HMM-based highway traffic volume prediction method

Abstract

The invention discloses an HMM-based highway traffic volume prediction method. The method comprises the following steps: dividing a training set on the basis of collected historical traffic volume data, analyzing frequency domain characteristics by adopting FFT, and determining average turning time on the basis of attenuation characteristics of residual noise distribution; fitting the digital features of the historical traffic volume data training set by using a Markov model to obtain Markov model parameters; analyzing the peak characteristic of the traffic volume during holidays and festivals by using the updating process, and obtaining an updating process model and model parameters; and predicting future monthly traffic flow data by using the constructed hidden Markov model according to historical average peak traffic volume data. According to the invention, under the condition that the multi-dimensional traffic state information acquisition condition is limited, high-precision traffic volume prediction can be completed only depending on historical traffic volume data.

Description

technical field [0001] The invention belongs to traffic volume prediction technology, in particular to an HMM-based expressway traffic volume prediction method. [0002] technical background [0003] With the vigorous development of the transportation industry, the number of domestic automobiles is increasing day by day. According to the public data of the Chinese government website, the huge travel demand has led to a sharp increase in the traffic flow on the expressway. Because the growth rate of the traffic flow far exceeds the expected planning of the expressway, the vehicle congestion on the expressway has become more and more serious. According to the "2020 China Urban Traffic Report" released by Baidu Maps, among cities with 2 million to 3 million car populations, Guangzhou, Kunming, Nanjing, Jinan, Hangzhou, Changsha, Foshan, Hefei, Shenyang, Harbin, etc. Cities make it into the top ten. Based on the increasingly complex traffic congestion problem, scientific and ac...

AI Technical Summary

Problems solved by technology

[0004] The traditional traffic flow prediction adopts the four-stage method. Although the existing research has improved it to a certain extent, the accuracy is still low and it is no longer applicable.

Images