CSE 594: Spatial Data Science &
Engineering
Overview of Moving Object Data
Trajectory of
Moving Objects
A type of spatiotemporal data generated by moving objects
A trajectory is a polyline in three-dimensional space
Two dimensions refer to the space and the third dimension
refers to the time
Represented as a sequence of position points Tr(P1, P2, …, Pn)
Each position point Pi is represented as (lat, lon, t), where
lat and lon are location coordinates and t is the timestamp
Queries on Trajectory Data
Given a dataset, a moving object id, a time range [ts, te], the query returns all the trajectory
line segments of the corresponding object such that all position points Pi of the line segments
have the timestamp value within the given time range
If the object id is not given, return all trajectory line segments satisfying the above constraints
Temporal Query
Given a dataset, a spatial range {latmin, lonmin, latmax, lonmax}, the query returns all trajectory
line segments such that all position points Pi of the line segments are located in the spatial
range
Spatial Range Query
Queries on Trajectory Data
Given a dataset, a spatial range {latmin, lonmin, latmax, lonmax}, and a time range [ts, te], the
query returns all trajectory line segments such that all position points Pi of the line segments
are within the spatial and temporal range
Spatiotemporal Range Query
Given a dataset, a query trajectory q, a distance function, and a distance threshold, the
similarity query returns all trajectories Tri where the distance between q and Tri is not greater
than the given distance threshold.
Similarity Query
Queries on Trajectory Data
Given a dataset, a query trajectory q, a positive integer k, and a distance function, the K-NN
query returns k nearest trajectories of q, where the distance from q to other trajectories is
measured based on the given distance function
K-NN Query
Popular Moving Object Databases
MobilityDB
Secondo
TrajMesa
TrajStore
MovingPandas
Sci-kit mobility
Libraries for Analyzing Moving Objects