Robust Median Trajectories


Update content

Title Robust Median Trajectories
Period 12 / 2011 - 12 / 2015
Status Current
Research number OND1343406
Data Supplier NWO


Trajectories represent the movement of entities like animals, people, vehicles, hurricanes, etc. They have become a type of data that is more and more collected and analyzed. This is due to technological advances like GPS, RFID tags, and mobile phones, and an increase in demand for analysis possibilities. New analysis methods for trajectory data have been developed in the last years, but a number of basic concepts are still lacking a satisfactory study. Two of these concepts are the median trajectory and the medoid trajectory, for a collection of trajectories. Intuitively, a median trajectory for a set of trajectories is some (new) trajectory that is in the middle of the set at all times. A medoid trajectory is a type of median that is also in the middle but always coincides with some trajectory of the set. This project addresses the question of the definition and computation of the median and medoid trajectories for a collection. In particular, we are interested in robust definitions that do not deteriorate if some parts of some trajectories in the collection deviate significantly. A good definition of median and medoid will aid clustering of trajectories, detection of outliers, and visualization of collections of trajectories. Algorithms to compute such median or medoid trajectories efficiently are essential, and will be developed. We will implement these algorithms, and analyze the suitability of different median and medoid trajectories for the tasks clustering, outlier detection, and visualization. To this end, we use trajectory data from birds, pedestrians, and vessels. null

Related organisations

Related people

Project leader Prof.dr. M.J. van Kreveld

Go to page top
Go back to contents
Go back to site navigation