The Study Of Short-term Traffic Flow Prediction Based On Combined Model

With the development of the economy and the acceleration of urbanization, carownership increases year by year and the number of travelers is also rapid growth. This givesthe already very tense traffic tremendous pressure. Traffic problems have become thebottleneck restricting the development of the national’s economy. Among them, the greatestimpact in people’s lives is the traffic congestion problem. Intelligent Transportation System isan effective means to solve the problem of traffic congestion. Intelligent TransportationSystems (ITS) integrates application of information technology, communications transmissiontechnology, electronic control technology and computer processing technology and othertechnologies effectively to collect traffic information, process and handle. The purpose is torealize real-time, accurately informations to reflect the current and future traffic state, and toorganize and control the traffic scientific.The forecasting of short-term traffic flow is the core of ITS. The purpose of the article isto get real-time and accurate traffic prediction information. Real-time and accurate trafficprediction information can guide travelers to choose the right travel routes. It can save traveltime; and it also can ease road congestion, reduce environmental pollution, energyconservation and so on.Based on the basic theory of traffic flow, for a section of road, respectively, using themoving average model, GRNN neural network model and combined model establish trafficforecast model. The article uses several evaluation to compare and analysis the three models.It illustrates the feasibility of the linear model, and verified the advantages of the combinedmodel.The main work and conclusions of the article are as follows:(1)Introduces the background and significance, summarizes on the results of previousstudies, and analyzes the basic characteristics of the traffic flow. This illustrates the feasibilityof short-term traffic flow forecasting;(2)This article introduces some data acquisition and pre-processing methods and thebasic parameters,which can describe traffic flow characteristics,;(3)This article describes the knowledge of moving average model, neural network modeland combined model in detail.It establishes three short-term traffic flow forecastingmodels.They are the moving average mod, the GRNN neural network model, and thecombined model which is based on the moving average model and GRNN neural networkmodel.And the artile validates with experimental data and obtains satisfactory results.