Study On Link Speed Estimation In Urban Arteries Based On GPS Equipped Floating Vehicle

Traffic detection and traffic parameters estimation are essential to modern traffic management, controls, guidance and planning. However, Existing approaches to detect traffic condition provide traffic data only at particular points along the roads and can perform limited range. As a result, such approaches can not perform well in modern traffic control that being based on traffic condition detection of the whole road network. Moreover, the implementation and maintenance of such equipment can be costly. As the vehicles equipped with GPS receiver have increased in past few years, a new way to collect traffic data and estimate traffic parameters is the use of GPS equipped floating vehicle method. The research described in this paper is concerned with estimating link speed of urban arteries from traffic data that collected by existing relatively fewer and a variety of GPS equipped float vehicles in city according to existing bigger sampling cycle. Based on the character of traffic detection of GPS equipped floating vehicle, the purpose of the paper is to utilize limited data from GPS equipped floating vehicle to acquire reliable and accurate link speed estimation and provide traffic section with credible traffic parameter.By the analysis of existing models to estimate traffic parameters from GPS equipped floating vehicle, the paper proposes a solution to estimate link speed by integrating road traffic information and traffic detective data from GPS equipped floating vehicle based on spatio-temporal data mining. The solution consist of three phases: the feature extraction of traffic data from a floating vehicle passing a link, the mining of feature from a type of floating vehicle in a link, and the mining integrating feature from all type of floating vehicle in a link and link feature. According to the solution, an effective estimating model is proposed.According to the estimating model, firstly, traffic information study and data preprocessing are conducted to provide data mining with normative data. Secondly, based on spatial mining, first phase of data mining is conducted. The average speed of a floating vehicle passing a link is estimated by an algorithm, which integrates GPS positional data of the vehicle and the link node positional data in GIS map.Finally, based on temporal mining, the section speed of a type of floating vehicle passing a link in a period of estimating time is estimated by the weighed averaging, which combines the average speed estimation of every floating vehicle and its...