Study On Traffic State Identification Of Expressway Based On Statistical Properties

The expressway is not only affected by its own state,but also closely related to the running status of the roads and highways in the city.With the increase of traffic demand,the problem of congestion is becoming more and more serious.Alleviate the problem of quick road congestion,the need to express the status of the expressway to determine the premise of an accurate understanding of a comprehensive understanding of expressway traffic conditions,to provide effective support for the late management and control.This paper starts from the expressway traffic data level,and analyzes the statistical characteristics of urban expressway traffic data.With the help of the basic principle of statistical data analysis,the traffic volume Q,velocity V and time occupancy rate O are selected.Based on the change of these three parameters(35)Q?(35)V?(35)O,the distribution of the parameters is analyzed from the two aspects of parameter distribution and parameter distribution.Secondly,we choose the method of combining the gray correlation degree with the system cluster to estimate the fuzzy state of the traffic state parameters and determine the gray correlation matrix.At the same time,select the parameter selection statistics RSQ and semi-partial SPRSQ correlation statistics,and state the state parameters The best classification of the number of fuzzy estimates to determine the best traffic classification results;Next,the common two-phase,three-phase and other state classification based on the combination of the macro chapter obtained in the fuzzy classification of the best results,The traffic condition is further refined into four distinct characteristic stages: stable free flow,metastable convergence,metastable capture,stable obstruction,and combined with each change feature in each feature stage,select different parameters and different analytical methods to extract each state separately Type the characteristics and expand the analysis.After mastering the characteristics of various traffic conditions,it needs to be identified and applied to the theory and practice of traffic management control.Therefore,in the last chapter of this paper,based on the D-S evidence pattern recognition principle,based on the statistical characteristics of traffic parameters and the division of four types of traffic states,the traffic single parameter pattern recognition model is constructed to determine the basic probability assignment function,The range of Q?V?O change,the multi-state parameter data fusion,so as to complete the overall identification of traffic status,combined with examples,expressway traffic status characteristics of the stage of the law to identify the application.