity of the next trip segment is displayed on the vehicle display screen. The risk warning intensity is shown in Table 1:

[0152] Table 1 Driver driving risk warning intensity

[0153]

[0154] The risk warning intensity is divided into 5 levels. When the risk warning intensity is level 1 and level 2, a red signal warning of bad driving behavior will be issued on the vehicle display screen; when the risk warning intensity is level 3, a continuous short vehicle prompt sound will be used to provide a warning; when the risk warning intensity is level 4 and level 5, a continuous and uninterrupted vehicle prompt sound will be used to provide a warning, and the vehicle will be automatically forced to slow down.

[0155] This embodiment uses LSTM long short-term memory network to establish a vehicle speed prediction model, processes the acquired traffic information and vehicle speed information, and divides the processed data into a training set and a test set;

[0156] Build an LSTM vehicle speed prediction model, input the training set into the model for training, and verify the trained model;

[0157] This embodiment uses three gate control units, namely the forget gate, input gate and output gate:

[0158] Forget gate: According to the output of the previous moment, the forget gate uses the sigmoid layer to filter the information that can pass through the cell state. Its expression is:

[0159] f t =σ(W f ·[h t-1 ,x t ]+b f )

[0160] Where: W f ,b f f t The weight matrix and bias of h t-1 is the hidden state at the last moment; x t Enter information for the current moment.

[0161] Input gate: The process of discarding unnecessary information and adding new information includes two steps: first, the input gate uses sigmoid to determine the update value; second, the tanh layer is used to generate new candidate values ​​and add them to obtain the candidate value. Its expression is:

[0162]

[0163] Where: f t ,i t C t-1 and The weight of i t =σ(W i ·[h t-1 ,x t ]+b i ),W i ,b i i t The weight matrix and bias of W c is the weight matrix of tanh, b c is bias; C t-1 and C t are the unit states at the previous moment and the current moment respectively.

[0164] Output gate: The initial output is obtained through the sigmoid layer, and then the value is scaled to between -1 and 1 using tanh, and then multiplied pairwise with the output obtained by sigmoid to obtain the output of the model. Its expression is:

[0165] o t =σ(W o ·[h t-1 ,x t ]+b o )

[0166] h t =o t tanh(C t )

[0167] Where: W o is the weight matrix of the output layer; b o For bias.

[0168] The vehicle speed prediction is performed based on the above vehicle speed prediction model.

[0169] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.

[0170] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for warning driver driving danger in an intelligent network environment, characterized by: The following steps are involved: S1. Obtain the natural driving data of drivers during each trip under different road types and traffic flows, and divide each trip into trip segments; S2, calculating characteristic parameters of two bad driving behaviors, speeding and sudden speed change, in each travel segment according to the natural driving data obtained in S1; S3. Under different road types and traffic flow conditions, the interquartile range method is used to determine the dynamic safety threshold of the characteristic parameter of bad driving behavior, and the bad driving behavior score of the driver is calculated based on the dynamic safety threshold of the characteristic parameter; S4. Use k-means clustering method to divide the degree of exceeding the dynamic safety threshold of characteristic parameters into three risk levels: low-level danger, medium-level danger, and high-level danger; S5: Based on the bad driving behavior score obtained in S3 and the risk level divided in S4, the entropy weight-matter element method is used to obtain the weights of each indicator under different traffic flow conditions, and the driver's driving risk evaluation model is obtained, thereby obtaining the comprehensive score and risk level of each driver's trip segment; S6: By predicting the speed of the next trip, the traffic flow information and speed of the next trip are substituted into S5 to obtain the comprehensive score and risk level of the driver's next trip segment, and the risk warning intensity is obtained to achieve driving hazard warning.

2. The method for warning driver driving danger in an intelligent network environment according to claim 1 is characterized by: The S1 is specifically: Obtain natural driving data of m trip samples of n drivers under different road types and traffic flow conditions, regard the driver's departure from point A to point B as a trip sample, and divide the m trip samples into x a trip segments, where a=1,2,3...,m, and the traffic flow status on each trip segment is obtained through the network connection system; Among them, road types include expressways, first-class highways, second-class highways, third-class highways, fourth-class highways, urban expressways, and urban ordinary roads; Traffic flow states include free flow, saturated flow, and congested flow; Natural driving data includes speed, acceleration and deceleration.

3. The method for warning driver driving danger in an intelligent network environment according to claim 2 is characterized by: The S2 is specifically: The characteristic parameters include speed, speed standard deviation, acceleration, acceleration standard deviation, deceleration, and deceleration standard deviation; the acquisition frequency of the characteristic parameters speed and acceleration and deceleration is set to an interval of 1s, and the speed v at the jth second is obtained. j and acceleration a j , which means acceleration when the direction is the same as the vehicle's motion direction In the opposite direction of the vehicle's movement, it indicates deceleration Calculate the speed standard deviation ε for each travel segment v And the standard deviation of acceleration and deceleration Select it as the value of the speed standard deviation and acceleration standard deviation per second of a travel segment; obtain the driving behavior characteristic parameters of each travel segment, and the i-th characteristic parameter value of each travel segment at the jth second is defined as r ij , for r ij Normalize it, eliminate the limitation of data unit, and transform it into dimensionless form r′ ij , the formula is as follows: When i is a positive index: When i is a negative index: The specific calculation of driving behavior characteristic parameters is as follows: Where: is the mean speed of the travel segment; ε v is the speed standard deviation; is the mean acceleration of the travel segment; is the mean deceleration of the travel segment; is the standard deviation of acceleration; is the standard deviation of deceleration.

4. The method for warning driver driving danger in an intelligent network environment according to claim 3 is characterized by: The S3 is specifically: Select the interquartile difference method to divide the safety threshold of bad driving behavior, the safety threshold of the i-th bad driving behavior characteristic parameter for: Where: X i is the upper quartile of the sample distribution of the i-th bad driving behavior characteristic parameter; Y i is the interquartile difference, i.e., the difference between the upper quartile and the lower quartile of the sample distribution of the i-th bad driving behavior characteristic parameter; Get the dynamic safety threshold function of driving behavior: Where: h1, h2, h3 represent free flow, saturated flow, and congested flow respectively; θ represents different road types, and θ is 1, 2, 3, 4, 5, 6, and 7, representing expressway, first-class highway, second-class highway, third-class highway, fourth-class highway, urban expressway, and urban ordinary road respectively; represents the safety threshold when the traffic flow state is h on the θth road type; They represent the time periods of the three traffic flow states on the θth road type; when When , it means that the driver does not have the i-th bad driving behavior; when When , it means that the driver has the i-th bad driving behavior; Calculate the number of bad driving behaviors and r′ of drivers at each moment under different road types and traffic flow conditions ij Exceeding safety threshold The degree of risk accumulation is carried out to quantify the driving risk of drivers under different traffic flow conditions, and the unit time behavior risk score q of each bad driving behavior of the driver is calculated. ij : The behavioral risk score Q of the driver’s i-th bad driving behavior in a trip segment i for: Where c represents the first moment when bad driving behavior occurs in a trip segment, and d represents the last moment when bad driving behavior occurs in a trip segment.

5. The method for warning driver driving danger in an intelligent network environment according to claim 4 is characterized in that: The S4 is specifically: As the magnitude of the characteristic parameter value exceeding the safety threshold increases, the impact of the i-th bad driving behavior on driving safety increases. The k-means clustering method is used to divide the degree of exceeding the safety threshold into three threshold ranges, representing three risk levels: low-level danger Medium risk Advanced Danger i represents the i-th bad driving behavior. Calculate the distance from each point to the corresponding cluster center and find the minimum distance d from each point to the center in each type of data. min and the maximum distance d max , the middle value of the corresponding point with the maximum distance of the previous category and the corresponding point with the minimum distance of the next category is taken as the upper boundary of the previous category and the lower boundary of the next category.

6. The method for warning driver driving danger in an intelligent network environment according to claim 5 is characterized by: The S5 includes: S501: Based on the bad driving behavior score obtained in S3, the entropy weight method is used to obtain the comprehensive score of each trip segment of the driver; S502: Based on the risk level divided in S4, the matter-element method is used to evaluate the risk of driving behavior in each trip segment.

7. The method for warning driver driving danger in an intelligent network environment according to claim 6, characterized in that: The S501 is specifically as follows: The entropy weight method is used to determine the weights of each indicator under different traffic flow states. The natural driving data is divided into three data sets according to the three traffic flow states of free flow, saturated flow, and congested flow. The entropy weight method is used to calculate the weights of indicators under each traffic flow state. Speed, speed standard deviation, acceleration, acceleration standard deviation, deceleration, and deceleration standard deviation are determined as evaluation indicators. Assume that there are e indicators to be evaluated and f evaluation samples in each of the three traffic flow states, forming the original evaluation matrix r h , h represents three traffic flow states, 1, 2, 3: Where r′ iu Represents the value of the i-th indicator in the u-th sample; The steps to obtain the index weights of matrix r using the entropy weight method are as follows: Calculate the entropy of the i-th indicator under the h-th traffic flow state in In the formula, is the proportion of the index value of the i-th index in the u-th sample; Calculate the entropy weight of the i-th indicator under the h-th traffic flow state The driver's composite score for each short trip is:

8. The method for warning driver driving danger in an intelligent network environment according to claim 6, characterized in that: The S502 is specifically as follows: Speed, speed standard deviation, acceleration, acceleration standard deviation, deceleration, and deceleration standard deviation are determined as evaluation indicators; Construct a matter-element extension model to determine the risk level of the event to be evaluated: Determine the classic domain of each indicator In the formula, Indicates that the divided classic domain is in the sth risk level under the hth traffic flow state, G i represents the risk assessment index, Indicates the risk assessment level value; Determine the section domain of each indicator In the formula, For g in P i The value range of Determine the object to be evaluated In the formula, N0 is the thing to be evaluated, Indicates the actual value of the evaluation index, g i represents the risk assessment index; Determine the relevance of the indicator: The distance between point k0 and the finite real interval K = (a, b) is: Determine the comprehensive relevance of indicators: Determine the risk level of the object to be evaluated: K s0 =maxK s (N0) Then it is determined that the matter-element to be evaluated belongs to level j0.

9. The method for warning driver driving danger in an intelligent network environment according to claim 1, characterized in that: The S6 is specifically: In the intelligent network environment, the traffic flow information h of the next trip segment is obtained, and a vehicle speed prediction model is established. The vehicle speed of the next trip segment is predicted by the vehicle speed prediction model to obtain the vehicle speed V of the next trip segment; the traffic flow information h and the vehicle speed V of the next trip segment are substituted into S5 to calculate the driving behavior score and risk level of the driver in the next trip segment, and the risk warning intensity is obtained, and the driver is warned according to different risk warning intensities.

Images