CN1932923A - Road traffic accident multi-happening section identifying method - Google Patents

Abstract

The invention is concerned with the distinguishing method for the road with more traffic accident. The invention uses double target of alpha-gamma to distinguish the road with more traffic accident, assigns the remarkable level alpha, according to the Poisson probability formula or the frequency accumulation chart, after computes the critical value gamma of the high accident number and the critical value R of the high accident rate, compares the gamma and the R of the un-distinguish road to get the distinguishing result of the road with more traffic accident.

Description

Road traffic accident multi-happening section identifying method

Technical field

The present invention relates to the traffic safety field, more relate to a kind of complete reliable road traffic accident multi-happening section identifying method in this field.

Background technology

The first step in road safety improvement project also is a step of most critical, is exactly the position of determining the accident-prone road section of urgent need improvement.Present most widely used method is the casualty data statistical analysis method, and different according to casualty data statistical treatment discriminant criterion and method of discrimination be divided into various method again, but these methods all have separately applicable elements and deficiency.

The absolute number method is divided into accident frequency method and accident rate method again.The accident frequency method is not considered factors such as the volume of traffic, severity of injuries, uses separately the reflection accident situation is had one-sidedness, needs a large amount of and long investigation statistics; Simultaneously the accident number of times how much lack theoretic tightness, dividing the accident-prone road section, influenced by artificial subjective factor bigger.The accident rate method has been considered the influence of the volume of traffic and two key factors of road section length on the basis of accident frequency method, but does not find the foundation of tight division accident-prone road section yet in theory.

The probability statistics distribution is theoretical tight, but practical condition harshness.The one, the accident number that counts by average road section length is on some roads and do not meet Poisson distribution, therefore can not be applied widely; In addition, can not determine a numerical value stereotypedly, consider the difference that the whole nation is regional, regional.

The cumulative frequency method is not considered the order of severity of volume of traffic factor and accident, though the catastrophe point with the cumulative frequency matched curve has certain mathematical theory foundation as the standard that defines the accident-prone road section simultaneously, but this method only is suitable for the similar matched curve of road conditions has the situation of flex point (sudden change), can't explain its practical significance in practice.

Summary of the invention

The purpose of this invention is to provide a kind of road traffic accident multi-happening section identifying method, it is more complete, reliable to make road traffic accident multi-happening section differentiate.

The present invention is a kind of discrimination method of road traffic accident multi-happening section, it is characterized in that: it may further comprise the steps:

1) utilize the checkout equipment that is arranged on the road to detect road traffic accident number and the traffic data of collecting highway section to be identified, the traffic hazard number is meant that highway section to be identified is in a measurement period and the corresponding traffic hazard number of times of highway mileage number, with the volume of traffic of traffic hazard number of times, be the road accident rate in this highway section divided by this highway section;

2) utilize computing machine to read road traffic accident number and traffic data, utilize statistical analysis software differentiation traffic hazard number and accident rate whether to meet Poisson distribution, can utilize the Poisson distribution Chi-square Test module (Chi-square) among the existing SAS of statistical software to realize;

3) when traffic hazard number and accident rate meet Poisson distribution, given level of signifiance α calculates high accident number of times critical value γ and high accident rate critical value R under the level of signifiance α according to Poisson new probability formula (1);

$P\{X\≥\mathrm{\γ}\}=1-\underset{K=0}{\overset{\mathrm{\γ}-1}{\mathrm{\Σ}}}\frac{{\mathrm{\λ}}^K}{K!}{e}^{-\mathrm{\λ}}=\underset{k=r}{\overset{\∞}{\mathrm{\Σ}}}\frac{{\mathrm{\λ}}^K}{K!}{e}^{-\mathrm{\λ}}\≤\mathrm{\α}---\left(1\right)$

Wherein: X-accident frequency,

γ-accident number of times critical value,

λ-Parameter for Poisson Distribution, K is a nonnegative integer,

When traffic hazard number and accident rate do not meet Poisson distribution, statistical method according to routine, calculate the frequency of identical accident number and identical accident rate generation respectively, and according to accident number and accident rate from small to large order respectively with its frequency accumulative total addition, obtain accumulative total accident frequency and accumulative total accident rate frequency, behind the given level of signifiance α, cumulative frequency pairing accident number and the accident rate equivalent with (1-α) are critical value γ, R;

4) traffic hazard number of times and the road accident rate with highway section to be identified compares with level of signifiance α corresponding down critical value γ and R, if traffic hazard number of times and road accident rate are all greater than level of signifiance α corresponding down critical value γ and R, perhaps the two one of greater than the respective value under the level of signifiance α, judge that promptly this highway section is the accident-prone road section.

The described checkout equipment that is arranged on the road is unit, MetroCount5600 roadside or the vehicle magnetic reflection magnitude of traffic flow and analyser NC97.

The present invention uses the two indexs of α-γ to differentiate road traffic accident multi-happening section, and its beneficial effect is mainly reflected in:

1. when accident number and accident rate meet Poisson distribution, consider the influence of the volume of traffic simultaneously, avoided the subjectivity of only differentiating with the accident number of times, thereby avoided the highway section erroneous judgement of the low accident rate of high accident number of times is decided to be the accident-prone road section, strengthened discriminating simultaneously the low high accident rate of accident number of times highway section.

2 when accident number and accident rate do not meet Poisson distribution, and this method has been widened traffic hazard cumulative frequency method.

Description of drawings

Fig. 1 is an operational flowchart of the present invention;

Traffic hazard cumulative frequency figure among Fig. 2 embodiment 2.

Embodiment

Embodiment 1 accident number and accident rate meet the example of Poisson distribution

The traffic hazard number is meant highway section to be identified in a measurement period and the corresponding traffic hazard number of times of highway mileage number, is statistical unit with the accident number that takes place on the long 1Km highway section in this example, and road type is an expressway.The division of other road types is as shown in table 1.The corresponding second class road type urban road, when differentiating road traffic accident multi-happening section with " length less than L and the highway section of the number 〉=2 that had an accident " replacements " length 1Km highway section ", the constant multiple decision criteria of putting of urban road that is of other steps.L wherein is meant the city intersection spacing: subsidiary road 30m, major trunk roads 60m.

Table 1 road type

Type	Roadway characteristic
		1 2 3 4	Expressway comprises that highway, Class I highway, secondary expressway urban road do not comprise suburbs and counties under the jurisdiction of a large city's highway, it is characterized in that crossing density is big by two, Class III highway level Four, substandard highway

1) given level of signifiance α=0.05 time, the Poisson new probability formula of calculating high accident number of times critical value γ is as shown in Equation (2):

$P\{X\&GreaterEqual;3\}=1-P\{X=0\}-P\{X=1\}-P\{X=2\}=1-\underset{K=0}{\overset{2}{\mathrm{\&Sigma;}}}\frac{{\left(0.8\right)}^k}{K!}{e}^{0.8}=0.0474<0.05---\left(2\right)$

Wherein: X-accident frequency;

γ-high accident number of times critical value;

λ ₁-Parameter for Poisson Distribution, the total number of accident of the interior generation of institute's road type one-period of study obtains λ divided by the type road total kilometrage in the Free Region ₁Estimated value.λ in this example ₁Estimated value is 0.8.

So γ=3, promptly accident number of times critical value is 3.Then the highway section of all accident generation number 〉=3 just needs further to investigate the accident rate situation, thereby determines the accident-prone road section fully.

2) high accident rate critical value R is calculated in given α=0.01 time, as shown in Equation (3):

$P\{Y\&GreaterEqual;4\}=1-P\{Y=0\}-P\{Y=1\}-P\{Y=2\}-P\{Y=3\}=1-\underset{N=0}{\overset{3}{\mathrm{\&Sigma;}}}\frac{{\left(0.8\right)}^N}{N!}{e}^{0.8}=0.0091<0.01---\left(3\right)$

Wherein: the Y-accident rate;

The high accident rate critical value of R-;

λ ₂-Parameter for Poisson Distribution, the interior accident rate of institute's road type one-period of studying obtains λ divided by the type road total kilometrage in the Free Region ₂Estimated value.

λ in this example ₂Estimated value be 0.8.So R=4.

This example adopts for avoiding accident rate numerical value less than normal Form.

Accident number of times 〉=3 when highway section to be identified, accident rate 〉=4 o'clock, we just judge this accident-prone road section, highway section.Avoided the subjectivity of only differentiating with the accident number of times; Consider the influence of the volume of traffic simultaneously, avoided the highway section erroneous judgement of the low accident rate of high accident number of times is decided to be the accident-prone road section, strengthened discriminating simultaneously the low high accident rate of accident number of times highway section.

Embodiment 2 accident numbers and accident rate do not meet the example of Poisson distribution:

Do not meet the situation of Poisson distribution for accident number and accident rate, be divided into some little highway sections waiting to study road by every 1Km length, find out the path section number that the number that has an accident equates, be that a certain accident number is waiting to study the frequency that takes place on the road, try to achieve the frequency of this accident number with this accident frequency divided by total path section number, frequency is added up addition from small to large according to the accident number get traffic hazard and count cumulative frequency.Behind the given level of signifiance α, the cumulative frequency pairing accident number equivalent with (1-α) is critical value γ.The accident number is got accident rate divided by the volume of traffic, find out the path section number that accident rate equates, try to achieve the accident rate observing frequency, from small to large the observing frequency cumulative addition is got the road accident rate cumulative frequency according to accident rate divided by total path section number.Behind the given level of signifiance α, the cumulative frequency pairing accident rate equivalent with (1-α) is critical value R.

The accident number of domestic 109 national highways in Ningxia is analyzed the Poisson distribution that does not show on the probability statistics by statistics, its accident statistics and calculate as shown in table 2ly, and Fig. 2 is traffic hazard number of times frequency accumulative total figure.

Given level of signifiance α=0.05,1-α=0.95 is critical value γ with transverse axis cumulative frequency 95% pairing accident number of times promptly, from table 2 and Fig. 2, gets γ=15.

Given level of signifiance α=0.10,1-α=0.90 is critical value γ with transverse axis cumulative frequency 90% pairing accident number promptly, from table 2 and Fig. 2, gets γ=13.

The computing method of high accident rate critical value R are similar.

After the given level of signifiance 0.05, the highway section of the number of times that has an accident≤15 accounts for 95% of sum, thus these highway sections, accident number of times＞15 as abnormity point, compare the position of finally definite accident-prone road section again with high accident rate critical value R.This method has been widened the cumulative frequency method, no longer is subject to the situation that the cumulative frequency curve has flex point.

Domestic 109 national highway accident statistics and the reckoners in table 2 Ningxia

Accident is counted X	0	1	2	3	4	5	6	7	8	9
												Accident frequency f	8	19	20	41	31	20	18	13	10	8
Frequency (%)	3.3	7.9	8.3	17	12.9	8.3	7.5	5.4	4.1	3.3
												Cumulative frequency (%)	3.3	11.2	19.5	36.5	49.4	57.7	65.2	71	75	78
Accident is counted X	10	11	12	13	14	15	16	17	18	19
												Accident frequency f	10	7	4	4	6	6	4	4	4	4	∑f＝241
Frequency (%)	4.1	2.9	1.7	1.7	2.5	2.5	1.7	1.7	1.6	1.6
												Cumulative frequency (%)	82.1	85	86.7	88.4	90.9	93.4	95.1	96.8	98.4	100

Claims (2)

1. the discrimination method of a road traffic accident multi-happening section is characterized in that, it may further comprise the steps:

1) utilizes the checkout equipment that is arranged on the road to detect road traffic accident number and the traffic data of collecting highway section to be identified,, obtain the road accident rate in this highway section the volume of traffic of traffic hazard number of times divided by this highway section;

2) utilize computing machine to read road traffic accident number and traffic data, and differentiate the traffic hazard number and whether accident rate meets Poisson distribution;

3) when traffic hazard number and accident rate meet Poisson distribution, given level of signifiance α calculates high accident number of times critical value γ and high accident rate critical value R under the level of signifiance α according to Poisson new probability formula (1);

$P\{X\&GreaterEqual;\mathrm{\&gamma;}\}=1-\underset{K=0}{\overset{\mathrm{\&gamma;}-1}{\mathrm{\&Sigma;}}}\frac{{\mathrm{\&lambda;}}^K}{K!}{e}^{-\mathrm{\&lambda;}}=\underset{k=r}{\overset{\&infin;}{\mathrm{\&Sigma;}}}\frac{{\mathrm{\&lambda;}}^K}{K!}{e}^{-\mathrm{\&lambda;}}\&le;\mathrm{\&alpha;}---\left(1\right)$

Wherein: X-accident frequency,

γ-accident number of times critical value,

λ-Parameter for Poisson Distribution, K is a nonnegative integer,

When traffic hazard number and accident rate do not meet Poisson distribution, calculate the frequency of identical accident number and identical accident rate generation respectively, and according to accident number and accident rate from small to large order respectively with its frequency accumulative total addition, obtain accumulative total accident frequency and accumulative total accident rate frequency, behind the given level of signifiance α, cumulative frequency pairing accident number and the accident rate equivalent with (1-α) are critical value γ, R;

4) during given level of signifiance α, the traffic hazard number of times and the road accident rate in highway section to be identified are compared with level of signifiance α corresponding down critical value γ and R, if traffic hazard number of times and road accident rate are all greater than level of signifiance α corresponding down critical value γ and R, perhaps the two one of greater than the respective value under the level of signifiance α, judge that promptly this highway section is the accident-prone road section.

2. the discrimination method of a kind of road traffic accident multi-happening section according to claim 1 is characterized in that the described checkout equipment that is arranged on the road is unit, MetroCount5600 roadside or the vehicle magnetic reflection magnitude of traffic flow and analyser NC97.

Images