=Paper=
{{Paper
|id=Vol-1660/womocoe-paper3
|storemode=property
|title=Ontology Architectures for the Orbital Space Environment and Space Situational Awareness Domain
|pdfUrl=https://ceur-ws.org/Vol-1660/womocoe-paper3.pdf
|volume=Vol-1660
|authors=Robert John Rovetto
|dblpUrl=https://dblp.org/rec/conf/fois/Rovetto16
}}
==Ontology Architectures for the Orbital Space Environment and Space Situational Awareness Domain==
https://ceur-ws.org/Vol-1660/womocoe-paper3.pdf
Ontology Architectures for the Orbital
Space Environment and Space Situational
Awareness Domain
Robert John ROVETTOa,1
a
University of Maryland, College Park Alumnus (2007),
The State University of New York at Buffalo Alumnus (2011)
American Public University System/AMU, space studies
Abstract. This paper applies some ontology architectures to the space domain,
specifically the orbital and near-earth space environment and the space situational
awareness domain. I briefly summarize local, single and hybrid ontology
architectures, and offer potential space ontology architectures for each by showing
how actual space data sources and space organizations would be involved.
Keywords. Ontology, formal ontology, ontology engineering, ontological
architecture, space ontology, space environment ontology, space object ontology,
space domain ontology, space situational awareness ontology, data-sharing,
astroinformatics.
1. Introduction
This paper applies some general ontology architectures to the space domain,
specifically the orbital space environment and the space situational awareness (SSA)
domain. As the number of artificial satellites in orbit increases, the potential for orbital
debris and orbital collisions increases. This spotlights the need for more accurate and
complete situational awareness of the space environment. This, in turn, requires
gathering more data, sharing it, and analyzing it to generate knowledge that should be
actionable in order to safeguard lives and infrastructure in space and on the surface of
Earth. Toward this, ontologies may provide one avenue.
This paper serves to introduce the space community to some ontology architecture
options when considering computational ontology for their space data and space
domain-modeling needs. I briefly summarize local, single and hybrid ontology
architectures [1][2], and offer potential space domain architectures for each by showing
how actual space data sources and space organizations would be involved. For all
figures, each node represents a distinct ontology or ontology suite (ovals), vocabulary,
or data-source (rectangles). As the paper provides a cursory discussion, details, both
technical and methodological, are left to existing publications and future work.
1
Corresponding author. Email: rrovetto@terpalum.umd.edu ; ontologos@yahoo.com
2. The Space Domain
The domain of interest encompasses the phenomena in orbit, in near-Earth and deep-
space environments relative to Earth. These entities are of interest because we should
(i) protect persons and property in orbit and on the planetary surface, (ii) increase our
scientific knowledge of the space environment, and (ii) ensure the future of safe and
peaceful space flight. The domain includes observation, detection, identification,
tracking, and propagation (prediction of future motion and behavior) of orbital objects
[4]. As such, the sensors, sensor networks, accumulated data and the processes by
which we attain knowledge of the entities under study are part of the domain of interest.
Collectively, this domain has been described as the space situational awareness (SSA)
domain. Given a focus on the regions of space where objects orbit, the domain can also
be called the orbital space environment domain, which could arguably have a wider
scope and have SSA as a part. If spatial zones from Earth are delimited, this may be
distinct from other space environment ontologies. Various permutations are possible,
depending on scope or domain demarcation.
Whereas astronomy studies all astronomical phenomena, the SSA domain or space
domain awareness is concerned with those objects in closer proximity to Earth and the
processes by which we achieve awareness of them and their environment. SSA is
essentially about the space phenomena in relation to Earth, i.e., their potential effect on
Earth and our space-related assets.
Both SSA and astronomy are data-intensive disciplines. Ground- and space-borne
sensors accumulate data on natural and artificial objects in orbit, and in the further
reaches of our solar system. Optical, radar, and infrared sensors individually provide
one (or more) aspect(s) of the observed orbital or near-Earth object (NEO).
Collectively, they provide a broader picture of these space objects. Catalogues or
databases of satellites, orbital debris, near-Earth objects and space weather phenomena
are maintained from this data. Example sorts of data include: the orbital parameters
used to describe an orbit, positional and motion data (as in the Two-line Element Sets),
and physical property data (shape properties, reflectance, mass, etc.). As the volume of
datasets grows, big data and ontology engineering research and applications may serve
to achieve the goals of space data-exchange for improved SSA.
3. Ontology and Ontology Architectures
Ontology is the general study of a given subject matter, universe of discourse or
domain. A computational or applied ontology is a computable terminology with a
formal semantics, the totality of which expresses a theory or understanding of the given
domain. Ontology terms annotate data from space data sources toward fostering data-
exchange and interoperability. A variety of ontology development and engineering
architectures [1][2], and methodologies [6][7] exist for the space community to
consider. I summarize three architectures.
A local (or multiple) ontology architecture is one in which “each information
source is described by its own ontology” [1]. Each ontology can then be interconnected,
creating a link between distinct databases, thereby facilitating data-exchange. A method
to interconnect local ontologies is by mapping ontology terms to one another. A single
ontology architecture has one ontology providing a shared terminology to annotate data
from multiple databases, has been called a „global ontology approach‟ [3], and can “be
a combination of several specialized ontologies” [1]. A hybrid architecture is one that
incorporates design features from each. “Similar to multiple ontology approaches the
semantics of each source is described by its own ontology.”[1]. Local ontologies
directly annotate data, while also having a shared vocabulary. Each of the three
architectures can use parts (selected terms) of, or the entirety of, other ontologies. In
the next section I offer a potential space ontology architecture for each of the above
three general ontology development architectures.
4. Space Ontology Architectures
I now apply each ontology architecture to the orbital space and SSA domain, offering
potential scenarios for space community data-sharing and inter-organizational
cooperation. Space agencies and SSA databases are included as actors and data sources.
These architectures are subject to revision, but draw upon [5] and [4]. The potential or
actual ontologies discussed may be distinct ontologies in their own right, or may
forming modules of a larger ontology.
4.1 Local Orbital Space Domain Architecture
Figure 1 portrays a cooperative scenario in which four space actors use an
interconnected system of locally developed ontologies to share data. They are: the
European Space Agency with its SSA and near-Earth object data; the National
Aeronautics and Space Administration with its orbital debris data, one or more
universities with, say, asteroid, comet and other NEO data; and satellite operators with
their own satellite and observational data. An ESA NEO ontology may be part of a
broader ESA SSA ontology, and likewise for the local ontologies of other space actor
partners.
Figure 1. Local space environment ontology architecture (v1). Scenario: European Space Agency, NASA,
universities & satellite operators exchanging data via interconnected local ontologies that annotate local
databases.
This is an architecture in which each space actor has sovereignty to design, develop,
and test their own ontology (or ontology library) to represent and annotate their data.
As such, a local SSA ontology architecture is helpful where space actors seek to
exchange information without a mediating or bridging resource. By developing
together, they can make the interconnection (e.g. via mappings) and interoperability of
their ontologies and systems smoother. Given the space community‟s shared scientific
knowledge (astrophysics, astrodynamics, etc.), and given that some space actors will
observe and track the same (numerically identical) orbital object (e.g. a GPS satellite),
concepts, terminology and semantics will overlap.
Figure 2 adds ontologies with broader domain concepts, as well as sub-domain and
other related content. Additions include: an astrodynamics standards ontology, which
may consist of data formats, and computational models; science reference ontologies,
the space situational awareness domain ontology (SSAO) [4]2, and event and object
ontologies that may be modular parts of the latter. They provide some of the common
knowledgebase for local ontologies. Other demarcations of the overall domain (and
individual ontologies) include the SSAO being equivalent to or, alternatively, part of
what I call the Orbital Space Environment Ontology (OSEO) (or some variation
thereof). As a part, the SSAO would primarily represent the activities and object
involved in achieving and maintaining awareness of the space environment, e.g.
observational, tracking, and computational processes.
Figure 2. Local space ontology architecture (v2): local ontologies using generic ontologies.
The SSAO or OSEO includes general terms that can be applied to any of the local
ontologies. These terms may subsume local terms, be asserted as equivalent (depending
on the intended meaning of corresponding terms), or may be imported into the
ontology. They help annotate data about individual orbital parameters, Two-line
Element Sets, observations, various objects in orbit, orbital events (e.g. collisions), and
so on. Examples of common terms include: orbital debris, launch vehicle, orbit,
circular orbit, inclination, asteroid, optical sensor, orbital conjunction, Hohmann
Transfer Maneuver, etc. More general domain-specific terms include: orbital
occurrence, orbital object, space object, space artifact, and orbital property. Scientific
discipline ontologies, e.g., physics, orbital dynamics, are developed to provide the
formal representations of the relevant scientific knowledge and principles.
4.2 Single Orbital Space Domain Architecture
2
Under development. URL=https://github.com/rrovetto/space-situational-awareness-ontology/
A single orbital space ontology architecture (Fig.3) can take the form of the SSAO or
OSEO directly annotating space data from distinct data systems from similar space
actors. Given that the data is about objects in space, a variety of their properties and
observations thereof (among other things), the ontology should have domain terms for:
orbits, orbital parameters, orbital objects such as debris and satellites, physical features,
satellite operations activities (e.g. launches, navigation, maneuvering); tracking,
propagation, and so on.
Figure 3. Single space ontology architecture (version 1): National, academia and company space actors
utilize a more general domain ontology such as the SSAO from [4] or some variation.
The single ontology used for the domain would itself be domain-specific. Without
other ontologies, this architecture avoids mappings, but faces the challenge of
agreement on the structure and content of the ontology. Figure 4 depicts a variation of
this architecture, adding other potential ontologies, or alternatively, decomposing the
single ontology into sub-domain ontology parts (similar to Fig.2) to form a more
complete picture of the domain. Some include the Orbital Debris Ontology [5],
Astronomical Object Ontology (e.g. similar to [8]), and physics ontologies. Terms from
each could be imported into the single ontology. Local data elements are described
with the semantics and formalisms of these ontologies.
Figure 4. Single space domain ontology architecture (version 2): added other ontologies that may be
modularized parts of the single ontology or distinct.
4.1. Hybrid SDO Architecture
Finally figures 5 and 6 present the hybrid architecture, the latter adding additional
ontologies. We see a higher-level shared space vocabulary, or alternative domain
ontologies, e.g., the SSA domain or orbital space ontology. Space actors would use this
common resource to provide a backbone terminology to relate to their local ontologies.
As a shared ontology, it may subsume the local ontologies. The shared resource may
also be compositional, consisting of distinct sub-domain ontologies. This architecture
has the benefit that local domain professionals can help ensure veridical formal
descriptions. It has the challenge of agreement on the terminology, definitions and
formalization of the shared higher-level resource.
Figure 5. Hybrid space ontology architecture.
Figure 6. Hybrid Space Ontology Architecture (version 2): addition of other modular ontologies.
5. Other Considerations and Discussion
Other considerations in the development of an ontological framework are the
ontology languages, i.e., the computable formalisms used to represent the domain.
Each language has its own limits on expressivity. According to [1] “The role and the
architecture of the ontologies influence heavily the representation formalism of an
ontology.” We also read “Depending on the use of the ontology, the representation
capabilities differ from approach to approach.” Whether to use other ontologies at
different levels of abstraction is also a consideration.
As always, another option is to develop a novel architecture or approach
(ontological or otherwise). Moreover, depending on the space community‟s needs,
feasibility studies, and the ability of ontology engineering to address those needs,
ontology may or may not be the best research direction. Each of the architectures
would be an interesting pursuit for this domain, as would the development of a novel
approach. In any case, the goals remain: to solve space data problems, improve SSA for
space (and terrestrial) safety, and expand our knowledge of the space environment.
6. Conclusion
When researching ontology for space data needs, there are various possible
architectures and methodologies the space community may consider. This paper
applied the local/multiple, single and hybrid ontology architectures to space situational
awareness and orbital space environment domain. I offered ideas for space domain
ontology architectures toward stimulating both data-sharing and international and inter-
institutional cooperation.
References
[1] H. Wache, T. V¨ogele, U. Visser, H. Stuckenschmidt, G. Schuster, H. Neumann, and S. H¨ubner.
Ontology-Based Integration of Information - A Survey of Existing Approaches. In Proceedings of the
IJCAI-01 Workshop on Ontologies and Information Sharing, 2001. URL= http://ceur-ws.org/Vol-
47/wache.pdf
[2] Isabel F. Cruz; Huiyong Xiao, The Role of Ontologies in Data Integration, Journal of Engineering
Intelligent Systems, 13 (2005) 245-252.
URL= https://www.cs.uic.edu/~advis/publications/dataint/eis05j.pdf
[3] M.Gagnon. Ontology-Based Integration of Data Sources. In Information Fusion, 2007 10th International
Conference on Information Fusion, 9-12 July 2007, p.1-8. DOI: 10.1109/ICIF.2007.4408086, IEEE.
[4] R.Rovetto & T.S. Kelso. Preliminaries of a Space Situational Awareness Ontology. 26th AIAA/AAS
Space Flight Mechanics Meeting, Napa, CA, USA. Feb 14-18th, 2016. Forthcoming in Advances in the
Astronautical Sciences, Univelt.
[5] R.Rovetto. An Ontological Architecture for Orbital Debris Data. Earth Science Informatics, 9(1), 67-82
(Aug 6 2015). Springer Berlin Heidelberg. URL= http://link.springer.com/article/10.1007/s12145-015-
0233-3 , DOI: 10.1007/s12145-015-0233-3
[6] D.Jones, T.Bench-Capon, P.Visser, Methodologies For Ontology Development.
[7] F.M.López, Overview Of Methodologies For Building Ontologies, Proceedings of the IJCAI-99
workshop on Ontologies and Problem-Solving Methods (KRR5) Stockholm, Sweden, August 2, 1999
URL= http://sunsite.informatik.rwth-aachen.de/Publications/CEUR-WS/Vol-18/4-fernandez.pdf
[8] International Virtual Observatory Alliance (IVOA). Ontology of Astronomical Object Types. URL=
http://www.ivoa.net/documents/Notes/AstrObjectOntology/