Research On Link Travel Time Prediction Method Based On Data Collected By Floating Car

Link travel time prediction is an important part of the Intelligent Transportation System, forecasting results affect the function validity of ITS related subsystems( ATIS and ATMS etc) directly. Methods based on data gathered from loop and other fixed equipments are using one or more profiles' traffic conditions to predict the whole road's traffic condition, there are some errors. Floating cars distribute symmetrically on the road, and transfer Real-time information (such as time, speed, longitude, latitude and direction, etc) to the floating car information center. Therefore, the data gathered from floating car can efficiently make up the data gathered from fixed equipment.This paper reviews the research on floating car technology and link travel time prediction home and abroad, and researches the floating car date pretreatment methods and the methods to confirm link average travel time. On this basis, the paper researches the link travel time prediction modes and prediction methods based on floating car data character, and proposes short-term prediction model based on kalman filtering and long-term prediction model based on BP neural network., then researches on model input, parameter calibration method and parameter select range. Finally we give a demonstration of link travel time prediction to verify models' validity based on the floating car date gathered from Hangzhou city supported by The 10th Five Years Key Programs for Science and Technology Development of China: Urban road and traffic management based on floating car technology - demonstrations in Hangzhou city.