Traffic simulation model calibration method considering traffic conflict index distribution problems

Abstract

The invention discloses a traffic simulation model calibration method considering traffic conflict index distribution problems. According to the method, an efficiency index and a safety index are considered at the same time, an appropriate intersection efficiency index is selected to serve as a constraint condition of a first-round calculation target and a second-round calculation target of traffic simulation modeling, and an appropriate intersection safety index is selected to serve as the second-round calculation target of traffic simulation modeling; by calculating a standard difference ofsimulation indexes and selecting an appropriate confidence interval to estimate the minimum number of times of simulation needed for an established simulation model, the number of simulation tests ofparameter correction is effectively reduced; an ANOVA test or a correlation test are used to screen out simulation parameters with remarkable influences on the efficiency index and the safety index; amulti-constraint calibration optimization problem is established, and a specific strategy and a process for calibrating model parameters by use of a genetic algorithm are proposed; a fitness parameter selection scheme is proposed; a model constraint condition selection scheme is proposed; and model calculation is performed through traffic conflict parameter distribution.

Description

technical field [0001] The invention relates to a traffic simulation model calibration method considering the multi-constraint optimization problem of traffic conflict index distribution, and belongs to the technical field of intelligent traffic. Background technique [0002] The microscopic traffic simulation model uses a large number of independent parameters to describe the operation of the traffic system, traffic flow characteristics, and driver behavior, and the parameter values ​​of the model have a greater impact on the simulation results. Microscopic traffic simulation model calibration refers to the adjustment of model parameters to improve the accuracy of the model fitting the real traffic operation conditions and vehicle driving behavior. The goal of model calibration is to minimize the relative error between the index value obtained by simulation and the measured value to evaluate the actual traffic operation status, so as to reproduce the actual traffic operatio...

AI Technical Summary

Problems solved by technology

[0009] 1. The traditional simulation model calibration method based on efficiency index does not consider safety index;

Images