CN112100857B - Risk assessment method for distracted driving behaviors - Google Patents

Risk assessment method for distracted driving behaviors Download PDF

Info

Publication number
CN112100857B
CN112100857B CN202010980521.3A CN202010980521A CN112100857B CN 112100857 B CN112100857 B CN 112100857B CN 202010980521 A CN202010980521 A CN 202010980521A CN 112100857 B CN112100857 B CN 112100857B
Authority
CN
China
Prior art keywords
driving
vehicle
risk
behaviors
behavior
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202010980521.3A
Other languages
Chinese (zh)
Other versions
CN112100857A (en
Inventor
金立生
郭柏苍
华强
谢宪毅
张舜然
迟浩天
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Yanshan University
Original Assignee
Yanshan University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Yanshan University filed Critical Yanshan University
Priority to CN202010980521.3A priority Critical patent/CN112100857B/en
Publication of CN112100857A publication Critical patent/CN112100857A/en
Application granted granted Critical
Publication of CN112100857B publication Critical patent/CN112100857B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/10—Complex mathematical operations
    • G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • General Engineering & Computer Science (AREA)
  • Pure & Applied Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Computational Mathematics (AREA)
  • Mathematical Analysis (AREA)
  • Mathematical Optimization (AREA)
  • Algebra (AREA)
  • Probability & Statistics with Applications (AREA)
  • Operations Research (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Evolutionary Biology (AREA)
  • Databases & Information Systems (AREA)
  • Software Systems (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Computer Hardware Design (AREA)
  • Evolutionary Computation (AREA)
  • Geometry (AREA)
  • Traffic Control Systems (AREA)

Abstract

The invention provides a risk assessment method for distracted driving behaviors, which specifically comprises the following steps: determining driving behavior types and driving tasks, designing driving schemes and collecting parameters, extracting abnormal acceleration event frequency, primarily evaluating risk levels of driving behaviors, comparing double driving performances and cross-comparing triple driving performances. The method provided by the invention breaks through the limitation of only researching the distraction driving behaviors, improves the distraction driving behaviors to the height of risk analysis, strictly evaluates the driving performance of the selected distraction driving behaviors, finally obtains a strict risk grade sequencing result, and provides theoretical reference and practical guidance for technologies such as a risk driving behavior identification technology, the safety evaluation of the driving state of a driver, a control right switching scheme of a man-machine co-driving situation and the like.

Description

Risk assessment method for distracted driving behaviors
Technical Field
The invention relates to the field of traffic, in particular to a risk assessment method for distracted driving behaviors.
Background
During driving, a driver unconsciously executes certain actions or ideas, so that the running state of the vehicle tends to or is directly biased to an unstable state, the safety of the driver and other road users is threatened, and the actions are generally called as risky driving actions. The driving risk caused by the non-standard behavior of the driver is accurately predicted in time, and reasonable intervention measures are taken, so that traffic accidents caused by factors of the driver can be effectively avoided.
The distracted driving behavior is the most extensive and common driving behavior among various driving behaviors, and the accident quantity and the death quantity are the highest, the previous research is mostly limited to paying attention to the distracted driving behavior, and the research target is positioned for identifying and evaluating whether the driver has the distracted driving behavior and is not associated with the risk of driving; or evaluating the loss of the driving performance by comparing and analyzing the distraction behavior with the standard driving behavior. However, the distracted driving behavior of the driver does not necessarily lead to the driver being in a distracted state, and even in a distracted state, the driver does not necessarily directly lead to the vehicle being in a risk state; on the other hand, if the distracted driving behavior is only compared with the concentrated (i.e. non-interfering) driving behavior, the result will be limited to the loss of the characteristic distracted driving behavior to the driving performance, and the actual driving risk will not be considered. Therefore, it is out of the actual situation and lacks practical application value to recognize only the distracted driving behavior itself without considering the actual driving risk state and the loss of the driving performance.
In various distracted driving behaviors, the behavior and the driving condition of the vehicle are directly or indirectly influenced by different loads borne by the driver, and if the vehicle continuously drives under a higher load, the probability of misjudgment, decision or operation of the driver is increased, and the risk level of vehicle driving is increased. Excessive information processing type driving tasks weaken the information processing capability of a driver, and the perception judgment capability and the driving operation capability of the driver to the driving environment and the periphery are reduced.
Disclosure of Invention
The invention aims to solve the technical problems and provides a risk assessment method for decentralized driving behaviors, which designs a vehicle-vehicle relative motion state test scheme which is similar to a common urban road and has different investigation emphasis points, and collects vehicle operation parameters, eye movement state parameters, driver operation characteristic parameters and relative motion state parameters of a self vehicle and other vehicles in the whole test process; the method comprises the steps that the risk of the distracted driving behaviors is preliminarily evaluated and classified by using the abnormal motion state of a vehicle, and then the risk grade of the risky driving behaviors is divided; and (3) taking the vehicle-vehicle relative motion state parameters as driving performance characterization parameters, evaluating the driving performance of different distracted driving behaviors, further verifying the risk grade division result of the risky driving behaviors, and finally obtaining the risk grade of the selected distracted driving behaviors for quantitative analysis. The method specifically comprises the following steps:
(1) determining the driving behavior type and the driving task:
selecting normal driving behaviors and multi-class distraction driving behaviors as main evaluation contents of the risk driving behaviors; dividing the selected distracted driving behavior into high and low task loads; finally, the determined normal driving behaviors and the determined high-low load distraction driving behaviors are used as driving tasks to be sequentially numbered;
the normal driving behavior is focused driving, and the distracted driving behavior comprises one or more of handheld mobile phone communication, text message editing, text message reading, conversation and thinking;
(2) designing a driving scheme and acquiring parameters:
designing a driving scene and an interactive driving scheme covering active and passive risk driving events and common road emergency events which are centered by a driver, and completing the interactive driving scheme by a plurality of subjects in the driving scene; sequentially testing each driving task designed in the step (1), wherein each driving task needs to be executed in the whole driving process; collecting vehicle operation parameters, driving human eye movement state parameters, driving human manipulation characteristic parameters and relative movement state parameters of the own vehicle and other vehicles in the whole driving process;
(3) extracting abnormal acceleration event frequency:
extracting the frequency of abnormal acceleration events in the process of executing each driving task by the driver according to the acquired parameters; the abnormal acceleration event is a risk driving event causing an abnormal acceleration state of the vehicle, and comprises one or more of longitudinal rapid deceleration or rapid braking, longitudinal rapid acceleration or rapid starting, lateral rapid left turn, lateral rapid right turn, abnormal left yaw and abnormal right yaw;
the judgment standard of the abnormal acceleration event is as follows:
the above-mentionedLongitudinal sudden deceleration or sudden braking: the longitudinal acceleration a is less than or equal to-0.45 g; the longitudinal rapid acceleration or rapid starting is as follows: the longitudinal acceleration a is more than or equal to 0.35 g; the lateral sudden left turn: the transverse acceleration a is less than or equal to-0.50 g; the lateral sudden right turn: the transverse acceleration a is more than or equal to 0.50 g; abnormal left yaw and abnormal right yaw: the accumulated yaw angle change is more than 6 degrees or less than-6 degrees within 3 seconds; wherein, the acceleration value takes the forward direction and the right side of the vehicle as positive values, takes the reverse direction and the left side of the vehicle as negative values, and g is 9.80m/s2。
(4) And (3) carrying out preliminary evaluation on the risk level of driving behaviors:
the risk condition of each driving behavior is measured by using the average abnormal acceleration occurrence rate, and the average abnormal acceleration occurrence rate M is calculated by the following method:
Figure BDA0002687353220000031
dividing the risk levels of different driving behaviors into two groups, namely a high risk group and a low risk group, classifying the driving behaviors with M being more than or equal to 0.03 into the high risk group, classifying the driving behaviors with M being less than 0.03 into the low risk group, sequencing the driving behaviors in each group according to the average incidence rate M of abnormal acceleration, and further primarily dividing the risk levels of the driving behaviors;
(5) and (3) performing double-driving performance comparison:
introducing driving behavior risk characterization parameters to calculate a driving performance comparison coefficient, and comparing the driving performance difference with the normal driving behavior by calculating the driving performance difference of a driver when executing a certain risky driving behavior so as to analyze the driving performance difference degree of different risky driving behaviors;
the method for defining and calculating the driving behavior risk characterization parameters comprises the following steps:
a) braking reaction time: the time length from the turning on of the brake lamp of the front automobile to the braking of the driver of the vehicle to be tested comprises the perception and the action time of the participants;
b) vehicle speed fluctuation rate: the fluctuation of the speed of the vehicle is measured and defined as the standard deviation of the speed from the average speed at different sampling times along the lane:
Figure BDA0002687353220000041
in the formula (I), the compound is shown in the specification,
Figure BDA0002687353220000042
is the vehicle speed fluctuation value, viIs the speed of the own vehicle at the ith sampling moment,
Figure BDA0002687353220000043
is the average speed of the vehicle during sampling, and n is the number of sampling points;
c) yaw rate fluctuation rate: the stability of the driver to the lateral control of the self vehicle is measured, and the smaller the fluctuation rate is, the more stable the lateral control state of the driver is:
Figure BDA0002687353220000044
in the formula (I), the compound is shown in the specification,
Figure BDA0002687353220000045
is the yaw-rate fluctuation value, riIs the yaw rate of the own vehicle at the ith sampling time,
Figure BDA0002687353220000046
is the average yaw rate of the vehicle during sampling, and n is the number of sampling points;
d) inverse collision time: the collision time is the time which can be used by a driver of the self-vehicle at a certain moment to avoid collision with a front vehicle by adjusting the speed of the self-vehicle, the value of the time is equal to the following distance divided by the relative speed, the collision time TTC has the problem that the speed of the front vehicle and the speed of the rear vehicle are equal, and the concept of the reciprocal of the collision time is introduced:
Figure BDA0002687353220000047
Figure BDA0002687353220000048
in the formula: t is tTTCAs a time-to-collision value, tRTCAs inverse value of time of collision, diIs the following distance between the front and rear cars, vaIs the speed of the bicycle; v. ofbThe speed of the leading vehicle.
The driving performance comparison coefficient is gbiThe calculation method is that the risk characterization parameters of the driver and the normal driving behavior are compared when the driver executes certain risk driving behavior, and the calculation method comprises the following steps:
Figure BDA0002687353220000049
in the formula, thetabiIs a driving behaviour risk characterizing parameter, theta, at the time of execution of a certain driving behaviour b0iIs a driving behavior risk characterization parameter when focusing on driving behavior; i is the driver number, b is the driving behavior number;
mean value mu of a driving performance comparison coefficient using a certain driving behavior bbAnd standard deviation σbCarrying out quantitative analysis:
Figure BDA0002687353220000051
wherein n is the number of subjects; assuming the driving performance comparison coefficient of the driving behavior b as the mean value mubAnd standard deviation σbObeying a lognormal distribution;
difference in driving behavior b as compared to driving performance focused on driving (μ)b-1) analyzing the driving performance difference degrees of different risky driving behaviors, sorting according to the difference degree, wherein the greater the difference degree is, the greater the risk is, and verifying the risk assessment result of the driving behavior in the previous step;
if the sequence of the driving behaviors is different from the risk assessment result in the last step, or the difference degree is similar, or the difference degree of the driving performance of specific driving behaviors is difficult to distinguish, setting the corresponding driving behaviors as pending items, and further determining through next-step triple driving performance cross comparison;
the driving behavior risk characterization parameters comprise one or more of brake reaction time, vehicle speed fluctuation rate, yaw rate fluctuation rate and collision time reciprocal;
(6) and (3) cross-comparing the triple driving performances:
aiming at the distraction driving behaviors to be determined in the last step, determining the relative influence degree between different distraction driving behaviors, namely the difference degree of driving performances, by adopting a pairwise comparison method, and introducing a cross comparison coefficient eta of the driving performancesbciFor reflecting the relative influence of the drivers when executing a certain driving behavior b and a driving behavior c, i.e. the driving performance difference (eta) of the driving behavior b of the ith driver compared with the driving behavior cbci-1)；
Figure BDA0002687353220000052
In the formula gbiAnd gciDriving performance comparison coefficients, v, for driving behavior b and driving behavior c, respectivelybiAnd vciI is a driver number, i is a driving behavior risk characterization parameter when a certain driving behavior b and c is executed respectively;
the mean value of the driving performance cross-comparison coefficients for driving behaviors b and c is recorded as ζbc；
Figure BDA0002687353220000061
The influence of the risky driving behavior b on the driving performance is different from that of the driving behavior c (ζ)bc-1) to determine the degree of difference in driving performance between different driving behaviors at risk;
driving performance cross-comparison coefficient variation (ζ)bc-1) the positive value and the negative value respectively represent the large and small relationship of the risk degrees of the two distracted driving behaviors; due to the fact that the driving behavior risk characterization parameters comprise various driving behavior risk characterization parameters, the situation that positive values and negative values exist simultaneously can exist, and therefore, the driving behavior risk characterization parameters exceed halfNumber risk characterization parameter corresponds to (ζ)bc-1) positive, indicating that the overall driving performance of the driving behaviour b is higher than the driving behaviour c, whereby the driving behaviour b is inferred to be more risky than the driving behaviour c; whereas more than half (ζ)bc-1) is negative, indicating that the driving behaviour b is less risky than the driving behaviour c; if (ζ)bcThe positive and negative values of-1) are equal, which indicates that the driving behavior b and the driving behavior c have the same risk degree and are classified into the same risk level in parallel.
And analyzing the difference degree of the driving performance, and finally sequencing the risk assessment results according to the difference degree.
The invention has the beneficial effects that:
the method provided by the invention breaks through the limitation of only researching the distraction driving behaviors, improves the distraction driving behaviors to the risk analysis height, has more practical significance, and provides theoretical reference and practical guidance for technologies such as a risk driving behavior identification technology, safety evaluation of the driving state of a driver, a control right switching scheme of a man-machine co-driving situation and the like.
The method provided by the invention designs a road test scheme, contains more urban road driving scenes and driving events in a limited field, and comprehensively measures the influence effect of the distracted driving task on the driver from the aspects of the transverse and longitudinal control capability and the reaction capability of the driver, and the acquired data contains more information quantity dimensions and wider characteristics.
The method provided by the invention strictly evaluates the driving performance of the selected distracted driving behaviors through preliminary division, double driving performance comparison and triple driving performance cross comparison analysis in sequence, and finally obtains a strict risk grade sequencing result.
Drawings
FIG. 1 is a flow chart of the interactive behavior of a driving scenario contemplated by the present invention;
FIG. 2 is a statistical chart of the frequency of abnormal acceleration events according to the present invention;
FIG. 3 is a diagram of the average occurrence rate M of abnormal acceleration according to the present invention;
FIG. 4 is a schematic view of data visualization of a double comparison of variations in driving performance characterization parameters according to the present invention;
fig. 5 is a data visualization schematic diagram of the cross comparison of the variation of the driving performance characterization parameter according to the present invention.
Detailed Description
The invention provides a risk assessment method for distracted driving behaviors, which comprises the following steps:
(1) determining the driving behavior type and the driving task:
selecting 1 type of normal driving behaviors and five types of distraction driving behaviors as main evaluation contents of the risk driving behaviors; dividing the selected distracted driving behavior into high and low task loads;
the driving behaviour and tasks designed were as follows:
the normal driving behavior is focused driving: the subject does not have other behaviors irrelevant to driving and other external environment interferences in the driving process, and a driving test is carried out according to a trained test design scheme and a trained route;
the distracted driving behavior comprises the following steps:
and (3) carrying out conversation by the handheld mobile phone: the test subject answers the telephone of the test assistant by the habitual hand and keeps talking in the whole process during the driving process, wherein the high-load task requires the test subject to answer the mental arithmetic questions with higher difficulty, and the low-load task requires the test subject to answer the mental arithmetic questions with lower difficulty;
editing the text message: the test subject uses the habitual hand of the test subject in the instant chat software to give a test haircut text message in the whole driving process, wherein the high-load task requires the test subject to write text to report the current driving position, speed and surrounding vehicle state information, and the low-load task only requires the test subject to write smooth text at will;
reading the text message: the examinee opens the preset text content of the mobile phone by the habitual hand to read in the whole driving process, wherein the high-load task requires the examinee to understand and memorize the text content as much as possible and encourages the behavior in a reward form, and the low-load task only requires the examinee to read the text content;
talking: the test assistant who answers the copilot position in the whole driving process of the subject asks the subject to ask questions, wherein the high-load task asks the subject to answer the mental arithmetic questions with higher difficulty, and the low-load task asks the subject to answer the mental arithmetic questions with lower difficulty;
thinking: subjects solve a provisionally given problem throughout the course of driving, wherein a high-load task requires subjects to solve the problem in conjunction with their own profession and experience, and to try to give a more detailed protocol flow; the low load task requires only subjects to comb through a simple solution;
finally, the determined normal driving behaviors and the determined high-low load distraction driving behaviors are used as driving tasks to be sequentially numbered, and the sequence is shown in a table 1;
TABLE 1 Driving behavior orderliness chart
Figure BDA0002687353220000081
(2) Designing a driving scheme and acquiring parameters:
designing a driving scene and an interactive driving scheme covering active and passive risk driving events and common road emergency events which are centered by a driver, and completing the interactive driving scheme by a plurality of subjects in the driving scene;
the driving scene comprises a passive driving task section and an active driving task section, which are divided into four straight line sections, and at least 3 auxiliary vehicles, pedestrian simulation models and moving obstacles are used as road interference factors to assist in completing an interactive driving scheme, as shown in fig. 1 and table 2;
the interactive driving scheme is as follows:
passive driving task section:
section 1: the method comprises the following steps that a tested vehicle approaches a front auxiliary vehicle 1, the auxiliary vehicle 1 runs at a variable speed, the tested vehicle runs with the vehicle, the auxiliary vehicle 1 brakes emergently, the tested vehicle takes a corresponding measure autonomously, the front auxiliary vehicle 2 backs up, and the tested vehicle takes a corresponding measure autonomously;
section 2: the method comprises the following steps that a tested vehicle runs on a right lane, an auxiliary vehicle 1 approaches from an adjacent lane transversely, the tested vehicle autonomously takes a countermeasure, the auxiliary vehicle 1 cuts in forcibly, the tested vehicle autonomously takes the countermeasure, the auxiliary vehicle 1 brakes emergently, the tested vehicle autonomously takes the countermeasure, a pedestrian crosses a road, the tested vehicle autonomously takes the countermeasure, a moving obstacle appears on a lane, and the tested vehicle autonomously takes the countermeasure;
active driving task section:
section 3: the tested vehicle approaches the front auxiliary vehicle 1 on the right lane, the auxiliary vehicle 1 runs at a constant speed on the right lane, the tested vehicle overtakes from the left lane, the auxiliary vehicle 2 runs at a constant speed on the left lane, the tested vehicle overtakes from the right lane, the auxiliary vehicle 3 runs at a constant speed on the right lane, and the tested vehicle overtakes from the left lane;
section 4: the tested vehicle approaches the auxiliary vehicle 1 of the front vehicle, the auxiliary vehicle 1 runs at a constant speed, the tested vehicle changes lane and overtakes and switches back to the original lane, the auxiliary vehicle 2 runs at a constant speed, the tested vehicle changes lane and overtakes and switches back to the original lane, the auxiliary vehicle 3 runs at a constant speed, and the tested vehicle changes lane and overtakes and switches back to the original lane.
The method comprises the following steps that a test subject continuously receives the interference from external behaviors such as emergency braking, approaching of vehicles in adjacent lanes, forced cut-in, crossing of roads by pedestrians and the like when driving vehicles on a passive driving task road section, and the expression characteristics of the risk driving behaviors of a driver in a passive driving state are comprehensively considered as much as possible; the subject takes driving behaviors such as continuous overtaking, active cut-in and the like to finish the target of the subject on the active driving task section, and the risk driving behavior expression characteristics of the driver in a relatively aggressive state are inspected.
TABLE 2 Observation of Driving scenarios and Driving scenarios
Figure BDA0002687353220000101
The subject sequentially executes each driving task designed in the step (1) in the driving scene, and each driving task needs to be executed in the whole driving process; each person needs to travel 5 circles around the whole field for each test.
Vehicle running parameters, driver operation characteristic parameters and relative motion state parameters of the vehicle and other vehicles in the whole driving process are acquired through a vehicle-mounted sensor, wherein the vehicle running parameters, the driver operation characteristic parameters and the relative motion state parameters comprise data such as the speed, the acceleration, the yaw angle, the speed of the vehicle, the relative transverse longitudinal distance and the like of the vehicle, and the driver rotates a steering wheel, steps on an accelerator or a brake pedal and the like; acquiring the eye movement state parameters of a driver through an eye movement instrument;
(3) extracting abnormal acceleration event frequency:
extracting the frequency of abnormal acceleration events in the process of executing each driving task by the driver according to the acquired parameters; the dangerous driving event is mainly caused by that a driver rapidly turns a steering wheel, and suddenly steps on an accelerating or braking pedal to cause the vehicle to be in a sharp turn, a sharp acceleration, a sharp deceleration and the like, and the kinematic parameters are expressed by that the transverse or longitudinal acceleration of the vehicle is obviously different from that of the vehicle in a stable driving process. Such events are dangerous because they increase the possibility of vehicle runaway, shorten the time for the driver to respond to the danger, and make it difficult for other road users to predict and judge the state of the own vehicle, thereby reducing the safety margin. The types of the risky driving events are respectively from low risk to high risk: a safety threat event, an imminent collision event, a collision event, all of which have in common the fact that the vehicle is in an abnormal acceleration state, and therefore, the driving event that results in the abnormal acceleration state is referred to as an abnormal acceleration event, including a longitudinal hard deceleration or hard braking, a longitudinal hard acceleration or fast start, a lateral hard left turn, a lateral hard right turn, an abnormal left yaw, and an abnormal right yaw; abnormal acceleration events are common to driver and passenger discomfort and have a high probability of causing accidents.
The judgment standard of the abnormal acceleration event is as follows:
the longitudinal sudden deceleration or sudden braking: the longitudinal acceleration a is less than or equal to-0.45 g; the longitudinal rapid acceleration or rapid starting is as follows: the longitudinal acceleration a is more than or equal to 0.35 g; the lateral sudden left turn: the transverse acceleration a is less than or equal to-0.50 g; the lateral sudden right turn: the transverse acceleration a is more than or equal to 0.50 g; abnormal left yaw and abnormal right yaw: the accumulated yaw angle change is more than 6 degrees or less than-6 degrees within 3 seconds; wherein, addThe speed value is positive in the forward direction and the right side of the vehicle, negative in the reverse direction and the left side of the vehicle, and g is 9.80m/s2。
The frequency of the abnormal acceleration events is extracted, and the statistical result of the abnormal acceleration events of each distracted driving behavior is shown in fig. 2.
(4) And (3) carrying out preliminary evaluation on the risk level of driving behaviors:
in order to measure the safety of the behavior of the driver, the risk condition of each driving behavior is measured by using the average occurrence rate of abnormal acceleration, and the average occurrence rate M of abnormal acceleration is calculated by the following method:
Figure BDA0002687353220000111
the average occurrence rate M of abnormal acceleration is calculated by using the formula (1), and the calculation result is shown in fig. 3.
Dividing the risk levels of different driving behaviors into two groups of high risk and low risk by taking M as a boundary, classifying the driving behaviors with M being more than or equal to 0.03 into the high risk group, classifying the driving behaviors with M being less than 0.03 into the low risk group, classifying the distraction driving behaviors according to the risk classification standard and carrying out preliminary sequencing based on the statistical result of the abnormal acceleration event, sequencing the driving behaviors in each group according to the average incidence rate M of the abnormal acceleration, and further carrying out preliminary classification on the risk levels of the driving behaviors, wherein the classification is shown in the following table 3.
TABLE 3 Primary ranking of distraction behavior Risk levels based on abnormal acceleration discrimination
Figure BDA0002687353220000112
Figure BDA0002687353220000121
(5) And (3) performing double-driving performance comparison:
and introducing driving behavior risk characterization parameters, and analyzing the driving performance of the driver in different driving behavior tests through the characterization parameters.
The method for defining and calculating the driving behavior risk characterization parameters comprises the following steps:
e) brake Reaction Time (BRT): the time length from the turning on of the brake lamp of the front automobile to the braking of the driver of the vehicle to be tested comprises the perception and the action time of the participants;
f) vehicle speed fluctuation (DSF): fluctuations in the speed of the vehicle are measured and represent the driver's ability to stably follow the leading vehicle, which is defined as the standard deviation of the speed from the average speed at different sample times along the lane:
Figure BDA0002687353220000122
in the formula (I), the compound is shown in the specification,
Figure BDA0002687353220000123
is the vehicle speed fluctuation value, viIs the speed of the own vehicle at the ith sampling moment,
Figure BDA0002687353220000124
is the average speed of the vehicle during sampling, and n is the number of sampling points;
g) yaw Rate Fluctuation (YRF): the stability of the driver to the lateral control of the self vehicle is measured, and the smaller the fluctuation rate is, the more stable the lateral control state of the driver is:
Figure BDA0002687353220000125
in the formula (I), the compound is shown in the specification,
Figure BDA0002687353220000126
is the yaw-rate fluctuation value, riIs the yaw rate of the own vehicle at the ith sampling time,
Figure BDA0002687353220000127
is the average yaw rate of the vehicle during sampling, and n is the number of sampling points;
h) reciprocal collision time (RTC): the Time To Collision (TTC) is the time available for the driver of the own vehicle to avoid collision with the preceding vehicle by adjusting the speed of the own vehicle at a certain moment, and the value of the time to collision is equal to the following distance divided by the relative speed, and the time to collision can reflect the instant risk state; the problem that the collision time TTC cannot be solved when the speeds of front and rear vehicles are equal is solved, and the concept of reciprocal collision time is introduced, so that the problem that the TTC is infinite when the relative speed is small is solved:
Figure BDA0002687353220000131
Figure BDA0002687353220000132
in the formula: t is tTTCAs a time-to-collision value, tRTCAs inverse value of time of collision, diIs the following distance between the front and rear cars, vaIs the speed of the bicycle; v. ofbThe speed of the leading vehicle.
In order to visually analyze the driving performance difference of different risky driving behaviors and further verify the risk evaluation result of the driving behaviors, a driving performance evaluation index, namely a driving performance comparison coefficient g, is providedbiMeaning that the driver has improved driving performance when performing a certain risky driving behavior compared to the standard driving behavior (g)bi-1). The calculation method is the ratio of the driving performance characteristic parameters of the driver in the state of executing the risky driving behaviors and the normal driving behaviors.
Figure BDA0002687353220000133
In the formula, thetabiIs a driving behaviour risk characterizing parameter, theta, at the time of execution of a certain driving behaviour b0iIs a driving behavior risk characterization parameter when focusing on driving behavior; i is drivingPerson number, b is driving behavior number;
taking into account the elimination of uncertainty factors due to demographic factors, the mean value mu of the comparison coefficient of driving performance for a certain driving behavior b is usedbAnd standard deviation σbCarrying out quantitative analysis:
Figure BDA0002687353220000134
wherein n is the number of subjects; assuming the driving performance comparison coefficient of the driving behavior b as the mean value mubAnd standard deviation σbObeying a lognormal distribution;
mean of lognormal distribution
Figure BDA0002687353220000141
And variance var (x) ═ σ2To define the confidence interval of a lognormal distribution, the parameters E (lnx) and Var (lnx) of the distribution are estimated from its mean and variance; expanding the taylor series of ln x to the mean to obtain:
Figure BDA0002687353220000142
taking the expectations of both sides of (8), ignoring higher order terms, we get:
Figure BDA0002687353220000143
due to the fact that
Figure BDA0002687353220000144
Is constant, and
Figure BDA0002687353220000145
taking the variance on both sides of (8) and ignoring higher order terms, it can be shown that:
Figure BDA0002687353220000146
due to the fact that
Figure BDA0002687353220000147
Is a constant, and the percentile values of 2.5% and 97.5% of the lognormal distribution can be inferred based on (7) and (8), forming a 95% confidence interval; thus, definition H0The comparison coefficient g does not influence the driving performance for the driving behavior bbiThe null hypothesis of (2) can be derived from the lognormal distribution
Figure BDA0002687353220000148
Figure BDA0002687353220000149
95% confidence interval of
Figure BDA00026873532200001410
If it is
Figure BDA00026873532200001411
There is not enough evidence to reject the null hypothesis; otherwise, H0Can be rejected; from this it can be concluded that: at a significance level of 5%, the driving behavior b differed from the driving performance of dedicated driving (μ;)b-1), comparing the driving performance difference with the normal driving behavior by calculating the driving performance difference when the driver executes a certain risky driving behavior, further analyzing the driving performance difference degrees of different risky driving behaviors, and sequencing according to the difference degrees, wherein the greater the difference degree is, namely the greater the risk is, the more the risk is, the risk evaluation result of the driving behavior in the previous step is verified;
the calculated driving performance comparison coefficient is shown in table 4.
TABLE 4 comparison of driving performance coefficient μ for different distracted driving behaviorsbResult of calculation of (2)
Figure BDA00026873532200001412
Figure BDA0002687353220000151
Further, the degree of reduction (mu) of the driving performance of various distracted driving behaviors relative to the case of focusing on driving is obtainedb-1) verifying and appropriately adjusting the "preliminary ranking of risk classes" according to the degree of reduction of the driving performance, as shown in table 5;
TABLE 5 distraction driving behavior grouping scheme based on driving performance characterization parameter variation
Figure BDA0002687353220000152
Data visualization is performed by using data in table 5 to obtain fig. 4, a center zero point is a driving performance characterization parameter of "focus on driving behavior (FD)", and four axes are respectively: the driving performance characteristic parameters of the distraction driving behaviors are increased relative to the characteristic parameters corresponding to the concentration driving behaviors.
If the sequence of the driving behaviors is different from the risk assessment result in the previous step, or the difference degree is similar, or the difference degree of the driving performances of specific driving behaviors is difficult to distinguish, the corresponding driving behaviors are set as pending items to obtain a double risk analysis sequence, and as shown in table 6, the pending items are further determined by the cross comparison of the triple driving performances in the next step:
TABLE 6 Risk level analysis results of double distraction driving behavior
Figure BDA0002687353220000161
(6) And (3) cross-comparing the triple driving performances:
combining table 5 and fig. 4, and comparing with table 3, the following conclusions can be drawn: firstly, from the overall perspective, the risk degree sorting mode of table 5 is substantially similar to the risk classification and sorting results of table 3, and the primary risk analysis result can be verified; secondly, from a detailed perspective, since the above steps are independent comparative analysis of the distracted driving behaviors and the dedicated driving behaviors, there are results with similar difference degrees, or it is difficult to distinguish the difference degrees of the driving performances of specific behaviors (for example, 2 groups and 3 groups in table 5), and the main reason of the results is that different kinds of distracted driving behaviors have different influences on the operation state of the driver, and sometimes it is difficult to avoid that the reduction degrees of the driving performances of a plurality of distracted driving behaviors are similar, but the internal causes are different, so that further triple driving performance cross comparison analysis is needed to solve the problem.
Aiming at the distraction driving behaviors to be determined in the last step, determining the relative influence degree between different distraction driving behaviors, namely the difference degree of driving performances, by adopting a pairwise comparison method, and introducing a cross comparison coefficient eta of the driving performancesbciCalculating the driving performance difference (eta) between the driving behavior b and the driving behavior c of the ith driverbci-1)，
Figure BDA0002687353220000171
In the formula gbiAnd gciDriving performance comparison coefficients, v, for driving behavior b and driving behavior c, respectivelybiAnd vciI is a driver number, i is a driving behavior risk characterization parameter when a certain driving behavior b and c is executed respectively;
the mean value of the driving performance cross-comparison coefficients for driving behaviors b and c is recorded as ζbc；
Figure BDA0002687353220000172
Assuming that the mean and standard deviation of the cross-comparison coefficients for driving performance are both subject to lognormal distribution, H1Represents a null hypothesis, i.e., the relative impact between driving behaviors b and c does not show a difference in performance indicators; distributed in a lognormal manner
Figure BDA0002687353220000173
Determination of 95% confidence intervals
Figure BDA0002687353220000174
If it is
Figure BDA0002687353220000175
The evidence is insufficient to reject the null hypothesis; otherwise, H1Can be rejected; it can thus be concluded that at a significance level of 5%, the impact of the risky driving behavior b on the driving performance differs compared to the driving behavior c (ζ)bc-1) to determine the degree of difference in driving performance between different risky driving behaviors.
Through pairwise comparison and analysis, the performance relative loss degree zeta of the two distraction driving behaviors can be obtainedbcAs shown in table 7.
TABLE 7 CROSS-COMPARISON COEFFICIENCY ZETAR FOR DRIVING PERFORMANCE OF DIFFERENT CENTRIC DRIVING PROPERTIESbcResult of calculation of (2)
Figure BDA0002687353220000176
Figure BDA0002687353220000181
Further, calculate (ζ)bc-1) obtaining table 8; fig. 5 is a data visualization effect of cross-comparison of the variation of the driving performance characterizing parameter.
TABLE 8 Cross-Compare results based on Driving Performance characterization parameter variance
Figure BDA0002687353220000182
Analyzing the degree of difference of the driving performances: sequentially carrying out cross comparison on the driving performances of every two distracted driving behaviors in the project to be determined, and carrying out cross comparison on the driving performances by using the variable quantity (zeta) of the coefficient of the cross comparisonbc-1) the positive value and the negative value respectively represent the large and small relationship of the risk degrees of the two distracted driving behaviors; because 4 driving behavior risk characterization parameters are included, the situation that positive and negative values exist simultaneously can exist, and more than half of the risk characterization parameters correspond toZeta ofbc-1) positive, indicating that the overall driving performance of the driving behaviour b is higher than the driving behaviour c, whereby the driving behaviour b is inferred to be more risky than the driving behaviour c; for example, if the four driving performance characterization parameters of HCC-H/HCC-L in table 8 are all positive, it indicates that the overall driving performance of HCC-H is higher than that of HCC-L, and thus the HCC-H risk degree is inferred to be higher than that of HCC-L. Whereas more than half (ζ)bc-1) is negative, indicating that the driving behaviour b is less risky than the driving behaviour c; if (ζ)bcThe positive and negative values of-1) are equal, which indicates that the driving behavior b is equivalent to the driving behavior c in risk degree, and the driving behavior b is classified into the same risk grade in parallel, for example, the BRT and RTC of HCC-L/TTM-L in table 8 are positive values, and the DSF and YRF are negative values, which indicates that the driving behavior b cannot be differentiated in risk degree despite the triple cross-comparison analysis, and the driving behavior b is classified into the same risk grade because the drivers participate in different types of distraction during the driving process, and the influence on the lateral and longitudinal control ability and the reaction time of the drivers are different.
Based on the above analysis and demonstration, the risk analysis results for the centric driving behavior are completed, and strict risk level ranking results are obtained, as shown in table 9.
TABLE 9 Risk assessment results for distracted driving behavior
Figure BDA0002687353220000191

Claims (8)

1. A risk assessment method for distracted driving behavior is characterized in that: the method comprises the following steps:
(1) determining the driving behavior type and the driving task:
selecting normal driving behaviors and multi-class distraction driving behaviors as main evaluation contents of the risk driving behaviors; dividing the selected distracted driving behavior into high and low task loads; finally, the determined normal driving behaviors and the determined high-low load distraction driving behaviors are used as driving tasks to be sequentially numbered;
the normal driving behavior is focused driving, and the distracted driving behavior comprises one or more of handheld mobile phone communication, text message editing, text message reading, conversation and thinking;
(2) designing a driving scheme and acquiring parameters:
designing a driving scene and an interactive driving scheme covering active and passive risk driving events and common road emergencies by taking a driver as a center, and completing the driving scheme by a plurality of subjects in the driving scene; sequentially testing each driving task designed in the step (1), wherein each driving task needs to be executed in the whole driving process; collecting vehicle operation parameters, driving human eye movement state parameters, driving human manipulation characteristic parameters and relative movement state parameters of the own vehicle and other vehicles in the whole driving process;
(3) extracting abnormal acceleration event frequency:
extracting the frequency of abnormal acceleration events in the process of executing each driving task by the driver according to the acquired parameters; the abnormal acceleration event is a risk driving event causing an abnormal acceleration state of the vehicle, and comprises one or more of longitudinal rapid deceleration or rapid braking, longitudinal rapid acceleration or rapid starting, lateral rapid left turn, lateral rapid right turn, abnormal left yaw and abnormal right yaw;
(4) and (3) carrying out preliminary evaluation on the risk level of driving behaviors:
the risk condition of each driving behavior is measured by using the average abnormal acceleration occurrence rate, and the average abnormal acceleration occurrence rate M is calculated by the following method:
Figure FDA0002687353210000011
dividing the risk levels of different driving behaviors into two groups, namely a high risk group and a low risk group, classifying the driving behaviors with M being more than or equal to 0.03 into the high risk group, classifying the driving behaviors with M being less than 0.03 into the low risk group, sequencing the driving behaviors in each group according to the average incidence rate M of abnormal acceleration, and further primarily dividing the risk levels of the driving behaviors;
(5) and (3) performing double-driving performance comparison:
introducing driving behavior risk characterization parameters to calculate a driving performance comparison coefficient, and calculating the driving performance difference between a driver executing a certain risky driving behavior and a normal driving behavior so as to analyze the driving performance difference degrees of different risky driving behaviors, sequencing according to the difference degrees, wherein the greater the difference degree is, the greater the risk is, and verifying the risk evaluation result of the driving behavior in the previous step;
if the sequencing of the driving behaviors is different from the risk assessment result in the previous step, or the difference degree is similar, or the difference degree of the driving performances of specific driving behaviors is difficult to distinguish, setting the corresponding driving behaviors as pending items, and further determining through next-step triple driving performance cross comparison;
the driving behavior risk characterization parameters comprise one or more of brake reaction time, vehicle speed fluctuation rate, yaw rate fluctuation rate and collision time reciprocal;
(6) and (3) cross-comparing the triple driving performances:
aiming at the distracted driving behaviors to be determined in the last step, determining the relative influence degree between different distracted driving behaviors by adopting a pairwise comparison method, namely the driving performance difference degree, introducing a driving performance cross comparison coefficient to calculate the driving performance difference between the driving behavior b and the driving behavior c of the ith driver, analyzing the driving performance difference degree, and finally sequencing the risk assessment results according to the difference degree.
2. The risk assessment method of distracted driving behavior according to claim 1, characterized in that: the driving task focusing on driving in the step (1) is as follows: the subject does not have other behaviors irrelevant to driving and other external environment interferences in the driving process, and a driving test is carried out according to a trained test design scheme and a trained route;
the driving task of the handheld mobile phone call is as follows: the test subject answers the telephone of the test assistant by the habitual hand and keeps talking in the whole process during the driving process, wherein the high-load task requires the test subject to answer the mental arithmetic questions with higher difficulty, and the low-load task requires the test subject to answer the mental arithmetic questions with lower difficulty;
the driving task of editing the text message comprises the following steps: the test subject uses the habitual hand of the test subject in the instant chat software to give a test haircut text message in the whole driving process, wherein the high-load task requires the test subject to write text to report the current driving position, speed and surrounding vehicle state information, and the low-load task only requires the test subject to write smooth text at will;
the driving task of reading the text message is as follows: the examinee opens the preset text content of the mobile phone by the habitual hand to read in the whole driving process, wherein the high-load task requires the examinee to understand and memorize the text content as much as possible and encourages the behavior in a reward form, and the low-load task only requires the examinee to read the text content;
the driving task of the conversation is as follows: the test assistant who answers the copilot position in the whole driving process of the subject asks the subject to ask questions, wherein the high-load task asks the subject to answer the mental arithmetic questions with higher difficulty, and the low-load task asks the subject to answer the mental arithmetic questions with lower difficulty;
the driving tasks of thinking are as follows: subjects solve a provisionally given problem throughout the course of driving, wherein a high-load task requires subjects to solve the problem in conjunction with their own profession and experience, and to try to give a more detailed protocol flow; the low load task requires the subject to comb only a simple solution.
3. The risk assessment method of distracted driving behavior according to claim 1, characterized in that: the driving scene in the step (2) comprises a passive driving task section and an active driving task section, and at least 3 auxiliary vehicles, pedestrian simulation models and moving obstacles are used as road interference factors to assist in completing an interactive driving scheme;
the interactive driving scheme is as follows:
passive driving task section: the method comprises the following steps that a tested vehicle approaches a front auxiliary vehicle 1, the auxiliary vehicle 1 runs at a variable speed, the tested vehicle runs with the vehicle, the auxiliary vehicle 1 brakes emergently, the tested vehicle takes a corresponding measure autonomously, the front auxiliary vehicle 2 backs up, and the tested vehicle takes a corresponding measure autonomously;
the method comprises the following steps that a tested vehicle runs on a right lane, an auxiliary vehicle 1 approaches from an adjacent lane transversely, the tested vehicle autonomously takes a countermeasure, the auxiliary vehicle 1 cuts in forcibly, the tested vehicle autonomously takes the countermeasure, the auxiliary vehicle 1 brakes emergently, the tested vehicle autonomously takes the countermeasure, a pedestrian crosses a road, the tested vehicle autonomously takes the countermeasure, a moving obstacle appears on a lane, and the tested vehicle autonomously takes the countermeasure;
active driving task section: the tested vehicle approaches the front auxiliary vehicle 1 on the right lane, the auxiliary vehicle 1 runs at a constant speed on the right lane, the tested vehicle overtakes from the left lane, the auxiliary vehicle 2 runs at a constant speed on the left lane, the tested vehicle overtakes from the right lane, the auxiliary vehicle 3 runs at a constant speed on the right lane, and the tested vehicle overtakes from the left lane;
the tested vehicle approaches the auxiliary vehicle 1 of the front vehicle, the auxiliary vehicle 1 runs at a constant speed, the tested vehicle changes lane and overtakes and switches back to the original lane, the auxiliary vehicle 2 runs at a constant speed, the tested vehicle changes lane and overtakes and switches back to the original lane, the auxiliary vehicle 3 runs at a constant speed, and the tested vehicle changes lane and overtakes and switches back to the original lane.
4. The risk assessment method of distracted driving behavior according to claim 1, characterized in that: the discrimination criteria of the abnormal acceleration event in the step (3) are as follows:
the longitudinal sudden deceleration or sudden braking: the longitudinal acceleration a is less than or equal to-0.45 g; the longitudinal rapid acceleration or rapid starting is as follows: the longitudinal acceleration a is more than or equal to 0.35 g; the lateral sudden left turn: the transverse acceleration a is less than or equal to-0.50 g; the lateral sudden right turn: the transverse acceleration a is more than or equal to 0.50 g; abnormal left yaw and abnormal right yaw: the accumulated yaw angle change is more than 6 degrees or less than-6 degrees within 3 seconds; wherein, the acceleration value takes the forward direction and the right side of the vehicle as positive values, takes the reverse direction and the left side of the vehicle as negative values, and g is 9.80m/s2。
5. The risk assessment method of distracted driving behavior according to claim 1, characterized in that: the method for defining and calculating the driving behavior risk characterization parameters in the step (5) is as follows:
a) braking reaction time: the time length from the turning on of the brake lamp of the front automobile to the braking of the driver of the vehicle to be tested comprises the perception and the action time of the participants;
b) vehicle speed fluctuation rate: the fluctuation of the speed of the vehicle is measured and defined as the standard deviation of the speed from the average speed at different sampling times along the lane:
Figure FDA0002687353210000051
in the formula (I), the compound is shown in the specification,
Figure FDA0002687353210000052
is the vehicle speed fluctuation value, viIs the speed of the own vehicle at the ith sampling moment,
Figure FDA0002687353210000053
is the average speed of the vehicle during sampling, and n is the number of sampling points;
c) yaw rate fluctuation rate: the stability of the driver to the lateral control of the self vehicle is measured, and the smaller the fluctuation rate is, the more stable the lateral control state of the driver is:
Figure FDA0002687353210000054
in the formula (I), the compound is shown in the specification,
Figure FDA0002687353210000055
is the yaw-rate fluctuation value, riIs the yaw rate of the own vehicle at the ith sampling time,
Figure FDA0002687353210000056
is the average yaw rate of the vehicle during sampling, and n is the number of sampling points;
d) inverse collision time: the collision time is the time which can be used by a driver of the self-vehicle at a certain moment to avoid collision with a front vehicle by adjusting the speed of the self-vehicle, the value of the time is equal to the following distance divided by the relative speed, the collision time TTC has the problem that the speed of the front vehicle and the speed of the rear vehicle are equal, and the concept of the reciprocal of the collision time is introduced:
Figure FDA0002687353210000057
Figure FDA0002687353210000058
in the formula: t is tTTCAs a time-to-collision value, tRTCAs inverse value of time of collision, diIs the following distance between the front and rear cars, vaIs the speed of the bicycle; v. ofbThe speed of the leading vehicle.
6. The risk assessment method of distracted driving behavior according to claim 1, characterized in that: the driving performance comparison coefficient in the step (5) is gbiThe calculation method is that the risk characterization parameters of the driver and the normal driving behavior are compared when the driver executes certain risk driving behavior, and the calculation method comprises the following steps:
Figure FDA0002687353210000059
in the formula, thetabiIs a driving behaviour risk characterizing parameter, theta, at the time of execution of a certain driving behaviour b0iIs a driving behavior risk characterization parameter when focusing on driving behavior; i is the driver number, b is the driving behavior number;
taking into account the elimination of uncertainty factors due to demographic factors, the mean value mu of the comparison coefficient of driving performance for a certain driving behavior b is usedbAnd standard deviation σbCarrying out quantitative analysis:
Figure FDA0002687353210000061
wherein n is the number of subjects; assuming the driving performance comparison coefficient of the driving behavior b as the mean value mubAnd standard deviation σbObeying a lognormal distribution;
mean of lognormal distribution
Figure FDA0002687353210000062
And variance var (x) ═ σ2To define the confidence interval of a lognormal distribution, the parameters E (ln x) and Var (ln x) of the distribution are estimated from its mean and variance; expanding the taylor series of ln x to the mean to obtain:
Figure FDA0002687353210000063
taking the expectations of both sides of (8), ignoring higher order terms, we get:
Figure FDA0002687353210000064
due to the fact that
Figure FDA0002687353210000065
Is constant, and
Figure FDA0002687353210000066
taking the variance on both sides of (8) and ignoring higher order terms, it can be shown that:
Figure FDA0002687353210000067
due to the fact that
Figure FDA0002687353210000068
Is a constant, and the percentile values of 2.5% and 97.5% of the lognormal distribution can be inferred based on (7) and (8), forming a 95% confidence interval(ii) a Thus, definition H0The comparison coefficient g does not influence the driving performance for the driving behavior bbiThe null hypothesis of (2) can be derived from the lognormal distribution
Figure FDA0002687353210000069
Figure FDA00026873532100000610
95% confidence interval of
Figure FDA00026873532100000611
If it is
Figure FDA00026873532100000612
There is not enough evidence to reject the null hypothesis; otherwise, H0Can be rejected; from this it can be concluded that: at a significance level of 5%, the driving behavior b differed from the driving performance of dedicated driving (μ;)b-1) to analyze the degree of difference in driving performance between different risky driving behaviors.
7. The risk assessment method of distracted driving behavior according to claim 1, characterized in that: in step (6), in order to further determine the relative influence degree between different distracted driving behaviors in the project to be determined, a driving performance cross comparison coefficient eta is providedbciFor reflecting the relative influence of the drivers when executing a certain driving behavior b and a driving behavior c, i.e. the driving performance difference (eta) of the driving behavior b of the ith driver compared with the driving behavior cbci-1)；
Figure FDA0002687353210000071
In the formula gbiAnd gciDriving performance comparison coefficients, v, for driving behavior b and driving behavior c, respectivelybiAnd vciI is a driver number, i is a driving behavior risk characterization parameter when a certain driving behavior b and c is executed respectively;
the mean value of the driving performance cross-comparison coefficients for driving behaviors b and c is recorded as ζbc；
Figure FDA0002687353210000072
Assuming that the mean and standard deviation of the cross-comparison coefficients for driving performance are both subject to lognormal distribution, H1Represents a null hypothesis, i.e., the relative impact between driving behaviors b and c does not show a difference in performance indicators; distributed in a lognormal manner
Figure FDA0002687353210000073
Determination of 95% confidence intervals
Figure FDA0002687353210000074
If it is
Figure FDA0002687353210000075
The evidence is insufficient to reject the null hypothesis; otherwise, H1Can be rejected; it can thus be concluded that at a significance level of 5%, the impact of the risky driving behavior b on the driving performance differs compared to the driving behavior c (ζ)bc-1) to determine the degree of difference in driving performance between different risky driving behaviors.
8. The risk assessment method of distracted driving behavior according to claim 7, wherein: sequentially carrying out cross comparison analysis on the driving performance of every two distracted driving behaviors in the project to be determined, and carrying out cross comparison on the variable quantity (zeta) of the coefficient of the driving performancebc-1) the positive value and the negative value respectively represent the large and small relationship of the risk degrees of the two distracted driving behaviors; since there may be a case where both positive and negative values exist due to the inclusion of a plurality of driving behavior risk characterization parameters, more than half of the risk characterization parameters correspond to (ζ)bc-1) positive, indicating that the overall driving performance of the driving behaviour b is higher than the driving behaviour c, whereby the driving behaviour b is inferred to be more risky than the driving behaviour c; inverse directionMore than half of (ζ)bc-1) is negative, indicating that the driving behaviour b is less risky than the driving behaviour c; if (ζ)bcThe positive and negative values of-1) are equal, which indicates that the driving behavior b and the driving behavior c have the same risk degree and are classified into the same risk level in parallel.
CN202010980521.3A 2020-09-17 2020-09-17 Risk assessment method for distracted driving behaviors Active CN112100857B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010980521.3A CN112100857B (en) 2020-09-17 2020-09-17 Risk assessment method for distracted driving behaviors

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010980521.3A CN112100857B (en) 2020-09-17 2020-09-17 Risk assessment method for distracted driving behaviors

Publications (2)

Publication Number Publication Date
CN112100857A CN112100857A (en) 2020-12-18
CN112100857B true CN112100857B (en) 2022-04-12

Family

ID=73759963

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010980521.3A Active CN112100857B (en) 2020-09-17 2020-09-17 Risk assessment method for distracted driving behaviors

Country Status (1)

Country Link
CN (1) CN112100857B (en)

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113859249A (en) * 2021-10-22 2021-12-31 合众新能源汽车有限公司 Intelligent feedback method and system based on vehicle running environment
CN114241753B (en) * 2021-12-03 2022-11-01 东南大学 Road safety evaluation method and system based on multi-dimensional influence factors
CN115099128B (en) * 2022-05-30 2023-07-07 同济大学 Abnormal driving behavior identification and cause analysis method and system

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109242227A (en) * 2017-07-10 2019-01-18 卢照敢 The driving risk and assessment models of car steering behavior
CN109690606A (en) * 2016-09-10 2019-04-26 瑞士再保险有限公司 Scoring driving measurement, triggering and the system intelligent, adaptive, based on telematics and its corresponding method signaled are carried out to the automatic guidance operation of associated automated system
US10407079B1 (en) * 2017-01-19 2019-09-10 State Farm Mutual Automobile Insurance Company Apparatuses, systems and methods for determining distracted drivers associated with vehicle driving routes
CN111274881A (en) * 2020-01-10 2020-06-12 中国平安财产保险股份有限公司 Driving safety monitoring method and device, computer equipment and storage medium

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109690606A (en) * 2016-09-10 2019-04-26 瑞士再保险有限公司 Scoring driving measurement, triggering and the system intelligent, adaptive, based on telematics and its corresponding method signaled are carried out to the automatic guidance operation of associated automated system
US10407079B1 (en) * 2017-01-19 2019-09-10 State Farm Mutual Automobile Insurance Company Apparatuses, systems and methods for determining distracted drivers associated with vehicle driving routes
CN109242227A (en) * 2017-07-10 2019-01-18 卢照敢 The driving risk and assessment models of car steering behavior
CN111274881A (en) * 2020-01-10 2020-06-12 中国平安财产保险股份有限公司 Driving safety monitoring method and device, computer equipment and storage medium

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
Association between cellphone use while driving legislation and self-reported behaviour among adult drivers in USA: a cross-sectional study;Toni Marie Rudisill等;《BMJ OPEN》;20190630;全文 *
Study on the Impact Degrees of Several Driving Behaviors When Driving While Performing Secondary Tasks;Lisheng Jin等;《IEEE》;20181031;全文 *
风险驾驶行为识别及干预研究综述;李艳等;《汽车与安全》;20200315;全文 *

Also Published As

Publication number Publication date
CN112100857A (en) 2020-12-18

Similar Documents

Publication Publication Date Title
CN112100857B (en) Risk assessment method for distracted driving behaviors
Chen et al. Key feature selection and risk prediction for lane-changing behaviors based on vehicles’ trajectory data
CN112487617B (en) Collision model-based risk prevention method, device, equipment and storage medium
Brombacher et al. Driving event detection and driving style classification using artificial neural networks
Wada et al. Characterization of expert drivers' last-second braking and its application to a collision avoidance system
Kovaceva et al. Identification of aggressive driving from naturalistic data in car-following situations
CN111813083B (en) Scene task complexity quantitative evaluation method
CN106781454B (en) The appraisal procedure and device of driving behavior
Muehlfeld et al. Statistical behavior modeling for driver-adaptive precrash systems
CN112508054B (en) Driving model training method, device, equipment and medium
Phan et al. Estimation of driver awareness of pedestrian based on Hidden Markov Model
JP6694995B1 (en) Lane change evaluation device and lane change evaluation method
Büyükyildiz et al. Identification of the driving style for the adaptation of assistance systems
Del Re et al. Implementation of road safety perception in autonomous vehicles in a lane change scenario
Akinapalli et al. Classification of motorized two-wheeler riders’ acceleration and deceleration behavior through short-term naturalistic riding study
CN115576224A (en) Simulation test and evaluation method for adaptive cruise control system
Beggiato et al. Sequence analysis of glance patterns to predict lane changes on urban arterial roads
CN112373482B (en) Driving habit modeling method based on driving simulator
Xu et al. Estimating risk levels of driving scenarios through analysis of driving styles for autonomous vehicles
Glaser et al. An Integrated Driver-Vehicle-Environment (I-DVE) model to assess crash risks
O'Brien et al. Analysis of driving data for autonomous vehicle applications
Li et al. Driving feature extraction from high and low skilled drivers in curve sections based on machine learning
SUZUKI et al. Analysis of the driving behavior when the system safety level during automated driving is presented to drivers
Cisneros Lombera et al. Driver Model Using Fuzzy Logic for Virtual Validation
CN116776204B (en) Driver risk sensitivity differentiation characterization method, device, equipment and medium

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant