Database Reference
In-Depth Information
Besides the moving data types, Hermes contains a TB-tree implementation.
This structure supports the standard operations for this index (point query and
range query), but also k-NN and similarity queries.
Hermes' query language is SQL extended by spatio-temporal operations.
Although SQL is familiar to most database systems' users, formulating complex
temporal queries in SQL is a hard task and queries tend to degenerate to deeply
nested function calls.
3.7 Conclusions
In this chapter, we have motivated a high-level conceptual model of trajectories
as continuous functions, represented by abstract data types. These serve as a
foundation to extend the data model and query language of a DBMS to support
representation and querying of movement data. We have shown how queries can
be formulated in this framework. The implementation within a DBMS prototype
was sketched.
3.8 Bibliographic Notes
The field of moving objects databases is covered in depth in the textbook by
Guting and Schneider
(
2005
). The data model of Section
3.2
was developed
in a series of papers. In
Guting et al.
(
2000
), the type system and operations
are carefully designed. Further papers define the discrete model and develop
algorithms for the operations (see
Guting and Schneider
,
2005
, for references).
The model was extended to a network-based representation of moving objects
(or trajectories) in
Guting et al.
(
2006
). Recently, it was generalized to model
objects moving in different environments (for example, road networks, public
transport, indoor spaces) and according to different transportation modes (
Xu
and Guting
,
2013
).
The SECONDO system is freely available for download from its Web site,
7
where a lot of further documentation can be found.
Survey articles on spatio-temporal indexing are
Mokbel et al.
(
2003
)and
Nguyen-Dinh et al.
(
2010
). The TB-Tree is described in
Pfoser et al.
(
2000
), the
MON-tree in
Almeida and Guting
(
2005
). The Hermes system, which also par-
tially implements the model of Section
3.2
, is decribed in
Pelekis and Theodor-
idis
(
2005
)and
Pelekis et al.
(
2008a
).
SECONDO supports further query types such as continuous nearest neighbor
queries (
Guting et al.
,
2010
) and spatio-temporal pattern queries (
Sakr and
Guting
,
2011
). The latter are discussed in Chapter
12
.
7
http://dna.fernuni-hagen.de/Secondo.html/
