Citation Context

...n. To name but a few methods: Meratnia et al. propose Top-Down Time Ratio (TD-TR) and OPen Window Time Ratio (OPW-TR) for the compression of spatiotemporal trajectories [23]. In addition, the work of =-=[26]-=- provides two sampling based compression methods: threshold-guided sampling and STTrace to deal with limited memory capacity. Recently, semantic-based trajectory model construction has emerged as a ho...

Citation Context

...Offline algorithms are presented in works [5] and [7]. In work [7], Meratnia and de By also suggest an algorithm for online data. Two algorithms for online data compression are also suggested in work =-=[9]-=-. For the map-matching problem, Brakatsoulas et al. [1] suggest an algorithm that works exclusively for offline data. Although there are suggestions for the compression and the map matching problem, t...

Citation Context

...ation. It is widely used in spatial databases, but it requires multiple passes over the data, which yields it inapplicable to on-line streaming applications. Sampling techniques have been proposed in =-=[23]-=- for compressing isolated trajectories. Furthermore, segmentation of multiple trajectories by fitting into axis-parallel rectangles is considered in [1, 11], where dynamic programming, greedy and heur...

Citation Context

...ce 1068-0314/12 $26.00 © 2012 IEEE DOI 10.1109/DCC.2012.14 62 Opening Window [1] algorithm was also proposed to solve the line simplificationsproblem sequentially in a greedy manner. Another solution =-=[3]-=- was the threshold-guidedsalgorithm via estimating the safe area of the next point using the position, speed andsorientation information. STTrace sampling algorithm was also implemented using asbottom...

Citation Context

...nd it occurs that |θ− θ′| > dθ. Such lightweight calculations can be performed by every object with negligible overhead, although a more sophisticated dead-reckoning [12] or threshold-guided strategy =-=[7]-=- could also be applied. The server registers a set F of focal points and at each time instant t inspects movements in their scope according to streaming object statuses. Status information is retained...

Citation Context

...d B-trees to host temporal intervals. This scheme maintains all successive stages of aggregation without any notion of time-decay; instead, our 3-tier approach is geared towards a streaming model. In =-=[10]-=- we proposed two single-pass sampling techniques to achieve effective trajectory compression, exploiting spatial locality and temporal timeliness inherent in trajectory streams. The underlying princip...

Citation Context

...nts is negligible. Future plans include development of additional features concerning construction and maintenance of synopses (e.g., by ignoring non-critical locations from trajectory representation =-=[7]-=-), as well as incorporation of complex trajectory joins for identifying kinetic operations, such as similarity and periodicity in movement [6]. Figure 1: Graphical interface of map application. 4. SYS...

Citation Context

...ctice is to oversample initially, and then to simplify by eliminating observations that add little information. In view of this, several algorithms have been developed for simplifying trajectory data =-=[17, 6, 21, 18, 11]-=-. All these algorithms make the natural assumption that the goal should be to simplify trajectories such that the position information captured in the simplified trajectories is “similar” to the posit...

Citation Context

...an error threshold. The DouglasPeucker algorithm has the limitation of ignoring temporalsdata. Top-Down Time Ratio (TD-TR) [5] overcomes thisslimitation by using Synchronized Euclidean Distances(SED) =-=[6]-=- instead of spatial error. The characteristics ofsthe vehicle trajectory are described in the proposedsalgorithm [7], in which the advantages and disadvantagessof judgment method with point by point a...