Map-Matching for Trajectory Prediction

Where are you going? Should you be going that way?

This article presents a method to predict vehicle trajectories on a digital road network using a database of past trips sampled from noisy GPS sensors. Besides predicting future directions, this method also assigns a probability to an arbitrary sequence of locations.

Central to this idea is using a digital map unto which we project all sampled locations by aggregating them into individual trajectories and …

data science editors pick geospatial gps