Accident black-spot identification optimizing method

Abstract

The invention discloses an accident black-spot identification optimizing method. Improvement is made based on an empirical Bayes method, and an accident model is constructed to obtain a mean value and a variance of before-accident distribution, so the shortcoming that a great number of reference road units are needed in the empirical Bayes method is overcome. On the part of an accident regression model, a variable is an expected value of the number of accidents happening on a certain road unit in a certain period and independent variables are values representing the traffic of the road unit and road characteristics, and parameters of the before-accident distribution can be obtained by using the expected value of the number of the accidents and the variance so as to determine the before-accident distribution. Simultaneously, through optimization, expectation obtained from after-accident distribution consists of two parts which are a predicted value and a correction term of the predicted value respectively, so the correction of the predicted value can be obtained. Accident back-spot identification results calculated by using the method are more accurate than those calculated by the conventional method, so casualties, vehicle damages, road property losses and the like caused by the accidents are reduced.

Description

technical field [0001] The invention relates to an optimization method for identifying accident black spots, and designs and implements a multiple regression empirical Bayesian safety evaluation method. It belongs to the field of traffic safety. Background technique [0002] The optimization method of accident black spot identification is an important key technology of highway traffic safety. At present, the level of research in this area in our country is uneven. At present, there are many identification methods for accident black spots in China, and the method commonly used by people is the empirical Bayesian method, because the Bayesian method fully considers the factors affecting accidents, and the expected number of accidents obtained from this is more accurate, which is beneficial to accident detection. Multiple point identification. [0003] In 1992, Hauer proposed that the empirical Bayesian method can overcome the regression effect of observed traffic accidents. ...

AI Technical Summary

Problems solved by technology

We must strictly select the reference road unit to ensure that the reference road unit has the same characteristics as the road unit whose safety we want to study. The higher the similarity requirement, the more difficult it is to find a sufficiently similar road unit. In fact, this It is difficult to achieve a large enough number of highly similar road units

In view of the above problems, how to overcome the defect that the empirical Bayesian method requires a large amount of road unit data is an urgent problem to be solved in the identification of accident black spots

Images