The Prediction Of Highway Traffic Volume And Distribution

Traffic volume fore cast plays an important role i n the highway constructi on study.In view of the status quo that one of t he direction of highway in our country tends to reform and widened, which has become one of the the main deve lopment trends in the future, in the article, ba sed on the existing facts characteristics of Hi ghwa y, highway traffic volume is predicted fr om “macroscopical” and “mi crocosmic ”, mainly in heavy traffic, and t he article is trying t o find new ide as for the t r affic volume prediction on widing the highway.In the begi nning,traffic volume i s pr edicted with t raditional four sta ges forecast method from “micro” view, according to the hi ghwa y traffic characteristics and t he influence of existing highways.It is difficult to get t he accurate result using t he traditional theoretical model, since the distribution of vehicle s has a large randomness.Littl e st udy has done on the prediction of the driving laws for divided lane vehic les.However, the influence of hea vy-duty vehicles on highway i s muc h la rger t han general passenger cars. The refore, this pape r predict the distribution of divide lanes for heavy-duty vehicles from the "micro" aspect s. To do this,first proposed the theory of grey system GM( 1,1) model to calculate the distri bution coefficients of accumulate d survey volume in two consecutive wee ks. Then take a re ference to BP neural network, using MATLAB pl atform, after several a mendments, to c orrect residuals of the gre y pre dicted r esults. Finally, combined the two methods to get the final result, namely the l ane di str ibuti on law of highway "six-lane".By t he site fiel d investigation for highway, the total traffic dis tribution was obtained and t hen the division lane law of heavy vehicles was predicte d. It can expand the traffic capacity, save economic cost and r educe engineering volume. It can he lp decisi on-makers to make ma cro-cont rol effectively, and provide a refere nce for people to t r avel. At the same time, it will pr ovide a t heoretical basis for highway expansion pr ojec t a nd new const ruction project i n the future. The improvement of s ub-lane approach can effectively ra ise t he prediction accuracy, and reduce the rela tive error. Pr ovide a r eference for the relevant research of traffic forecasts a s wel l as further study in the fut ure.