Introduction - Computational Movement Analysis - page 4

￿

Descriptive parameters are inherently scale-specific . Goodchild ( 2001 ,p.11)
makes a very illustrative case that in essence there is “no such thing as the slope
of a geographic surface, only a slope at a specific scale or grid spacing”. Simi-
larly, movement properties such as speed, acceleration or path sinuosity require
a sampling scale, and the choice of this scale is in most cases far from being
self-evident.

For all those reasons this topic argues for the need of new scientific fundamentals
for CMA. Following Antony Galton, this new theory shall enable GIScience to bridge
the “perceived gulf between, on the one hand, low-level observational [movement]
data that constitutes the “raw material” of our science, and on the other hand, the
high-level conceptual schemes through which we as humans interpret, understand,
and use that data” (Galton 2005 , p. 300).

1.2 Introducing Computational Movement Analysis

Computational Movement Analysis (CMA) draws concepts and methods from three
methodological research areas: (1) geographic information science or GIScience,
(2) computer science, and (3) statistics. Additional important contributions emerge
application fields studying movement, such as, for example, movement ecology,
surveillance and crowd management, as well as transportation research.
From geographic information science CMA inherits concepts for modeling space
and the movement within, as well as a suite of spatio-temporal operations inter-
relating space, time, and movement. From computer science CMA draws on the
database theory on how to store, index, and query inherently dynamic movement
data. Also from computer science CMA profits from developments around analyt-
ical tools such as data mining, knowledge discovery, and simulation for numerical
experiments. Finally, from statistics CMA inherits many techniques for descriptive
statistics, exploratory data analysis, and stochastic models for movement simulation
(for instance, random walk, states space models). Clearly, these fields overlap. For
instance, data mining and visualization reappear in visual analytics approaches for
movement data, and mapping spatio-temporal movement requires innovative visu-
alization approaches.
Although most applied research fields studying movement do not explicitly focus
on the development of computational movement analysis methods, they still signif-
icantly contribute to the respective theory. In movement ecology, for instance, there
is a very active community developing statistics based tools for movement analysis.
Similarly, many relevant developments for moving object databases emerge an active
community addressing fleet management problems.

Definition ( Computational Movement Analysis ) (CMA) is the interdisciplinary
research field studying the development and application of computational tech-
niques for capturing, processing, managing, structuring, and ultimately analyzing