Compressing Spatio-temporal Trajectories

Abstract

A trajectory is a sequence of locations, each associated with a timestamp, describing the movement of a point. Trajectory data is becoming increasingly available and the size of recorded trajectories is getting larger. In this paper we study the problem of compressing planar trajectories such that the most common spatio-temporal queries can still be answered approximately after the compression has taken place. In the process, we develop an implementation of the Douglas-Peucker path-simplification algorithm which works efficiently even in the case where the polygonal path given as input is allowed to self-intersect. For a polygonal path of size n, the processing time is O(n log k n) for k = 2 or k = 3 depending on the type of simplification.

Direction-preserving trajectory simplification

“…We adopt the Douglas-Peucker algorithm [6] for PPTS which is the most popular existing algorithm for PPTS [17,3,8], and we use our SP algorithm for DPTS. We vary ǫt for DPTS.…”

Visualization of Vessel Traffic

“…The basis for our trajectory compression is a Piecewise Linear Segmentation (PLS) method, such as studied by Gudmundsson et al [64], and Cao et al [31]. Also in Chapter 3 we proposed a trajectory compression algorithm, which is similar, but reconstructs the position using the velocity in the second step of the algorithm.…”

“…From earlier work on trajectory compression algorithms we selected the two error measures that retained the most stops. We used E µ from Gudmundsson et al [64], for which we selected the best value for µ (in terms of stop retention) that we could find, and E t from Cao et al [31].…”

Reporting Leaders and Followers among Trajectories of Moving Point Objects

“…Simplified trajectories are likely to be non-synchronous, yet they can approximate the original trajectory within a fixed specified error bound (see e.g. [6,19]). …”