Navigation Road Data Update Based On Taxi GPS Track

A routable road map with the road location information, road lane direction, and road traffic rules is the essential to intelligent transportation systems(ITS). The traditional mapping with a long term and complex process among multi-agencies, makes map updating slowly, and are unable to deliver the changes of traffic rules on existing roads and to reflect the latest information of newly constructed roads. In recent years, The method of the road map updating through an unique geospatial process of mining GPS traces collected from public vehicles is becoming more and more attention at home and abroad that A vehicle trace is a time series of positioning points, in which a point represents road location while the order among points indicates traffic rules. The dataset of GPS traces are wide sources and large amount of data although it has a low accuracy of positioning. Therefore, it could be obtained extensively from an inexpensive GPS receiver installed on public vehicles such as taxicab, bus, patrol car and so on.The methods of extracting and updating road by using GPS trajectory have been researched in this paper by taking kunming taxi GPS trajectory data for experimental data. And the regression algorithms, the algorithm of least squares support vector regression machine have been attemped for the extraction of road network in this paper. In addition, based on buffer and skeleton line, a new method of extracting the road network structure from from the GPS trajectory has been presented And at last the contrast experiment for the three methods have been done.On the basis of the extraction of road network, the method of judging the road updates by using the theory of change detection is simple and quick but can’t tell the specific way of updating.Therefore, on the basis of this, the spatial analysis and updating method is proposed based on incremental judgment and the experiments are done. The experimental result shows that the method is simple and efficient, and has strong practicability.