We are gathering trajectories of GPS points of moving objects (cars, bikes, etc.) and would like to compute approximate curves (that may self-intersect) that would represent those clouds of points. What I am looking is a literature lists (preferably textbooks if not, papers) that I need to familiarize myself in order to solve this problem.

Any comments?

  • It will be helpful to distinguish versions of this problem. First, you have more than a point cloud: you also have an inherent temporal sequence to the points. That's crucial--among other things, self-intersection becomes a non-problem (because it occurs at different times). Second, in the applications you mention the points are expected to be on known features: roads and bike paths. This is such an advantage that you probably should ignore or carefully adapt solutions that use only the trajectory data (such as Kalman filters). – whuber Jul 5 '13 at 15:41
  • This is a good point. Yes, in fact we have a sequence of points that form paths. Drawn an curves they would intersect. Unfortunatelly there's no road-network available in the area, so no map-matching algorithms can be used. Still I would like to draw paths as curves (with some smoothing applied). Therefore I am looking for some literature that covers topics like curve fitting (with constant curvature) or some sort of smart smoothing (because the trajectory may intersect itself)... – arthur Jul 9 '13 at 11:10

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.