Moving objects are objects like cars, persons or animals equipped with a GPS devise that have a
geometry changing over the time: they produce trajectories, that is to say descriptions of the
movements of those objects. We can define a trajectory as a time-space function that records
the changing of the position of an object moving in space during a given time interval.
I follow the model presented in the article "A conceptual view on trajectories"
[S. Spaccapietra et al.], where a trajectory is a sequence of stops and moves.
A stop is a part of a trajectory in which the moving object does not move, while a move is a
part
of a trajectory in which the moving object changes its position.
Starting from this model, I analyze all kinds of uncertainty involved, trying so give some
solution
and representation and developing an algorithm for trajectories classification, taking care
about
uncertainty.
An algorithm for Trajectories Classification: We are interested in trajectories of people driving cars, that stop near some points of interest and reach them walking. We describe a conceptual model for those trajectories. We describe a probabilistic algorithm to classify a trajectory on the basis of the points of interest visited by a person during her trajectory.
Uncertainty in Trajectories Classification: Definition of uncertainty and its representation in our model.
Trajectories Classification (slides): Uncertainty in trajectory classification. Point of Interest. conceptual Model. Fuzzy Geometries. Algorithm for trajectory classification. Annotations and semantics.