A simple approach to traffic density estimation by using Kernel Density Estimation

It is so crucial to know traffic density to manage traffic effectively and act on roads accordingly for safer driving. In this paper, Kernel Density Estimation is used to estimate traffic density by assuming velocity information is known for a given region. Also, kernel weights will be found by using Kolmogorov-Smirnov Tests, since its complexity is less than other methods, and linear Least Square, because it makes approach simpler. We are able to estimate the density with the proposed method and the error between CDF is so small, for an example system which has twelve weights, the error is 8.7667 × 10^-7. By using our approach, methods can be developed to predict traffic density.