Driving behavior spectrum construction method in auxiliary driving environment

Abstract

The invention discloses a method for constructing a driving behavior spectrum in an auxiliary driving environment, and the method comprises the following steps: collecting natural driving data: collecting the natural driving data of a plurality of driving behaviors when a driver drives a self-adaptive cruise control system vehicle, the natural driving data comprises driver behavior information and vehicle operation and state parameter information; acquiring dangerous driving behavior spectrum characteristic parameters; calculating a dangerous driving behavior spectrum characteristic value G of the driving behavior; clustering the dangerous driving behavior spectrum characteristic values of the plurality of driving behaviors; identifying the driving behaviors of each category after clustering to obtain a driving behavior spectrum in an auxiliary driving environment; the driving behavior of the driver is judged through the driving behavior spectrum construction method in the auxiliary driving environment, and the danger degree of the driving behavior is confirmed.

Description

Technical Field

[0001] The present invention relates to the field of adaptive cruise control, and in particular to the technical field of a method for constructing a driving behavior spectrum in an assisted driving environment. Background Art

[0002] Adaptive cruise control is now widely used in society. In order to improve the safety performance of drivers when driving vehicles with adaptive cruise control systems, the driver's driving behavior is evaluated and the degree of danger of the driving behavior is determined;

[0003] Therefore, how to invent a method for constructing a driving behavior spectrum in an assisted driving environment, evaluate the driver's driving behavior through the driving behavior spectrum construction method in an assisted driving environment, and confirm the degree of danger of the driving behavior has become a difficult problem that needs to be solved urgently in this field. Summary of the Invention

[0004] In order to solve the above technical problems, a driving behavior spectrum construction method in an assisted driving environment is provided, which can be used to evaluate the driver's driving behavior and confirm the degree of danger of the driving behavior.

[0005] According to the present invention, a method for constructing a driving behavior spectrum in an assisted driving environment is provided, which is characterized by comprising the following steps:

[0006] Natural driving data collection, including collecting natural driving data of a number of driving behaviors of a driver when driving a vehicle with an adaptive cruise control system, the natural driving data including driver behavior information and vehicle operation and state parameter information;

[0007] Obtaining characteristic parameters of dangerous driving behavior spectrum;

[0008] Calculate the dangerous driving behavior spectrum characteristic value G of the driving behavior;

[0009] Clustering the dangerous driving behavior spectrum feature values ​​of several driving behaviors;

[0010] The clustered driving behaviors are labeled to obtain the driving behavior spectrum in the assisted driving environment;

[0011] By comparing the driving behavior spectrum in the assisted driving environment with the characteristic parameters of the dangerous driving behavior spectrum of the driving behavior to be determined, the dangerous category of the driving behavior to be determined is obtained.

[0012] Compared with the existing technology, the present invention has the following advantages: by collecting natural driving data including driver behavior information and vehicle operation and status parameter information, it is conducive to making the driving behavior spectrum in the assisted driving environment obtained by analyzing natural driving data closer to reality, thereby facilitating more accurate judgment of the degree of driving behavior risk;

[0013] Calculate the dangerous driving behavior spectrum characteristic value G of the driving behavior through the dangerous driving behavior spectrum characteristic parameter, thereby clustering the driving behavior through the dangerous driving behavior spectrum characteristic value G and dividing the driving behavior into several categories;

[0014] By identifying the clustered driving behaviors, we can obtain the driving behavior spectrum in the assisted driving environment, and thus obtain the range of dangerousness of different types of driving behaviors.

[0015] By comparing the driving behavior spectrum in the assisted driving environment with the characteristic parameters of the dangerous driving behavior spectrum of the driving behavior to be determined, the natural driving data of the driving behavior to be determined is converted into the spectral characteristic parameters of the driving behavior to be determined, and the dangerous driving behavior spectrum characteristic value G of the driving behavior to be determined is obtained through the spectral characteristic parameters of the driving behavior to be determined. Thus, the dangerous category of the driving behavior to be determined is obtained by comparing the dangerous driving behavior spectrum characteristic value G of the driving behavior to be determined with the driving behavior spectrum in the assisted driving environment;

[0016] Thus, the driving behavior of the driver can be evaluated by the driving behavior spectrum construction method in the assisted driving environment to confirm the dangerousness of the driving behavior; and the dangerous category determination of the driving behavior to be determined is closer to reality, which is conducive to achieving more accurate judgment of the dangerousness of the driving behavior.

[0017] Furthermore, the natural driving data is preprocessed to obtain feature information;

[0018] The characteristic information includes vehicle acceleration a, steering wheel angle θ, vehicle lateral offset L, longitudinal speed v1 of the vehicle, longitudinal speed v2 of the vehicle in front of the target lane, longitudinal speed v3 of the vehicle behind the target lane, longitudinal speed v1′ of the vehicle in front of the target lane, distance D2 between the vehicle in front of the target lane, distance D3 between the vehicle in front of the target lane, and distance D1 between the vehicle in front of the target lane.

[0019] Preferably, the driver behavior information includes: steering wheel angle, accelerator pedal opening and closing degree;

[0020] The vehicle operation and status parameter information includes: vehicle speed, acceleration, lateral offset, and distance to the front and rear vehicles;

[0021] The pretreatment method is:

[0022] ① First, align the driver behavior information data and vehicle operation and status parameter information data on the time axis to achieve data time synchronization;

[0023] ② Eliminate outliers from the data, fill in missing values, and finally standardize the data;

[0024] ③ Uniformly set the time window length, perform feature extraction according to the appropriate time window, and obtain feature information.

[0025] Furthermore, the dangerous driving behavior spectrum characteristic parameters include a sudden speed change characteristic parameter K1, an emergency steering characteristic parameter K2, a zigzag driving characteristic parameter K3, a dangerous car-following characteristic parameter K4 and a poor lane-changing characteristic parameter K5.

[0026] The beneficial effect of adopting the previous step is that it can take into account the dangerous conditions of sudden speed changes, emergency steering, Z-shaped lane changes, dangerous following and poor lane changes in driving behaviors, so that the driving behavior spectrum obtained in the assisted driving environment can be closer to reality when classifying driving behaviors of different degrees of danger.

[0027] Furthermore, the rapid speed change characteristic parameter K1 is calculated by the vehicle acceleration a according to Formula 1;

[0028] Formula 1 is:

[0029] K1(t) is the rapid speed change characteristic parameter K1 at time t; a(t) is the acceleration of the vehicle at time t; a(t-△t) is the acceleration of the vehicle at time t-△t; △t is the time interval.

[0030] The beneficial effect of adopting the previous step is that the rapid speed change characteristic parameter K1 is obtained, and the magnitude of K1 can be used to characterize the degree of danger during speed change. The larger the K1 value, the greater the degree of danger of rapid speed change.

[0031] Furthermore, the emergency steering characteristic parameter K2 is calculated by the steering wheel angle θ according to Formula 2;

[0032] Formula 2 is:

[0033] K2(t) is the emergency steering characteristic parameter K2 at time t; θ(t) is the steering wheel angle at time t; θ(t- △ t) is t- △ Steering wheel angle at time t; △ t is the time interval.

[0034] The beneficial effect of adopting the previous step is that the emergency steering characteristic parameter K2 is obtained, and the size of K2 can be used to represent the degree of danger during sudden steering. The larger the K2 value, the greater the degree of danger of sudden speed change.

[0035] Furthermore, the Z-shaped driving characteristic parameter K3 is calculated by the vehicle lateral offset L according to Formula 3;

[0036] Formula 3 is:

[0037] K3(t) is the Z-shaped driving characteristic parameter K3 at time t; Σ|L(t)| is the cumulative value of the vehicle's lateral offset distance per unit time starting from time t; d(t) is the cumulative longitudinal driving distance per unit time starting from time t; F is the data recording frequency.

[0038] The beneficial effect of adopting the previous step is that the Z-shaped driving characteristic parameter K3 is obtained, and the size of K3 can represent the degree of danger during Z-shaped driving. The larger the K3 value, the greater the degree of danger of sudden speed change.

[0039] Furthermore, the dangerous car-following characteristic parameter K4 is calculated according to Formula 4 using the longitudinal velocity v1 of the vehicle, the longitudinal velocity v1′ of the preceding vehicle, and the distance D1 between the vehicle and the preceding vehicle.

[0040] Formula 4 is:

[0041] K4(t) is the characteristic parameter of dangerous car-following at time t; v1(t) is the longitudinal velocity of the ego vehicle at time t; v1′(t) is the longitudinal velocity of the leading vehicle at time t; and D1(t) is the distance between the ego vehicle and the leading vehicle at time t.

[0042] The beneficial effect of adopting the previous step is that the dangerous car-following characteristic parameter K4 is obtained, and the size of K4 can represent the degree of danger during dangerous car-following. The larger the K4 value, the greater the degree of danger of sudden speed change.

[0043] Furthermore, the poor lane change characteristic parameter K5 is obtained according to Formula 5 using the longitudinal speed v1 of the ego vehicle, the longitudinal speed v2 of the vehicle in front of the target lane, the longitudinal speed v3 of the vehicle behind the target lane, the distance D2 between the ego vehicle and the vehicle in front of the target lane, and the distance D3 between the ego vehicle and the vehicle behind the target lane.

[0044] Formula 5 is:

[0045] K5(t) is the characteristic parameter of the poor lane change at time t; v1(t) is the longitudinal velocity of the ego vehicle at time t; v2 is the longitudinal velocity of the preceding vehicle in the target lane, v3 is the longitudinal velocity of the following vehicle in the target lane, D2 is the distance between the ego vehicle and the preceding vehicle in the target lane, and D3 is the distance between the ego vehicle and the following vehicle in the target lane.

[0046] The beneficial effect of adopting the previous step is that the characteristic parameter K5 of the bad lane change is obtained, and the value of K5 can be used to represent the degree of danger during the bad lane change. The larger the K5 value, the greater the danger of the sudden speed change.

[0047] Furthermore, the process of calculating the dangerous driving behavior spectrum characteristic value G of the driving behavior includes the following steps:

[0048] The dangerous driving behavior characteristic parameter threshold is calculated according to Formula 6 using the dangerous driving behavior characteristic parameter.

[0049] Formula 6 is:

[0050] Q i is the upper quartile of the characteristic parameter value of the i-th dangerous driving behavior, I i is the quartile difference, that is, the difference between the upper quartile and the lower quartile of the characteristic parameter value of the i-th dangerous driving behavior.

[0051] According to formula 7, the instantaneous risk score value S corresponding to each dangerous driving behavior characteristic parameter at time t is calculated i (t);

[0052] Formula 7 is:

[0053] S i (t) is normalized by formula 8 to obtain S i (t) corresponding to A Ni ;

[0054] Formula 8 is:

[0055] A i is the average risk score value corresponding to each dangerous driving behavior characteristic parameter of driving behavior; A i By S i (t) is averaged after accumulation;

[0056] According to A Ni The dangerous driving behavior spectrum characteristic value G of the driving behavior is calculated by formula 9;

[0057] Formula nine is:

[0058] W i is the weight of the characteristic parameter of the i-th dangerous driving behavior.

[0059] The beneficial effect of adopting the previous step is that the dangerous driving behavior spectrum characteristic value G of the driving behavior is obtained through Formula 6, Formula 7, Formula 8, and Formula 9.

[0060] Furthermore, the W i Calculation methods include:

[0061] Calculate p by formula 10 ij ;

[0062] Formula 10 is:

[0063]

[0064] p ji A is the characteristic parameter of the i-th dangerous driving behavior of the j-th driving behavior sample Ni , accounting for the characteristic parameters of the ith dangerous driving behavior of all driving behavior samples A Ni and the proportion;

[0065] The entropy value E of the characteristic parameter of the i-th dangerous driving behavior is calculated by formula 11: i ;

[0066] Formula 11 is:

[0067] The weight W of the characteristic parameter of the i-th dangerous driving behavior is calculated by formula 12: i ;

[0068] Formula 11 is:

[0069] and / or

[0070] For several driving behaviors whose dangerous driving behavior spectrum characteristic value G is not 0, the dangerous driving behavior spectrum characteristic value G is clustered using the K-means clustering method, and the several driving behaviors are divided into four categories;

[0071] The dangerous driving behavior spectrum characteristic values ​​G of several driving behaviors included in each type of driving behavior are averaged to obtain the dangerous driving behavior spectrum characteristic value G of each type of driving behavior. j ';

[0072] The dangerous driving behavior spectrum characteristic value G of each type of driving behavior j 'Sort by size, from largest to smallest G j The corresponding different categories of driving behaviors are marked as high-risk driving behavior, medium-risk driving behavior, low-risk driving behavior, and safe driving behavior;

[0073] The maximum deviation value is removed from the dangerous driving behavior spectrum characteristic value G of each driving behavior included in the safe driving behavior, and the range of the remaining dangerous driving behavior spectrum characteristic value G is the safe driving behavior range;

[0074] Remove the maximum deviation value from the dangerous driving behavior spectrum characteristic value G of each driving behavior included in the low-risk driving behavior, and use the maximum value of the remaining dangerous driving behavior spectrum characteristic value G as the maximum value of the low-risk driving behavior range, and use the maximum G value of the safe driving behavior range as the maximum value of the low-risk driving behavior range;

[0075] Remove the maximum deviation value from the dangerous driving behavior spectrum characteristic value G of each driving behavior included in the medium-risk driving behavior, and use the maximum value of the remaining dangerous driving behavior spectrum characteristic value G as the maximum value of the low-risk driving behavior range, and use the maximum G value of the low-risk driving behavior range as the maximum value of the medium-risk driving behavior range;

[0076] The maximum deviation value is removed from the dangerous driving behavior spectrum characteristic value G of each driving behavior included in the high-risk driving behavior, and the maximum value of the remaining dangerous driving behavior spectrum characteristic value G is taken as the maximum value of the low-risk driving behavior range, and the maximum G value of the medium-risk driving behavior range is taken as the maximum value of the high-risk driving behavior range.

[0077] The beneficial effect of adopting the previous step is that by classifying according to the driving behavior spectrum characteristic value G, driving behavior categories with different driving risk levels are obtained, driving behavior categories with different driving risk levels are identified, and driving behaviors are divided into safe driving behaviors, low-risk driving behaviors, medium-risk driving behaviors, and high-risk driving behaviors; the driver's driving behavior to be determined is evaluated and the risk level of the driving behavior to be determined is confirmed. DETAILED DESCRIPTION

[0078] In order to better understand the technical solution of the present invention, the present invention is further described below in conjunction with specific embodiments.

[0079] Example 1:

[0080] This embodiment provides a method for constructing a driving behavior spectrum in an assisted driving environment, comprising the following steps:

[0081] Natural driving data collection, including collecting natural driving data of a number of driving behaviors of a driver when driving a vehicle with an adaptive cruise control system, the natural driving data including driver behavior information and vehicle operation and state parameter information;

[0082] The driver behavior information includes: steering wheel angle, accelerator pedal opening and closing degree;

[0083] The vehicle operation and status parameter information includes: vehicle speed, acceleration, lateral offset, and distance to the front and rear vehicles;

[0084] The natural driving data is preprocessed to obtain characteristic information; the preprocessing method is:

[0085] ① First, align the driver behavior information data and vehicle operation and status parameter information data on the time axis to achieve data time synchronization;

[0086] ② Eliminate outliers from the data, fill in missing values, and finally standardize the data;

[0087] ③ Uniformly set the time window length, perform feature extraction according to the appropriate time window, and obtain feature information.

[0088] The characteristic information includes vehicle acceleration a, steering wheel angle θ, vehicle lateral offset L, longitudinal speed v1 of the vehicle, longitudinal speed v2 of the vehicle in front of the target lane, longitudinal speed v3 of the vehicle behind the target lane, longitudinal speed of the vehicle in front, distance D2 between the vehicle in front of the target lane, distance D3 between the vehicle in front of the target lane, and distance D1 between the vehicle in front of the vehicle.

[0089] The method for constructing a driving behavior spectrum in an assisted driving environment includes: obtaining characteristic parameters of a dangerous driving behavior spectrum; calculating a characteristic value G of the dangerous driving behavior spectrum of a driving behavior; clustering the characteristic values ​​of the dangerous driving behavior spectrum of a plurality of driving behaviors; and labeling each category of clustered driving behaviors to obtain a driving behavior spectrum in an assisted driving environment.

[0090] The characteristic parameters of the dangerous driving behavior spectrum include a sudden speed change characteristic parameter K1, an emergency steering characteristic parameter K2, a Z-shaped driving characteristic parameter K3, a dangerous car-following characteristic parameter K4 and a bad lane-changing characteristic parameter K5.

[0091] The rapid speed change characteristic parameter K1 is calculated by the vehicle acceleration a according to Formula 1;

[0092] Formula 1 is:

[0093] K1(t) is the rapid speed change characteristic parameter K1 at time t; a(t) is the acceleration of the vehicle at time t; a(t-△t) is the acceleration of the vehicle at time t-△t

[0094] Spend; △ t is the time interval.

[0095] The Z-shaped driving characteristic parameter K3 is calculated using the vehicle lateral offset L according to Formula 3.

[0096] The emergency steering characteristic parameter K2 is calculated by the steering wheel angle θ according to formula 2;

[0097] Formula 2 is:

[0098] K2(t) is the emergency steering characteristic parameter K2 at time t; θ(t) is the steering wheel angle at time t; θ(t-△ t) is t- △ Steering wheel angle at time t; △ t is the time interval.

[0099] The Z-shaped driving characteristic parameter K3 is calculated by the vehicle lateral offset L according to Formula 3;

[0100] Formula 3 is:

[0101] K3(t) is the Z-shaped driving characteristic parameter K3 at time t; ∑|L(t)| is the cumulative value of the vehicle's lateral offset distance per unit time starting from time t; d(t) is the cumulative longitudinal driving distance per unit time starting from time t; F is the data recording frequency.

[0102] The dangerous car-following characteristic parameter K4 is calculated using the longitudinal velocity v1 of the vehicle, the longitudinal velocity v1′ of the preceding vehicle, and the distance D1 between the vehicle and the preceding vehicle according to Formula 4.

[0103] Formula 4 is:

[0104] K4(t) is the characteristic parameter of dangerous car-following at time t; v1(t) is the longitudinal velocity of the ego vehicle at time t; v1′(t) is the longitudinal velocity of the leading vehicle at time t; and D1(t) is the distance between the ego vehicle and the leading vehicle at time t.

[0105] The bad lane change characteristic parameter K5 is obtained according to Formula 5 using the longitudinal speed v1 of the ego vehicle, the longitudinal speed v2 of the vehicle in front of the target lane, the longitudinal speed v3 of the vehicle behind the target lane, the distance D2 between the ego vehicle and the vehicle in front of the target lane, and the distance D3 between the ego vehicle and the vehicle behind the target lane.

[0106] Formula 5 is:

[0107] K5(t) is the characteristic parameter of the poor lane change at time t; v1(t) is the longitudinal velocity of the ego vehicle at time t; v2 is the longitudinal velocity of the preceding vehicle in the target lane, v3 is the longitudinal velocity of the following vehicle in the target lane, D2 is the distance between the ego vehicle and the preceding vehicle in the target lane, and D3 is the distance between the ego vehicle and the following vehicle in the target lane.

[0108] The process of calculating the dangerous driving behavior spectrum characteristic value G of driving behavior includes:

[0109] Next steps:

[0110] The dangerous driving behavior characteristic parameter threshold is calculated according to Formula 6 using the dangerous driving behavior characteristic parameter.

[0111] Formula 6 is:

[0112] Q iis the upper quartile of the characteristic parameter value of the i-th dangerous driving behavior, I i is the quartile difference, that is, the difference between the upper quartile and the lower quartile of the characteristic parameter value of the i-th dangerous driving behavior.

[0113] According to formula 7, the instantaneous risk score value S corresponding to each dangerous driving behavior characteristic parameter at time t is calculated i (t);

[0114] Formula 7 is:

[0115] S i (t) is normalized by formula 8 to obtain S i (t) corresponding to A Ni ;

[0116] Formula 8 is:

[0117] A i is the average risk score value corresponding to each dangerous driving behavior characteristic parameter of driving behavior; A i By S i (t) is averaged after accumulation;

[0118] According to A Ni The dangerous driving behavior spectrum characteristic value G of the driving behavior is calculated by formula 9;

[0119] Formula nine is:

[0120] W i is the weight of the characteristic parameter of the i-th dangerous driving behavior.

[0121] The W i Calculation methods include:

[0122] Calculate p by formula 10 ij ;

[0123] Formula 10 is:

[0124]

[0125] p ij A is the characteristic parameter of the i-th dangerous driving behavior of the j-th driving behavior sample Ni , accounting for the characteristic parameters of the ith dangerous driving behavior of all driving behavior samples A Ni and the proportion;

[0126] The entropy value E of the characteristic parameter of the i-th dangerous driving behavior is calculated by formula 11: i ;

[0127] Formula 11 is:

[0128] The weight W of the characteristic parameter of the i-th dangerous driving behavior is calculated by formula 12: i ;

[0129] Formula 11 is:

[0130] For several driving behaviors whose dangerous driving behavior spectrum characteristic value G is not 0, the dangerous driving behavior spectrum characteristic value G is clustered using the K-means clustering method, and the several driving behaviors are divided into four categories;

[0131] The dangerous driving behavior spectrum characteristic values ​​G of several driving behaviors included in each type of driving behavior are averaged to obtain the dangerous driving behavior spectrum characteristic value G of each type of driving behavior. j ';

[0132] The dangerous driving behavior spectrum characteristic value G of each type of driving behavior j 'Sort by size, from largest to smallest G j The corresponding different categories of driving behaviors are marked as high-risk driving behavior, medium-risk driving behavior, low-risk driving behavior, and safe driving behavior;

[0133] The maximum deviation value is removed from the dangerous driving behavior spectrum characteristic value G of each driving behavior included in the safe driving behavior, and the range of the remaining dangerous driving behavior spectrum characteristic value G is the safe driving behavior range;

[0134] Remove the maximum deviation value from the dangerous driving behavior spectrum characteristic value G of each driving behavior included in the low-risk driving behavior, and use the maximum value of the remaining dangerous driving behavior spectrum characteristic value G as the maximum value of the low-risk driving behavior range, and use the maximum G value of the safe driving behavior range as the maximum value of the low-risk driving behavior range;

[0135] Remove the maximum deviation value from the dangerous driving behavior spectrum characteristic value G of each driving behavior included in the medium-risk driving behavior, and use the maximum value of the remaining dangerous driving behavior spectrum characteristic value G as the maximum value of the low-risk driving behavior range, and use the maximum G value of the low-risk driving behavior range as the maximum value of the medium-risk driving behavior range;

[0136] The maximum deviation value is removed from the dangerous driving behavior spectrum characteristic value G of each driving behavior included in the high-risk driving behavior, and the maximum value of the remaining dangerous driving behavior spectrum characteristic value G is taken as the maximum value of the low-risk driving behavior range, and the maximum G value of the medium-risk driving behavior range is taken as the maximum value of the high-risk driving behavior range.

[0137] By comparing the driving behavior spectrum in the assisted driving environment with the characteristic parameters of the dangerous driving behavior spectrum of the driving behavior to be determined, the dangerous category of the driving behavior to be determined is obtained;

[0138] The natural driving data of the driving behavior to be determined is converted into the spectral characteristic parameters of the driving behavior to be determined, and the dangerous driving behavior spectrum characteristic value G of the driving behavior to be determined is obtained through the spectral characteristic parameters of the driving behavior to be determined. Then, the dangerous category of the driving behavior to be determined is obtained by comparing the dangerous driving behavior spectrum characteristic value G of the driving behavior to be determined with the driving behavior spectrum in the assisted driving environment.

[0139] The above description is merely a preferred embodiment of the present application and an illustration of the technical principles employed. Those skilled in the art should understand that the scope of the invention herein is not limited to the technical solutions formed by the specific combination of the above-mentioned technical features, but also encompasses other technical solutions formed by any combination of the above-mentioned technical features or their equivalents without departing from the inventive concept. For example, a technical solution formed by replacing the above-mentioned features with (but not limited to) technical features having similar functions disclosed in this application.

Claims

1. A method for constructing a driving behavior spectrum in an assisted driving environment, characterized in that: The following steps are involved: Natural driving data collection, including collecting natural driving data of a number of driving behaviors of a driver when driving a vehicle with an adaptive cruise control system, the natural driving data including driver behavior information and vehicle operation and state parameter information; Obtaining characteristic parameters of dangerous driving behavior spectrum; Calculate the dangerous driving behavior spectrum characteristic value G of the driving behavior; Clustering the dangerous driving behavior spectrum feature values ​​of several driving behaviors; The clustered driving behaviors are labeled to obtain the driving behavior spectrum in the assisted driving environment; By comparing the driving behavior spectrum in the assisted driving environment with the characteristic parameters of the dangerous driving behavior spectrum of the driving behavior to be determined, the dangerous category of the driving behavior to be determined is obtained.

2. The method for constructing a driving behavior spectrum in an assisted driving environment according to claim 1, characterized in that: Preprocessing the natural driving data to obtain feature information; The characteristic information includes vehicle acceleration a, steering wheel angle θ, vehicle lateral offset L, longitudinal speed v1 of the vehicle, longitudinal speed v2 of the vehicle in front of the target lane, longitudinal speed v3 of the vehicle behind the target lane, longitudinal speed v1′ of the vehicle in front of the target lane, distance D2 between the vehicle in front of the target lane, distance D3 between the vehicle in front of the target lane, and distance D1 between the vehicle in front of the target lane.

3. The method for constructing a driving behavior spectrum in an assisted driving environment according to claim 2, characterized in that: The characteristic parameters of the dangerous driving behavior spectrum include a sudden speed change characteristic parameter K1, an emergency steering characteristic parameter K2, a Z-shaped driving characteristic parameter K3, a dangerous car-following characteristic parameter K4 and a bad lane-changing characteristic parameter K5.

4. The method for constructing a driving behavior spectrum in an assisted driving environment according to claim 3, characterized in that: The rapid speed change characteristic parameter K1 is calculated by the vehicle acceleration a according to Formula 1; Formula 1 is: K1(t) is the rapid speed change characteristic parameter K1 at time t; a(t) is the acceleration of the vehicle at time t; a(t-△t) is the acceleration of the vehicle at time t-△t; △t is the time interval.

5. The method for constructing a driving behavior spectrum in an assisted driving environment according to claim 3, characterized in that: The emergency steering characteristic parameter K2 is calculated by the steering wheel angle θ according to formula 2; Formula 2 is: K2(t) is the emergency steering characteristic parameter K2 at time t; θ(t) is the steering wheel angle at time t; θ(t-△t) is the steering wheel angle at time t-△t; △t is the time interval.

6. The method for constructing a driving behavior spectrum in an assisted driving environment according to claim 3, characterized in that: The Z-shaped driving characteristic parameter K3 is calculated by the vehicle lateral offset L according to Formula 3; Formula 3 is: K3(t) is the Z-shaped driving characteristic parameter K3 at time t; ∑|L(t)| is the cumulative value of the vehicle's lateral offset distance per unit time starting from time t; d(t) is the cumulative longitudinal driving distance per unit time starting from time t; F is the data recording frequency.

7. The method for constructing a driving behavior spectrum in an assisted driving environment according to claim 3, characterized in that: The dangerous car-following characteristic parameter K4 is calculated using the longitudinal velocity v1 of the vehicle, the longitudinal velocity v1′ of the preceding vehicle, and the distance D1 between the vehicle and the preceding vehicle according to Formula 4. Formula 4 is: K4(t) is the characteristic parameter of dangerous car-following at time t; v1(t) is the longitudinal velocity of the ego vehicle at time t; v1′(t) is the longitudinal velocity of the leading vehicle at time t; and D1(t) is the distance between the ego vehicle and the leading vehicle at time t.

8. The method for constructing a driving behavior spectrum in an assisted driving environment according to claim 3, characterized in that: The bad lane change characteristic parameter K5 is obtained according to Formula 5 using the longitudinal speed v1 of the ego vehicle, the longitudinal speed v2 of the vehicle in front of the target lane, the longitudinal speed v3 of the vehicle behind the target lane, the distance D2 between the ego vehicle and the vehicle in front of the target lane, and the distance D3 between the ego vehicle and the vehicle behind the target lane. Formula 5 is: K5(t) is the characteristic parameter of the poor lane change at time t; v1(t) is the longitudinal velocity of the ego vehicle at time t; v2 is the longitudinal velocity of the preceding vehicle in the target lane, v3 is the longitudinal velocity of the following vehicle in the target lane, D2 is the distance between the ego vehicle and the preceding vehicle in the target lane, and D3 is the distance between the ego vehicle and the following vehicle in the target lane.

9. The method for constructing a driving behavior spectrum in an assisted driving environment according to claim 3, characterized in that: The process of calculating the dangerous driving behavior spectrum characteristic value G of the driving behavior includes the following steps: The dangerous driving behavior characteristic parameter threshold is calculated according to Formula 6 using the dangerous driving behavior characteristic parameter. Formula 6 is: Q i is the upper quartile of the characteristic parameter value of the i-th dangerous driving behavior, I i is the quartile difference, that is, the difference between the upper quartile and the lower quartile of the characteristic parameter value of the i-th dangerous driving behavior. According to formula 7, the instantaneous risk score value S corresponding to each dangerous driving behavior characteristic parameter at time t is calculated i (t); Formula 7 is: S i (t) is normalized by formula 8 to obtain S i (t) corresponding to A Ni ; Formula 8 is: A i is the average risk score value corresponding to each dangerous driving behavior characteristic parameter of driving behavior; A i By S i (t) is averaged after accumulation; According to A Ni The dangerous driving behavior spectrum characteristic value G of the driving behavior is calculated by formula 9; Formula nine is: W i is the weight of the characteristic parameter of the i-th dangerous driving behavior.

10. The method for constructing a driving behavior spectrum in an assisted driving environment according to claim 9, characterized in that: The W i Calculation methods include: Calculate p by formula 10 ij ; Formula 10 is: j=1,2,…,m;i=1,2,…,n; p ji A is the characteristic parameter of the i-th dangerous driving behavior of the j-th driving behavior sample Ni , accounting for the characteristic parameters of the ith dangerous driving behavior of all driving behavior samples A Ni and the proportion; The entropy value E of the characteristic parameter of the i-th dangerous driving behavior is calculated by formula 11: i ; Formula 11 is: The weight W of the characteristic parameter of the i-th dangerous driving behavior is calculated by formula 12: i ; Formula 11 is: i=1,2,…,n; and / or For several driving behaviors whose dangerous driving behavior spectrum characteristic value G is not 0, the dangerous driving behavior spectrum characteristic value G is clustered using the K-means clustering method, and the several driving behaviors are divided into four categories; The dangerous driving behavior spectrum characteristic values ​​G of several driving behaviors included in each type of driving behavior are averaged to obtain the dangerous driving behavior spectrum characteristic value G of each type of driving behavior. j '; The dangerous driving behavior spectrum characteristic value G of each type of driving behavior j 'Arrange by size, from largest to smallest G j The corresponding different categories of driving behaviors are marked as high-risk driving behavior, medium-risk driving behavior, low-risk driving behavior, and safe driving behavior; The maximum deviation value is removed from the dangerous driving behavior spectrum characteristic value G of each driving behavior included in the safe driving behavior, and the range of the remaining dangerous driving behavior spectrum characteristic value G is the safe driving behavior range; Remove the maximum deviation value from the dangerous driving behavior spectrum characteristic value G of each driving behavior included in the low-risk driving behavior, and use the maximum value of the remaining dangerous driving behavior spectrum characteristic value G as the maximum value of the low-risk driving behavior range, and use the maximum G value of the safe driving behavior range as the maximum value of the low-risk driving behavior range; Remove the maximum deviation value from the dangerous driving behavior spectrum characteristic value G of each driving behavior included in the medium-risk driving behavior, and use the maximum value of the remaining dangerous driving behavior spectrum characteristic value G as the maximum value of the low-risk driving behavior range, and use the maximum G value of the low-risk driving behavior range as the maximum value of the medium-risk driving behavior range; The maximum deviation value is removed from the dangerous driving behavior spectrum characteristic value G of each driving behavior included in the high-risk driving behavior, and the maximum value of the remaining dangerous driving behavior spectrum characteristic value G is taken as the maximum value of the low-risk driving behavior range, and the maximum G value of the medium-risk driving behavior range is taken as the maximum value of the high-risk driving behavior range.

Images