CN104392328A - Uncertainty evaluation method of vehicle-pedestrian traffic accident - Google Patents

Abstract

The invention discloses an uncertainty evaluation method of a vehicle-pedestrian traffic accident. A probability density function of a vehicle-pedestrian state parameter before crash is recognized through the probability density function PDF of a measurement parameter after the crash. According to the method, the vehicle-pedestrian state parameter range before the crash is effectively recognized, and the probability of occurrence of the state parameters is objectively given, so that important basic data and basis are provided for more rationally analyzing and evaluating the traffic event.

Description

A kind of uncertainty assessment method of vehicle-pedestrian's traffic hazard

Technical field

The invention belongs to field of traffic, particularly relate to a kind of uncertainty assessment method of vehicle-pedestrian's traffic hazard, particularly in a kind of vehicle-pedestrian's traffic hazard, collide probabilistic identification and the evaluation method of front state parameter.

Background technology

The frequent generation of road traffic accident has become very serious social concern, in order to protect pedestrains safety, must make a careful inquiry to accident and analyse scientifically.Therefore, after traffic hazard occurs, how reconstruct the state of collision vehicle in front-pedestrian according to scene of the accident situation, for analyzing the collision rule of vehicle with pedestrian, rationally carrying out traffic hazard qualification and divisions of responsibility can etc. and there is important research value and social effect.

Traditional accident reconstruction is all carry out based on deterministic models, and traditional accident reconstruction method can only provide state parameter before one group of collision determined.Such as, but traffic hazard is the coefficient result of many factors, more or less there is various uncertain factor, comprising: the friction, braking time etc. on the error of field measurement data, vehicle and road surface.If do not consider the coupling of these uncertain factors in accident reconstruction, then may cause the front state parameter of collision in the vehicle-pedestrian's traffic hazard identified and actual deviation comparatively greatly, even correctly cannot carry out accident responsibility division.

Summary of the invention

For the difficulty of accident reconstruction under consideration uncertain factor, the object of the invention is to propose a kind of probability density function (Probabilitydensity function utilizing collision rift vehicle-pedestrian to measure parameter, PDF) identify the method for the PDF of collision vehicle in front-pedestrian's state parameter, thus provide foundation for the rational evaluation of traffic hazard.

According to an aspect of the present invention, propose a kind of uncertainty assessment method of vehicle-pedestrian's traffic hazard, comprising:

Steps A: the first state parameter obtaining the collision rift vehicle-pedestrian measured, and determine first each rank square of the first probability density function PDF of described first state parameter, wherein said first state parameter comprises: the height of the quality of the position of the Distance geometry angle of all directions of collision rift pedestrian relative vehicle, braking mark length, vehicle body local deformation, vehicle and pedestrian and centroid position, pedestrian;

Step B: utilize reliability degree calculation method, according to the vehicle-pedestrian collision's realistic model built by described first state parameter, calculate the 2nd PDF of first state parameter of collision rift vehicle-pedestrian, and determine second each rank square of the 2nd PDF of described first state parameter, wherein said reliability degree calculation method comprises: FOSM, second-order second-moment method, JC method and Monte Carlo method;

Step C: according to the reverse objective function constructed by described first each rank square and described second each rank square, determine to collide second state parameter of vehicle in front-pedestrian, wherein said second state parameter comprises: the friction factor on vehicle and ground in the relative angle of the position of the speed of Vehicle Speed, pedestrian's walking, pedestrian and vehicle collision, pedestrian and vehicle, the time of car brakeing, braking procedure.

Wherein, steps A comprises:

In the scene of a traffic accident, first state parameter of repetitive measurement collision rift vehicle and pedestrian;

Statistics obtains a PDF of described first state parameter;

Integral Processing is carried out to a described PDF, to determine first each rank square of a described PDF.

Wherein, step B comprises:

According to described first state parameter, build vehicle-pedestrian collision's realistic model;

Utilize reliability degree calculation method, according to described vehicle-pedestrian collision's realistic model, calculate the 2nd PDF of first state parameter of collision rift vehicle-pedestrian;

Integral Processing is carried out to described 2nd PDF, to determine second each rank square of described 2nd PDF.

Further, vehicle-pedestrian collision's realistic model is Multi-body dynamic model.

Wherein, step C comprises:

According to the quadratic sum of the difference of described first each rank square and described second each rank square, set up the reverse objective function of vehicle-pedestrian's accident reconstruction;

Utilize genetic algorithm, described reverse objective function is optimized and solves;

When described reverse objective function meets optimal conditions, determine the 3rd PDF of the second state parameter colliding vehicle in front-pedestrian;

According to described 3rd PDF, determine probability density distribution and the span thereof of described second state parameter.

According to a further aspect in the invention, a kind of uncertainty assessment method of vehicle-pedestrian's traffic hazard is provided, comprises the steps:

Step 1: according to scene of a traffic accident situation, repetitive measurement collision rift vehicle-pedestrian state parameter, and statistics obtains the PDF of the state parameter measured;

Step 2: according to vehicle and pedestrian's type, sets up the vehicle-pedestrian collision realistic model corresponding with traffic hazard;

Step 3: parametric description is carried out to the PDF of collision vehicle in front-pedestrian's state parameter, and the span of setup parameter and initial value;

Step 4: based on the vehicle-pedestrian collision's realistic model set up, repeatedly call the reliability degree calculation method such as first-order reliability method, solve the PDF obtaining collision rift vehicle-pedestrian state parameter;

Step 5: the PDF of the collision rift vehicle-pedestrian state parameter measured in the PDF of the collision rift vehicle-pedestrian state parameter calculated in step 4 and step 1 is contrasted, sets up the reverse objective function of vehicle-pedestrian's accident reconstruction;

Step 6: utilize optimization method to carry out iterative to reverse objective function, if do not meet iteration convergence condition, the next generation then producing the PDF describing collision vehicle in front-pedestrian's state parameter within the scope of parameter value is newly worth, then step 4 is gone back to, if meet the condition of convergence, then carry out step 7;

Step 7: utilize the PDF optimizing and obtain parameter value structure collision vehicle in front-pedestrian's state parameter, realize the uncertainty assessment of traffic hazard.

Wherein, collision rift vehicle-pedestrian state parameter comprises: the height etc. of the quality of the position of the Distance geometry angle of all directions of collision rift pedestrian relative vehicle, braking mark length, vehicle body local deformation, vehicle and pedestrian and centroid position, pedestrian.

Wherein, collide the state parameter of vehicle in front-pedestrian to comprise: the friction factor etc. on vehicle and ground in the relative angle of the position of the speed of Vehicle Speed, pedestrian's walking, pedestrian and vehicle collision, pedestrian and vehicle, the time of car brakeing, braking procedure.

Wherein, the PDF of collision vehicle in front-pedestrian's state parameter is described by parameterized form, and recognition result is not common one group of determined value, but the possible range of collision vehicle in front-pedestrian's state parameter and probability density distribution thereof.

Alternatively, in step 2 above, vehicle-pedestrian collision's realistic model is Multi-body dynamic model.

Alternatively, in above-mentioned steps 4, the reliability degree calculation method adopted for the PDF obtaining collision rift vehicle-pedestrian state parameter comprises: FOSM, second-order second-moment method, JC method and Monte Carlo method etc.

Alternatively, in above-mentioned steps 5, when setting up reverse objective function, the square or semi-invariant etc. of finite point and PDF in the probability density curve that the data of contrast comprise the PDF of the PDF of the collision rift vehicle-pedestrian state parameter of measurement and the collision rift vehicle-pedestrian state parameter of calculating.

Alternatively, in above-mentioned steps 6, utilize genetic algorithm as optimization method, wherein, realized the generation of new value of future generation by the operation such as selection, intersection, variation of genetic algorithm.

As from the foregoing, a kind of vehicle-pedestrian's traffic hazard uncertainty assessment method that the present invention proposes, consider uncertain factor (the uncertain factor particularly measured) impact on accident reconstruction result, be intended to carry out parametric description to the PDF of collision vehicle in front-pedestrian's state parameter.By measurement of comparison to the PDF of collision rift vehicle-pedestrian state parameter and the PDF of collision rift vehicle-pedestrian state parameter that calculates set up reverse objective function, and the estimation of PDF utilizing optimization method to realize collision vehicle in front-pedestrian's state parameter, thus provide important foundation data and foundation for more reasonably analysis and inspection traffic hazard.

Accompanying drawing explanation

In order to be illustrated more clearly in the technical scheme of the embodiment of the present invention, be briefly described to the accompanying drawing used required in the embodiment of the present invention below, apparently, accompanying drawing described is below only some embodiments of the present invention, for those of ordinary skill in the art, under the prerequisite not paying creative work, other accompanying drawing can also be obtained according to these accompanying drawings.

Fig. 1 is the process flow diagram of the uncertainty assessment method of vehicle-pedestrian's traffic hazard of the present invention;

Fig. 2 is the flow chart of steps of vehicle-pedestrian's traffic hazard uncertainty assessment in the specific embodiment of the invention;

Fig. 3 (a) is the state of colliding vehicle in front-pedestrian in the present invention;

Fig. 3 (b) is the state of colliding vehicle in front-pedestrian in the present invention;

Fig. 4 (a) is the probability density curve of the collision vehicle in front speed υ identified in the present invention;

Fig. 4 (b) is the probability density curve of the coefficientoffrictionμ on vehicle and ground in the braking procedure identified in the present invention;

Fig. 4 (c) is the probability density curve of the angle [alpha] between collision vehicle in front and pedestrian identified in the present invention;

Fig. 5 (a) is that the X at the collision rift pedestrian relative vehicle front-wheel center of surveying and calculating in the present invention is to the distance l that dishes out _xprobability density curve;

Fig. 5 (b) is that the Y-direction at the collision rift pedestrian relative vehicle front-wheel center of surveying and calculating in the present invention is dished out distance l _yprobability density curve;

Fig. 5 (c) be the collision rift pedestrian of surveying and calculating in the present invention and vehicle front-wheel center X between the probability density curve of angle theta.

Embodiment

Below in conjunction with the accompanying drawing in the embodiment of the present invention, be clearly and completely described the technical scheme in the embodiment of the present invention, obviously, described embodiment is a part of embodiment of the present invention, instead of whole embodiment.Based on the embodiment in the present invention, the every other embodiment that those of ordinary skill in the art obtain under the prerequisite not making creative work, all should belong to the scope of protection of the invention.

In view of traditional accident reconstruction cannot consider the situation of uncertain factor, the present invention takes into full account the impact of uncertain factor on reverse result, do not only give the reasonable value scope of collision vehicle in front-pedestrian's state parameter, and the more important thing is and give its probability density curve, this provides important evidence for more reasonably evaluating traffic hazard.

Further, the present invention carries out parametric description by the PDF of collision vehicle in front-pedestrian's state parameter, and utilizes reliability degree calculation method to solve the PDF of collision rift vehicle-pedestrian state parameter, to improve efficiency and the precision of the uncertain reconstruction of traffic hazard.

In addition, the present invention can the PDF of state parameter before the multiple collision of disposable acquisition, comprises the PDF of the parameters such as the friction factor on vehicle and ground in Vehicle Speed, the speed of pedestrian's walking, the position of pedestrian and vehicle collision, the relative angle of pedestrian and vehicle, time of car brakeing, braking procedure.

Below in conjunction with accompanying drawing 1, to consider that measuring the uncertain lower vehicle-pedestrian traffic accident of response is evaluated as example, is described in detail to the evaluation method of vehicle-pedestrian's traffic hazard of the present invention.

As shown in Figure 1, the uncertainty assessment method of vehicle-pedestrian's traffic hazard of the present invention comprises the following steps A-C.

Steps A: the first state parameter obtaining the collision rift vehicle-pedestrian measured, and determine first each rank square of a PDF of the first state parameter.

Here, the first state parameter comprises the height of the Distance geometry angle of all directions of collision rift pedestrian relative vehicle, braking mark length, the position of vehicle body local deformation, the quality of vehicle and pedestrian and centroid position, pedestrian.

Particularly, in the scene of a traffic accident, first state parameter of repetitive measurement collision rift vehicle and pedestrian; Then, a PDF of acquisition first state parameter is added up; And Integral Processing is carried out to a PDF, to determine first each rank square of a PDF.

Step B: utilize reliability degree calculation method, according to the vehicle-pedestrian collision's realistic model built by the first state parameter, calculate the 2nd PDF of first state parameter of collision rift vehicle-pedestrian, and determine second each rank square of the 2nd PDF of the first state parameter.

Specifically, according to the first state parameter, build vehicle-pedestrian collision's realistic model.Here vehicle-pedestrian collision's realistic model can be Multi-body dynamic model.Then, utilize reliability degree calculation method, according to vehicle-pedestrian collision's realistic model, calculate the 2nd PDF of first state parameter of collision rift vehicle-pedestrian.

Wherein, reliability degree calculation method comprises FOSM, second-order second-moment method, JC method and Monte Carlo method.Integral Processing is carried out to described 2nd PDF, to determine second each rank square of described 2nd PDF.

Step C: according to the reverse objective function constructed by first each rank square and second each rank square, determines the second state parameter colliding vehicle in front-pedestrian.

Wherein, the second state parameter comprises the friction factor on vehicle and ground in Vehicle Speed, the speed of pedestrian's walking, the position of pedestrian and vehicle collision, the relative angle of pedestrian and vehicle, time of car brakeing, braking procedure.

Particularly, according to the quadratic sum of the difference of first each rank square and second each rank square, the reverse objective function of vehicle-pedestrian's accident reconstruction is set up; Utilize genetic algorithm, reverse objective function is optimized and solves; When reverse objective function meets optimal conditions, determine the 3rd PDF of the second state parameter colliding vehicle in front-pedestrian; According to the 3rd PDF, determine probability density distribution and the span thereof of the second state parameter.

As from the foregoing, a kind of vehicle-pedestrian's traffic hazard uncertainty assessment method that the present invention proposes, consider uncertain factor (the uncertain factor particularly measured) impact on accident reconstruction result, be intended to carry out parametric description to the PDF of collision vehicle in front-pedestrian's state parameter.By measurement of comparison to the PDF of collision rift vehicle-pedestrian state parameter and the PDF of collision rift vehicle-pedestrian state parameter that calculates set up reverse objective function, and utilize optimization method to realize the estimation of the PDF to collision vehicle in front-pedestrian's state parameter, thus provide important evidence for more reasonably analyzing traffic hazard.

After vehicle-pedestrian's traffic hazard occurs, there is determinacy scarcely in the measurement due to scene of the accident data, in order to consider the impact of this kind of uncertainty on vehicle-pedestrian collision's accident reconstruction, can according to scene of the accident situation (namely, the collision rift vehicle-pedestrian state parameter measured) first set up vehicle-pedestrian collision's realistic model, and parametric description is carried out to the PDF of collision vehicle in front-pedestrian's state parameter; Utilize reliability degree calculation method can obtain the PDF of collision rift vehicle-pedestrian state parameter on this basis; Measurement of comparison with each rank square of collision rift vehicle-pedestrian state parameter PDF calculated, form reverse objective function, each rank square of PDF can by obtaining PDF integration here; Utilize genetic algorithm to be optimized to solve, thus obtain the PDF colliding vehicle in front-pedestrian's state parameter.

Fig. 2 shows the steps flow chart of vehicle-pedestrian's traffic hazard uncertainty assessment method in the specific embodiment of the invention, and concrete implementation step is as follows:

Step 1: after vehicle-pedestrian's traffic hazard occurs, the state parameter Z of repetitive measurement collision rift vehicle-pedestrian ^u, the X comprising collision rift pedestrian relative vehicle front-wheel center to dish out distance l to Y-direction _x, l _y, and pedestrian and vehicle front-wheel center X between angle theta, and statistics obtains l _x, l _ywith the probability density curve of avatars as shown in Fig. 5 (a)-5 (c) of the PDF of θ, PDF.Meanwhile, can according to l _x, l _ycorresponding l is calculated with the PDF of θ _x, l _ywith each rank square of the PDF of θ wherein i represents i-th state parameter, and j represents j rank square, and m represents that each rank square obtains by measuring.

Step 2: according to the state parameter Z of the collision rift vehicle-pedestrian measured ^u, set up the vehicle-pedestrian collision's accident multi-rigid body realistic model as shown in Fig. 3 (a)-3 (b).The object of carrying out uncertain reconstruction to traffic hazard is the state parameter before identifying collision, and the state parameter wherein before collision comprises vehicle and the coefficientoffrictionμ on ground and the angle [alpha] between pedestrian and vehicle in collision vehicle in front speed υ, braking procedure;

Step 3: utilize formula (1) to describe the PDF of the state parameter of collision vehicle in front-pedestrian:

$P_{X^u}^{\mathrm{\λ}}\left(x^u\right)=\{\frac{{\mathrm{\κ}}_{\mathrm{\λ}}}{\sqrt{b_1^2-{4b}_2(b_0-x^u)}}\·{(1-{\left(\frac{{-b}_1+\sqrt{b_1^2-{4b}_2(b_0-x^u)}}{{2b}_2}\right)}^2)}^{\mathrm{\λ}-\frac{1}{2}}---\left(1\right)$

X in formula ^ufor colliding the state parameter of vehicle in front-pedestrian, namely represent υ, μ and α; Here, b ₀, b ₁, b ₂, λ, κ _χfor describing X ^uthe coefficient of parametrization PDF.If can identify and obtain parameter b ₀, b ₁, b ₂, λ, then obtain colliding the PDF of vehicle in front-pedestrian's state parameter by formula (1).

Step 4: as any given one group of b ₀, b ₁, b ₂, λ time, based on set up vehicle-pedestrian collision's accident multi-rigid body realistic model, repeatedly call the reliability degree calculation method of such as first-order reliability method method, calculate the PDF of collision rift vehicle-pedestrian state parameter, detailed process is as described below.

First, the collision rift vehicle-pedestrian state parameter Z will measured ^uspan carry out n decile, getting step-length is Δ z=(z _u-z _l)/n, here z _land z _urespectively define Z ^uthe lower limit of value and the upper limit, n be more than or equal to 2 positive integer; Secondly, for kth step, ε=z is got _l+ k Δ z, conformation function function G ^u=g (X ^u)-ε, wherein g (X ^u) represent the numerical model corresponding according to vehicle-pedestrian collision's accident multi-rigid body realistic model; Again, utilize the reliability degree calculation method of FOSM, calculate collision rift vehicle-pedestrian state parameter Z ^uaccumulated probability F (z ^u), namely utilize first-order reliability method method to solve formula (2):

$F\left(z^u\right)={\∫}_{G^u\≤0}{P}_{X^u}^{\mathrm{\λ}}\left(x^u\right){\mathrm{dx}}^u---\left(2\right)$

Here, k, from 0 to n, repeatedly calls the PDF that first-order reliability method method can calculate collision rift vehicle-pedestrian state parameter, and calculates each rank square of the PDF of corresponding collision rift vehicle-pedestrian state parameter on this basis wherein i represents i-th state parameter, and j represents j rank square, and c represents that each rank square is by calculating.

Step 5: by calculate collision rift vehicle-pedestrian state parameter PDF corresponding to each rank square with each rank square corresponding to the collision rift vehicle-pedestrian state parameter PDF measured contrast, and using the quadratic sum of the two difference as reverse objective function, shown in (3):

$\begin{array}{cc}\min & \underset{i=1}{\overset{n_2}{\mathrm{\Σ}}}\underset{j=1}{\overset{n_1}{\mathrm{\Σ}}}{(M_{\mathrm{ij}}^c\left(X^u\right)-M_{\mathrm{ij}}^m)}^2---\left(3\right)\end{array}$

In formula with be respectively measurement with calculate collision rift vehicle-pedestrian state parameter PDF corresponding to each rank square; n ₁for the order of square, in this example, get 4; n ₂for the number of collision rift vehicle-pedestrian state parameter, be 3 in this example.

Step 6: utilize genetic algorithm to be optimized reverse objective function and solve, if do not meet iteration convergence condition, then produces parameter b by operations such as selection, intersection, variations ₀, b ₁, b ₂, λ the next generation be newly worth, then go back to step 4; If meet the condition of convergence, then carry out step 7.

Step 7: will the parameter value b as shown in table 1 obtained be optimized ₀, b ₁, b ₂, λ substitutes into formula (1), obtains the probability density curve of the collision vehicle in front-pedestrian's state parameter υ, μ and α identified, as shown in Fig. 4 (a) to 4 (c).Fig. 4 (a) does not only give the reasonable value scope of collision vehicle in front-pedestrian's state parameter to 4 (c) result, and the more important thing is and give its probability density curve, and this provides important evidence for more reasonably evaluating traffic hazard.Table 1 result is substituted into step 4, obtains the PDF of the collision rift vehicle-pedestrian state parameter of optimum calculating, as shown in Fig. 5 (a) to 5 (c).

Table 1 collides the recognition result of vehicle in front-pedestrian's state parameter PDF

If Fig. 5 is a) to shown in 5 (c), the PDF of collision rift vehicle-pedestrian state parameter calculated is substantially identical with both PDF of the collision rift vehicle-pedestrian state parameter of measurement, thus demonstrates correctness and the validity of the evaluation method of vehicle-pedestrian's traffic hazard of the present invention.

As from the foregoing, a kind of vehicle-pedestrian's traffic hazard uncertainty assessment method that the present invention proposes, consider uncertain factor (the uncertain factor particularly measured) impact on accident reconstruction result, be intended to carry out parametric description to the PDF of collision vehicle in front-pedestrian's state parameter.By measurement of comparison to the PDF of collision rift vehicle-pedestrian state parameter and the PDF of collision rift vehicle-pedestrian state parameter that calculates set up reverse objective function, and the estimation of PDF utilizing optimization method to realize collision vehicle in front-pedestrian's state parameter, thus provide important foundation data and foundation for more reasonably analysis and inspection traffic hazard.

The above; be only the specific embodiment of the present invention, but protection scope of the present invention is not limited thereto, is anyly familiar with those skilled in the art in the technical scope that the present invention discloses; change can be expected easily or replace, all should be encompassed within protection scope of the present invention.Therefore, protection scope of the present invention should described be as the criterion with the protection domain of claim.

Claims (5)

1. a uncertainty assessment method for vehicle-pedestrian's traffic hazard, comprising:

Steps A: the first state parameter obtaining the collision rift vehicle-pedestrian measured, and determine first each rank square of the first probability density function PDF of described first state parameter, wherein said first state parameter comprises: the height of the quality of the position of the Distance geometry angle of all directions of collision rift pedestrian relative vehicle, braking mark length, vehicle body local deformation, vehicle and pedestrian and centroid position, pedestrian;

Step B: utilize reliability degree calculation method, according to the vehicle-pedestrian collision's realistic model built by described first state parameter, calculate the 2nd PDF of first state parameter of collision rift vehicle-pedestrian, and determine second each rank square of the 2nd PDF of described first state parameter, wherein said reliability degree calculation method comprises: FOSM, second-order second-moment method, JC method and Monte Carlo method;

Step C: according to the reverse objective function constructed by described first each rank square and described second each rank square, determine to collide second state parameter of vehicle in front-pedestrian, wherein said second state parameter comprises: the friction factor on vehicle and ground in the relative angle of the position of the speed of Vehicle Speed, pedestrian's walking, pedestrian and vehicle collision, pedestrian and vehicle, the time of car brakeing, braking procedure.

2. evaluation method according to claim 1, is characterized in that, described steps A comprises:

In the scene of a traffic accident, first state parameter of repetitive measurement collision rift vehicle and pedestrian;

Statistics obtains a PDF of described first state parameter;

Integral Processing is carried out to a described PDF, to determine first each rank square of a described PDF.

3. evaluation method according to claim 1, is characterized in that, described step B comprises:

According to described first state parameter, build vehicle-pedestrian collision's realistic model;

Utilize reliability degree calculation method, according to described vehicle-pedestrian collision's realistic model, calculate the 2nd PDF of first state parameter of collision rift vehicle-pedestrian;

Integral Processing is carried out to described 2nd PDF, to determine second each rank square of described 2nd PDF.

4. evaluation method according to claim 3, is characterized in that, described vehicle-pedestrian collision's realistic model is Multi-body dynamic model.

5. method according to any one of claim 1 to 4, is characterized in that, step C comprises:

According to the quadratic sum of the difference of described first each rank square and described second each rank square, set up the reverse objective function of vehicle-pedestrian's accident reconstruction;

Utilize genetic algorithm, described reverse objective function is optimized and solves;

When described reverse objective function meets optimal conditions, determine the 3rd PDF of the second state parameter colliding vehicle in front-pedestrian;

According to described 3rd PDF, determine probability density distribution and the span thereof of described second state parameter.