Study On Early Warning Method Of Road Traffic Congestion Based On The Detected Dynamic Traffic Information

With the continuous city development, urban traffic contradictionmainlyshowsthat urban traffic supply can not meet the increasing traffic demand, urban roadtraffic congestion problem is more and more serious, so that it has become a globalsocial problem that restricting the urban harmonious development, and alleviate urbanroad traffic congestion is more and more important. Urban trunk road is the urbantraffic arteries, timely, accurately predict and identify the urban road traffic congestion,targettlytake traffic control and traffic induced to the congestion point, can relieve roadtraffic congestion degree in the important traffic point and reduce negative effect oftraffic congestion. Therefore establish scientific and effective early warning method forurban road traffic congestion has important practical value.This paperanalyzed the changing feature and the space-time characteristics amongflow, speed and density in the process of formation, continuous, dissipate by usingtraffic flow model, identify the future traffic congestiondegree and adopt the slowblocking strategy by predicting the future traffic flow parameters of a certain roadsection. On the basis of the analysis of traffic flow parameters, compared several maintraffic data collecting technology and studied on the processing method of trafficinformation data.Urban trunk road traffic status is closely related to the adjacent section of trafficstatus, this paper presented anearly warning method of traffic congestion based on themulti-point status parameters, established a prediction model of critical section and therelated test section, forecastedfuture traffic flow state in the key section through thepresent and the historical traffic flow state parameter data series. Using ARIMA timeseries prediction model and improved BP neural network prediction method, a linearcombination of the traffic state parameter prediction model is established, based on theminimum error squares, equal rights and entropy value method to solve the weightvalue of thecombined model, and verifiedthe minimum error prediction method has theoptimal prediction effect throughsimulation example.Finally, comfirmedaverage speed, saturation, and the average delay as theevaluation indexes of traffic congestion in urban trunk road, dividedtraffic congestioninto unbloked, slight congestion, congestion and severe congestion, combined the prediction results of traffic state parameter, use the improved fuzzy comprehensiveevaluation method to identify the traffic congestion degree, and givethe early warninginformationof traffic congestion, then realizethe early warning of urban trunk roadtraffic jams.The traffic congestion early warning method proposed in this paper can forecasttraffic congestion timely and effective, and be used for early warning of traffic stateand traffic guidance in intelligent transportation system.