Resarch On Analysis And Application Of Monitoring Data In Intelligent Transportation System

The number of vehicles increasing rapidly makes traffic problem a worldwide problem. In order to solve traffic problem efficiently all governments are developing the intelligent transportation system. Intelligent transportation system is a system using modern technology for collecting and processing traffic information to supervise and monitor traffic condition. Intelligent transportation systems often contain a lot of traffic data from different sources with complex structures. The traffic data contains many inherent laws of the traffic system. The analysis of traffic data plays an important role in traffic monitoring and alleviating traffic pressure.This paper is based on the vehicle monitoring system of intelligent transportation system. Vehicle monitor system can collect vehicle information with a terminal using GPS and SIM card and store information in the database. There are three main objectives:1.Analysis the vehicle traffic data and predict average speed for specific position in the future.2.Add a speed prediction module based on vehicle monitoring system and storage speed prediction results. Driver and administer can see predicting result through particular query conditions.3.Make a standard to measure the driver’s driving behavior. This standard can be regarded as the risk coefficient of the driver. Administers can monitor it and decide whether to send alarm messages to the driver.Main work completed is as follows:(1) Transfer data from the original Oracle database to the local MySQL database to simplify the process of sampling.(2) Cluster traffic data and get the traffic time division result. In order to simplify the prediction data matrix, use each traffic period as a feature of the neural network. Get sample data of each feature to arrange the predicting matrix.(3) BP neural network is used to predict vehicle speed in the future with traffic information in each period as a category. BP neural network is trained after the principle of choosing parameters of hidden layer. Then a neural network with smaller error can be used in predicting vehicle speeds.(4) Add speed prediction module in the original system to display speed prediction results for administrator and driver. It uses spring framework, struts2framework and open source data persistence framework. Data transferring between platforms is based on web service. Front-end pages are completed with JavaScript, HTML and CSS. The whole module is developed with tomcat server and MySQL database. Driver and administer can see the predicting result by input query conditions.(5) The speed prediction module is based on the predicting results stored in the database. Users can input different conditions (time or position) to see the predicting results.(6) The definition of dangerous coefficient should be combined with related research of vehicle braking distance and predicting vehicle speed. Define risk coefficient and display it to administer, with which administer can decide whether to send warning messages to the driver.