Estimating Suitable Probability Distribution Function of Risk for Different Traffic States

摘要

Generally, time to collision (TTC) is used to quantify traffic risk. In this paper, we aimed at studying suitable probability distribution of TTC under different traffic states. By using maximum likelihood estimation (MLE), the parameters of different states are found. Normal distribution and generalized extreme value distribution are used to assume probability distribution function, and chi-squared test is used to test the results. It is found that the normal distribution is more suitable for the data.