Although much progress in high performance computing has been made in recent years, there still lacks a mechanism to enable global-scale integration and sharing of large quantities of da
Trang 1Sharing of Distributed Geospatial Data through Grid Technology
national climate data center (Ramapriyan et al.,
2006) The data volume will increase significantly
if similar models of finer spatial resolutions, such
as 1 km, are used The models are being changed
and refined from time to time and new geospatial
data, the NASA EOS data and NOAA climate
data, are being collected by satellites
continu-ously A fixed computing environment that
con-tains only static data sources will not fulfill such
kind of geospatial applications Consequently, a
capability of seamless and dynamic accessing to
large quantities of distributed geospatial data is
the key to the success of today’s and tomorrow’s
geospatial applications
Although much progress in high performance
computing has been made in recent years, there
still lacks a mechanism to enable global-scale
integration and sharing of large quantities of
data, such as geospatial data, from large-scale,
heterogeneous, and distributed storage systems
Fortunately, the emerging Grid technology might
be able to solve this problem Grid technology is
a form of distributed computational technology
that involves the coordination and sharing of
computing, application, data, storage, and network
resources across dynamic and geographically
dis-persed organizations (Foster et al., 2001) Resource
sharing in a Grid is highly controlled Resource
providers and consumers define clearly and
care-fully what is shared, who is allowed to share, and
the conditions under which the sharing occurs
Individuals and/or institutions agreeing to follow
such sharing rules form a virtual organization
(VO) The resource sharing across multiple VOs
is enabled by the Grid technology The intrinsic
advantages of the Grid technology fit the problems
of the sharing of distributed geospatial data very
well (Di, 2005) The Globus Toolkit, currently at
version 4, is an open source toolkit for building
Grids provided by the Globus Alliance It provides
many useful components and services that make
the use of Grid technology easier
sh Ar Ing of geosp At IAL dAt A through gr Id techno Logy
To enable the sharing of distributed geospatial data, a large-scale infrastructure that can integrate the currently dispersed data together and enable the efficient sharing of those huge amounts of geospatial data in a secure and controllable man-ner is crucial But because geospatial data are huge in quantity and geographically distributed across heterogeneous environments, there are still a lot of problems need to be faced with and solved in order to create such an infrastructure Those major problems and how they can be ad-dressed by Grid technology are discussed in the following section
System heterogeneity There are hundreds of
large geospatial data centers and countless small
or personal data centers around the world forms and systems used to store and manage the geospatial data in each center may vary greatly There are many types of high performance stor-age systems used, such as the Distributed Parallel Storage System (DPSS), the High Performance Storage System (HPSS), and the Storage Resource Broker (SRB) Unfortunately, these storage sys-tems typically use incompatible protocols for data access (Allcock et al., 2002) Also, the diversity
Plat-of platforms and systems on which geospatial applications are running greatly increase the data sharing difficulty Thus, geospatial applications should be presented with a uniform view of data and uniform mechanisms for accessing the data independent from the platforms and systems used Grid technology addresses this problem by providing storage system abstraction and uniform API for data accessing Several components and tools have been provided in the Globus Toolkit, including GridFTP and OGSA-DAI, to integrate heterogeneous systems and make the geospatial data accessible throughout the Internet
Uniform mechanism to publish and discover
geospatial data Usually geospatial data are
Trang 2
Sharing of Distributed Geospatial Data through Grid Technology
lished by extracting their attributes – geospatial
metadata, storing and managing them within
catalogues, and making the metadata queryable
Heterogeneity exists in this process because,
currently, different models are used to describe
geospatial metadata and different methods are
used to query geospatial metadata For example,
Earth Observation System (EOS) ClearingHOuse
(ECHO) and EOS Data Gateway (EDG) both
provide the capabilities to publish and discover
NASA EOS data, each with a different model to
describe NASA EOS metadata and a different
approach for users to search NASA EOS data
To solve this problem, two issues need to be
ad-dressed One issue is the need for a widely accepted
domain metadata schema to eliminate semantic
heterogeneity of different metadata models There
are domain standards for geospatial metadata
schemas available to address this issue, such as
ISO 19115 – Geographic Information Metadata
(ISO, 2003a) and ISO 19115 part 2 – extensions
for imagery and gridded data (ISO, 2003b) The
other issue is the need for uniform interfaces for
publishing and discovering geospatial data from
different metadata catalogues An example of such
uniform interfaces is the Catalogue Service – Web
Profile (CSW) developed by the Open Geospatial
Consortium (OGC) (Nebert and Whiteside, 2005;
Wei et al., 2005) The intrinsic Service Oriented
Architecture (SOA) characteristic of the Grid
technology enables the cooperation of
differ-ent catalogues With Grid technology, legacy
catalogues can be wrapped and exposed as Web
services which provide uniform publishing and
discovering interfaces, while leaving the internal
mechanisms of the catalogues untouched Grid
technology also provides a mechanism for
creat-ing federations of distributed catalogue services
Queries to any single accessing point of such a
federation can be delivered to all the catalogue
services throughout the federation Thus the
discovery of geospatial data can be much more
efficient
Performance Geospatial data are not only
large in quantity but also huge in size Although the computing capability and network bandwidth are increasing rapidly, accessing and transfer-ring large amounts of geospatial data are still huge burdens Grid technology provides several mechanisms that can improve availability and accessing performance of geospatial data, one of which is an important component within a data-intensive Grid environment – Data Replication System (DRS) provided by Globus Toolkit A data replica is a full or partial copy of the original data (Chervenak et al., 2001) With the help of DRS, multiple replicas of the original geospatial data can be created, distributed, and managed across different storage systems and data centers DRS monitors the storage systems, computing plat-forms, and networks within a Grid environment
in real time If a user wants to access a specific geospatial data, DRS will choose one replica which provides the best accessing performance for the user DRS can even choose more than one replica for the user and provide the user with a stripped-style data accessing mechanism which enables the user to retrieve different parts of the original geospatial data from different replicas simultaneously and combine those different parts into a complete data after retrieving Multiple replicas are created to increase the availability
of geospatial data; otherwise a single failure will make those geospatial data unavailable The accessing performance is also improved by choosing optimized replicas Other mechanisms are also provided by Grid technology to improve the accessing performance and reliability for geospatial data, like GridFTP, which provides much more improved data transfer performance than the traditional FTP protocol
Security Security is a critical issue
associ-ated with the sharing of geospatial data Many
of the geospatial data are sensitive and restricted
to be accessed by only some special persons or organizations Some of the geospatial data are
to be shared for commercial purposes and are
Trang 3Sharing of Distributed Geospatial Data through Grid Technology
associated with an accessing fee Currently
dif-ferent organizations and communities are using
diverse mechanisms to handle security related
issues, such as authentication, authorization, and
access control Consequently, there is a need for
a uniform security mechanism to coordinate the
sharing of geospatial data across those naturally
untrusted organizations and user communities
while keeping the diverse local security
mecha-nism intact The Grid Security Infrastructure
(GSI) provided by the Grid technology can be
used to address this problem Based on GSI, each
geospatial organization or user community can
form a VO Each individual user, machine,
stor-age system, application, or a VO will have one
or more certificates as its identity Certain trust
relationships can be set up among different VOs
(Welch et al 2003) As a consequence, a larger
VO is formed Thus, fine-grain access control
policies on geospatial data can be issued to any
individual user, application, or VO that has one
or more certificates through Community
Autho-rization Service (CAS) provided by the Globus
Toolkit Currently, the X.509 certificates based
on Public Key Infrastructure (PKI) are used by
Grid technology and to provide high-level
au-thentication, authorization, and single sign-on
functionality (Welch 2005)
Today, efforts have been taken by some
geosci-ence communities to leverage Grid technology for
the sharing of geospatial data For example, Earth
System Grid (ESG) is a research project sponsored
by the U.S Department of Energy (DOE) Office of
Science to address the formidable challenges
as-sociated with enabling analysis of and knowledge
development from global earth system models
The goal of ESG is to provide a seamless and
powerful environment that enables next
genera-tion climate research by integrating distributed
federations of supercomputers and large-scale
data & analysis servers through a combination
of Grid technology and emerging community
technologies The Center for Spatial Information
University also developed a prototype system for efficient sharing, customization, and acquisition
of distributed NASA EOS data by integrating the Grid technology and Open Geospatial Consortium (OGC) Web Services technologies This prototype system involves three partners distributed across the United States: George Mason University, NASA Ames Research Center, and Lawrence Livermore National Lab Each partner forms
a VO and trust relationships are set up among those three VOs to create an integrated Grid environment About 20TB of remote sensing and climate simulation data are shared among this prototype Grid-enabled Catalogue Service for Web (CSW) was implemented to provide uniform mechanism for data publication and discovery Data Replication System and Resource Selection components were also implemented to improve the performance of data sharing The customization
of data was achieved by leveraging OGC Web Services, such as Web Coverage Service (WCS) and Web Map Service (WMS), to provide more options for geospatial data accessing
future trends
The goal of the Grid technology is to create a computing and data management infrastructure that will provide the electronic underpinning for a global society in business, government, research, science, and entertainment (Berman et al., 2003)
As an essential information source for scientific research and even people’s everyday life, distrib-uted geospatial data all over the world are also doomed to be integrated to form a global-scale warehouse to promote the sharing of geospatial data Grid technology is still young and there are many open issues to be addressed and missing functionalities to be developed New computing and network technologies are also emerging and advancing, such as the wireless and mobile computing technologies, which greatly extend the
Trang 4
Sharing of Distributed Geospatial Data through Grid Technology
tion With the maturation of Grid technology and
the advancement of computing and network
tech-nologies, this will not only be a dream: wherever
the geospatial data are, they can be shared and
accessed from almost anywhere at anytime
conc Lus Ion
With the rapid accumulation of geospatial data
and the advancement of geoscience, there is a
critical requirement for an infrastructure that
can integrate large-scale, heterogeneous, and
distributed storage systems for the sharing of
geospatial data within multiple user communities
The emerging Grid technology can address the
problems associated with the sharing of
distrib-uted geospatial data, including the heterogeneity
of computing platforms and storage systems,
uniform mechanism to publish and discover
geospatial data, performance issues, and security
and access control issues Some efforts within the
geospatial society have been taken to leverage
the Grid technology for the sharing of distributed
data With the maturation of Grid technology, the
integration and sharing of distributed geospatial
data will be easier and more efficient
references
Allcock, B., Bester, J., Bresnahan, J.,
Cherve-nak, L A., Foster, I., Kesselman, C., Meder, S.,
Nefedova, V., Quesnel, D., & Tuecke, S (2002,
May) Data Management and Transfer in High
Performance Computational Grid Environments
Parallel Computing Journal, 28(5), 749-771.
Berman, F., Fox, G., & Hey, T., (2003) The Grid:
past, present, future In Berman, F., Fox, G., and
Hey, A eds, Grid Computing: Making the Global
Infrastructure a Reality, 9-50 Wiley, New York,
NY, USA
Chervenak, A., Foster, I., Kesselman, C., bury, C., & Tuecke, S (2001) The Data Grid: Towards an Architecture for the Distributed Management and Analysis of Large Scientific
Salis-Datasets Journal of Network and Computer
Ap-plications, 23, 187-200.
Di, L (2005) The Geospatial Grid In Rana, S and Sharma, J (eds.), Frontiers of Geographic
Information Technology Springer-Verlag.
Foster, I., Kesselman, C., & Tuecke, S., (2001) The Anatomy of the Grid: Enabling Scalable
Virtual Organizations International Journal
Karimi, A H & Peachavanish, R., (2005) teroperability in Geospatial Information Systems
In-In Khosrow-Pour, M (eds.), Encyclopedia of
Information Science and Technology Hershey,
PA: Idea Group Reference
Lamberti, F., & Beco, S., (2002) SpaceGRID -
An international programme to ease access and dissemination of Earth Observation data/prod-ucts: How new technologies can support Earth
Observation Users Community 22nd EARSeL
Symposium & General Assembly, Prague, Czech
Republic, June 4-6, 2002
Lo, C P., & Yeung, A K W., (2002) Concepts and
techniques of geographic information systems
Upper Saddle River, NJ: Prentice Hall
Nebert, D., & Whiteside, A., 2005 OGCTM logue Services Specification (Version 2.0.0) OGC Document Number: 04-021r3, 187pp
Cata-Ramapriyan, H., Isaac, D., Yang, W., Bonnlander, B., & Danks, D., (2006) An Intelligent Archive Testbed Incorporating Data Mining – Lessons and
Trang 5Sharing of Distributed Geospatial Data through Grid Technology
Observations IEEE International Geoscience and
Remote Sensing Symposium (IGARSS) 2006 July
3- August 4, 2006, Denver, Colorado
Wei, Y., Di, L., Zhao, B., Liao, G., Chen, A., Bai,
Y., & Liu, Y (2005) The design and
implementa-tion of a Grid-enabled catalogue service IEEE
International Geoscience and Remote Sensing
Symposium (IGARSS) 2005 on July 25-29, 2005,
Seoul, Korea
Welch, V (2005) Globus Toolkit Version 4 Grid
Security Infrastructure: A Standards
Perspec-tive
Welch, V., Siebenlist, F., Foster, I., Bresnahan, J.,
Czajkowski, K., Gawor, J., Kesselman, C Meder,
S., Pearlman, L., & Tuecke, S (2003) Security for
Grid Services Twelfth International Symposium
on High Performance Distributed Computing
(HPDC-12), IEEE Press.
key ter Ms
Certificate: A public key and information
about the certificate owner bound together by
the digital signature of a CA In the case of a CA
certificate the certificate is self signed, i.e., it was
signed using its own private key
Data Replica: A complete or partial copy of
original data
DPSS: The Distributed-Parallel Storage
System (DPSS) is a scalable, high-performance,
distributed-parallel data storage system orginally
developed as part of the DARPA -funded MAGIC
Testbed, with additional support from the U.S
Dept of Energy, Energy Research Division,
Mathematical, Information, and Computational
Sciences Office
Grid Technology: Grid technology is an
emerging computing model that provides the ability to perform higher throughput computing
by taking advantage of many networked ers to model a virtual computer architecture that
comput-is able to dcomput-istribute process execution across a parallel infrastructure
GridFTP: Extension of traditional FTP
pro-tocol It is a uniform, secure, high-performance interface to file-based storage systems on the Grid
HPSS: High Performance Storage System
(HPSS) is hierarchical storage system software that manages and accesses terabytes to petabytes
of data on disk and robotic tape libraries
OGSA-DAI: Open Grid Services Architecture
– Data Accessing Interface It is a middleware product which supports the exposure of data resources, such as relational or XML databases,
on to Grids
SRB: The Storage Resource Broker (SRB)
is a Data Grid Management System (DGMS) or simply a logical distributed file system based on
a client-server architecture which presents the user with a single global logical namespace or file hierarchy
Virtual Organization: A Virtual
Organiza-tion is a group of individuals or instituOrganiza-tions who share the computing resources of a “Grid” for a common goal
X.509: In cryptography, X.509 is an ITU-T
standard for public key infrastructure (PKI) X.509 specifies, amongst other things, standard formats for public key certificates and a certifica-tion path validation algorithm
Trang 6Section VILocation-Based Services
Trang 70
Chapter XXIX Cognitively Ergonomic Route
an increasing number of behavioral studies have, for example, pointed to the following characteristics: the use of landmarks, changing levels of granularity, the qualitative description of spatial relations The authors detail these aspects and additionally introduce formal approaches that incorporate them
to automatically provide route directions that adhere to principles of cognitive ergonomics
c ogn It Ive Aspects of r oute
dIrect Ions
Route directions fascinate researchers in several
fields Since the 70s linguists and cognitive
scien-to cognition scien-to learn about cognitive processes that reflect structuring principles of environmen-tal knowledge (e.g., Klein, 1978) Over the last decade, the number of publications on various aspects of route directions has increased Next to
Trang 8
Cognitively Ergonomic Route Directions
tions and how to identify principles that allow us
to define what makes route directions cognitively
ergonomic, technical aspects of navigation support
systems have become an additional focus The
question required from the latter perspective is
part of a broader approach that aims to formally
characterize the meaning (semantics) of spatial
relations In other words, if we want to bridge the
gap between information systems and behavioral
analysis we have to answer how we perform the
transition from data to knowledge
Several key elements can be identified based
on psychological and linguistic literature on route
directions that are pertinent for cognitively
ergo-nomic route directions (Denis, 1997; Lovelace,
Hegarty, & Montello, 1999; Tversky & Lee,
1999) These comprise the conceptualization of
directions at decision points, the spatial chunking
of route direction elements to obtain hierarchies
and to change the level of granularity, the role
of landmarks, the communication in different
modalities, the traveling in different modes, and
aspects of personalization (see Table 1) Most
research on routes and route directions deals
with navigation in urban structures such as street
networks The results discussed in this article
focus on this domain
Appro Aches t o r epresent Ing
r oute k now Ledge
Behavioral studies have substantiated key ments of cognitively ergonomic route directions
ele-To implement these aspects in information systems detailed formal characterizations of route knowl-edge are required The approaches discussed below are a representative vocabulary that allows for the characterization of mental conceptualiza-tion processes reflecting the results from behav-ioral studies (see Table 1) In this sense we can
refer to them as Ontologies of Route Knowledge
(Chandrasekaran, Josephson, & Benjamins, 1999; Gruber, 1993) In Guarino’s terminology these
approaches would most likely be called domain
ontologies (Guarino, 1998).
One of the earliest approaches is the TOUR
model by Kuipers (Kuipers, 1978) that later
devel-oped into the Spatial Semantic Hierarchy (SSH)
(Kuipers, 2000) Kuipers and his collaborators developed this approach to add the qualitative-ness that can be found in the organization of a cognitive agent’s spatial knowledge to approaches
in robotics The latter classically relied more on quantitative spatial descriptions The SSH al-lows for modeling cognitive representations of space as well as for building a framework for robot navigation, i.e qualitative and quantita-
Table 1 Cognitive ergonomics of route directions
Cognitively ergonomic route directions
• are qualitative, not quantitative,
• allow for different levels of granularity and organize spatial knowledge hierarchically,
• reflect cognitive conceptualizations of directions at decision points,
• chunk route direction elements into larger units to reduce cognitive load,
• use landmarks to:
° disambiguate spatial situations,
° anchor turning actions,
° and to confirm that the right actions have been taken,
• present information in multimodal communication systems allowing for an interplay of language and graphics, but respecting for the underlying conceptual structure,
• allow for an adaptation to the user’s familiarity with an environment, as well as personal styles and different languages.
Trang 9Cognitively Ergonomic Route Directions
tive aspects are combined The SSH especially
reflects the aspect of hierarchical organization of
spatial knowledge by providing different levels of
information representation: the sensory, control,
causal, topological, and metrical level Ontological
characterizations are developed for each level to
match human cognitive processes
The Route Graph model (Werner,
Krieg-Brückner, & Herrmann, 2000) describes key
elements for route based navigation Similar to
the SSH, it allows representing knowledge on
different levels of granularity However, it is
much more abstract and does not provide any
processes for acquiring this knowledge It is
intended to provide a formalism expressing key
notions of route knowledge independent of a
particular implementation, agent, or domain Its
focus is on a sound formal specification of basic
elements and operations, like the transition from
route knowledge to survey knowledge by merging
routes into a graph-like structure
A linguistically grounded approach with the
aim to generate verbal route directions is the
CORAL project by Dale and coworkers (e.g.,
Dale, Geldof, & Prost, 2005) One of the central
aspects of their approach is the organization of
parts of a route into meaningful units, a process
they call segmentation Instead of providing
turn-by-turn directions, this approach allows for
a small number of instructions that capture the
most important aspects of a route The employed
modeling language is called Route Planning
Markup Language (RPML)
Formalisms that model route knowledge on
the conceptual level can be found in the theory of
wayfinding choremes (Klippel, Tappe, Kulik, &
Lee, 2005) and context-specific route directions
(Richter & Klippel, 2005) These approaches
model route knowledge modality-independent
on the conceptual level The wayfinding choreme
theory employs conceptual primitives—as the
result of conceptualization processes of a
cogni-tive agent incorporating functional as well as
basic as well as super-ordinate valid expressions
on different levels of granularity The approach
to context-specific route directions builds on this theory A systematics of route direction ele-ments determines which, and how, entities may
be referred to in route directions Accordingly, abstract relational specifications are inferred by optimization processes that adapt route directions
to environmental characteristics and inherent route properties
Human wayfinding, however, may not be restricted to a single mode of transportation
A typical example is public transport, where travelers frequently switch between pedestrian movement and passive transportation (trains, buses, etc.) Timpf (2002) analyzed route direc-tions for multi-modal wayfinding and developed two different ontologies of route knowledge: one representing knowledge from the perspective of the traveler and one taking the perspective of the transportation system The former focuses
on movement along a single route, i.e., actions
to perform to reach the destination, while the latter provides concepts referring to the complete transportation network
An industry approach for formalizing route knowledge can be found in Part 6: Navigation
Service of the OpenLS specification The
Open-GIS Location Services (OpenLS) Implementation Specification (Mabrouk, 2005) describes an open platform for location-based application services, the so called GeoMobility Server (GMS) proposed
by the Open Geospatial Consortium (OGC) It offers a framework for the interoperable use of mobile devices, services and location-related
data The Navigation Service described in Part 6
of the OpenLS specification provides the ing client, amongst other services, with prepro-cessed data that is required for the generation of route directions Based on XML specifications,
access-it defines a data structure that allows clients to generate their own route directions which may accord more to a user’s preferences The used
Trang 10
Cognitively Ergonomic Route Directions
(descriptions combining a turn at a decision point
and proceeding on the following route segment)
and enhances them with additional information
about route elements
c ore Aspects of c ogn It Ive Ly
ergono MIc r oute dIrect Ions
In the following, three aspects that are at the core
of cognitively ergonomic route directions will be
discussed in greater detail: cognitively adequate
direction concepts, the use of landmarks, and
spatial chunking to obtain hierarchies and change
the level of granularity
Conceptualization of Directions at
decision points
The specification of direction changes is the most
pertinent information in route directions While
current route information systems heavily rely on
street names to identify the proper direction to
take, behavioral research (Tom & Denis, 2003)
has shown that from a cognitive perspective, street names are not the preferred means to re-orient oneself People rather rely on landmarks (as discussed in the next section) and appropriate direction concepts On the most basic level we have to specify the correspondence between a direction change (in terms of the angle) and a direction concept For example, which sector is applicable to a concept like “turn right”? On a more elaborate level, we have to specify alterna-tive direction concepts and detail their scope
of application Figure 1 shows some examples
of how the same direction change can result in different direction concepts (and corresponding verbalizations) depending, among other things, on the spatial structure in which the change occurs
We need this level of specificity for two reasons First, a qualitative but precise direction model allows for verbally instantiating a situation model (Zwaan & Radvansky, 1998) of the encountered intersections Second, intersections can function
as landmarks Just like classical examples of marks, such as the Eiffel Tower, in the context
land-of a specific route, a salient intersection can be
Figure 1 A change of a direction is associated with different conceptualizations according to the section at which it takes place The ‘pure’ change may be linguistically characterized as take the second exit at the roundabout (a) At intersection (b) it might change to the second right; at intersection (c) it may change to fork right, and at (d) it becomes veer right.
Trang 11inter-Cognitively Ergonomic Route Directions
used to organize spatial knowledge This aspect
has not yet gained much attention
Enriching Route Directions with
Landmarks
Analyzing human route directions shows how
prominently landmarks are used to structure
the respective spatial knowledge, to give the
instructed the possibility to assure that they are
still following the correct route, and to anchor
required turning actions Since landmarks seem
to be such an important part of human-generated
route directions their integration is pertinent for
automatically generating cognitively ergonomic
instructions
Several classifications of landmarks and their
characteristics have been discussed in the
litera-ture One of the first assessments is presented by
Lynch (1960) who distinguishes Landmarks as one
of five elements that structure urban knowledge:
path, edges, districts, nodes, and landmarks It
is commonly agreed that the landmark account
should comprise all five elements, as according
to Presson and Montello (1988) everything that
stands out of the background may serve as a
landmark That is, given the right spatial context
different features of an environment may serve as
landmarks Sorrows and Hirtle (1999) distinguish
three characteristics important for making an
ob-ject a landmark: its visual, semantic, and structural
characteristics Additionally, landmarks can be
categorized according to their cognitive function
within route directions, their geometry, and their
spatial relation to the route Humans conceptualize
landmarks either as point-like, linear, or area-like
entities However, these conceptualizations do not
necessarily correspond to the geometric
charac-teristics of objects but reflect the schematization
processes cognitive agents apply (Herskovits,
1986) A detailed description of the different
roles of landmarks is necessary to allow for their
integration in an automatic generation process
way to enrich route directions with landmarks
is to include references to salient intersections, like T-intersections or roundabouts, which are easy to identify automatically This also reflects the direction concepts humans employ with such structures (see also Figure 1)
Spatial Chunking: Hierarchies and Levels of Granularity
The hierarchical organization of spatial mation and flexibly changing between levels
infor-of granularity are omnipresent in the cognitive organization of spatial knowledge (Hobbs, 1985; Kuipers, 2000) Chunking elementary wayfinding actions (such as turns at intersections) in order
to impose a hierarchical structure and to change the level of granularity reflects not only cogni-tive conceptualization processes but organizes route knowledge in a cognitively ergonomic way Especially users who are familiar with an environment can profit from such an approach
In general, providing a user with too much detail violates findings of cognitive science, as for ex-
ample formulated in Clark’s 007 Principle: “In
general, evolved creatures will neither store nor process information in costly ways when they can use the structure of the environment and their operations upon it as a convenient stand-in for the information-processing operations concerned That is, know only as much as you need to know
to get the job done.” (Clark, 1989, p 64)Structuring route descriptions by subsuming instructions gives users a coarse overview over a route, which is easier to perceive and quite often sufficient for successful wayfinding, especially
if the user is familiar with the environment Of course, the subsumed information still has to be accessible in case the user needs it (or, as discus-sions on positioning technologies in this volume show, the user may simply re-query a new route from his new position) This may either be pos-sible by zoom-in operations, i.e., by accessing the
Trang 12
Cognitively Ergonomic Route Directions
(mental) inference processes Such inferences, for
example, extract from an instruction like “turn left
at the dead end” information on which action to
perform at all intersections before the dead end,
namely to continue straight (e.g., Duckham &
Kulik, 2003) The following cognitive strategies
for spatial chunking are discussed in the
litera-ture (Dale et al., 2005; Klippel, Tappe, & Habel,
2003): numerical chunking, structure chunking,
landmark chunking, and chunking using the street
level hierarchy
t he MuLt IMod AL present At Ion
of r oute k now Ledge
The multimodal communication of spatial
in-formation is a core aspect of human cognition:
linguistic expressions, graphical representations
such as sketch maps, and gestures are channels
along which humans naturally communicate
(Ovi-att, 2003) Each representational medium—each
channel—has advantages in specific contexts but
may fail in other situations (Kray, Laakso, Elting,
& Coors, 2003) For example, natural language
expressions are inherently underspecified: a term
like turn right is applicable to a range of different
turning angles at an intersection and therefore
may be sufficient in many situations Figrue 2,
however, shows a situation that requires a complex
explanation if a description is provided in tic terms In this case, a graphic representation
linguis-is more suitable to communicate the situation at hand Communication channels also differ with respect to their suitability in the identification
of landmarks A salient object at an intersection
might be visually easily identifiable and nisable, but hard to describe linguistically An
recog-expression like follow the road to the dead end
on the other hand, may chunk a large part within
a route linguistically and therefore, communicate the spatial situation more efficiently if the dead end is a long way away and hard to depict on a small screen
The communication of route information, whether visually, linguistically, or in any other modality, has to follow the same guidelines as established for the structuring of route knowledge Cluttering any communication process has shown
to violate cognitive ergonomics and to slow down information processing This confinement
to sparseness has been shown for visual route directions, for example, by Agrawala and Stolte (2000), who based their route direction tool
on results obtained from sketch maps (Tversky
& Lee, 1999)
suMMAr y
In the last decades, research on route directions in linguistics and cognitive science revealed many underlying principles and processes of human route direction production and comprehension, and, thus, provides us with an understanding
of what constitutes cognitively ergonomic route directions However, this understanding has to
be formally specified to be implemented in formation systems for wayfinding assistance, like internet route-planners In essence, three cognitive principles need to be implemented in wayfind-ing assistance systems to generate cognitively ergonomic route directions: adequate direction concepts, the enrichment of route directions with
in-Figure 2 Complex intersection
Trang 13Cognitively Ergonomic Route Directions
landmarks, and spatial chunking which allows for
a hierarchical structuring of route knowledge and
representations on different levels of granularity
To this end, we need a thorough understanding of
which direction concept humans apply in which
situation, a detailed ontology of the different kinds
of landmarks and the role they may take in route
directions, as well as formal characterizations
that model hierarchical structures and guide the
changes of granularity
r eferences
Agrawala, M., & Stolte, C (2000) A design and
implementation for effective
computer-gener-ated route maps In AAAI Symposium on Smart
Graphics, March 2000 Stanford.
Chandrasekaran, B., Josephson, J R., &
Ben-jamins, V R (1999) What are ontologies, and
why do we need them? IEEE Intelligent Systems
and Their Applications, 14(1), 20-26.
Clark, A (1989) Microcognition: Philosophy,
cognitive science, and parallel distributed
pro-cessing Cambridge, MA: MIT Press.
Dale, R., Geldof, S., & Prost, J.-P (2005) Using
natural language generation in automatic route
description Journal of Research and practice in
Information Technology, 37(1), 89-105.
Denis, M (1997) The description of routes: A
cognitive approach to the production of spatial
discourse Cahiers de Psychologie Cognitive,
16, 409-458.
Duckham, M., & Kulik, L (2003) “Simples”
paths: Automated route selection for navigation
In W Kuhn, M Worboys, & S Timpf (Eds.),
Spatial Information Theory: Foundations of
Geographic Information Science Conference on
Spatial Information Theory (COSIT) 2003 (pp
182-199) Berlin: Springer
Gruber, T R (1993) A translation approach to
portable ontologies Knowledge Acquisition, 5(2),
199-220
Guarino, N (1998) Formal ontology and
infor-mation systems In N Guarino (Ed.), Formal
Ontology in Information Systems Proceedings of FOIS’98, Trento, Italy, 6-8 June 1998 (pp 3-15)
Amsterdam: IOS Press
Herskovits, A (1986) Language and Spatial
Cognition: An Interdisciplinary Study of the Representation of the Prepositions in English
Cambridge, UK: Cambridge University Press.Hobbs, J R (1985) Granularity In A Joshi (Ed.),
Proceedings of the 9th International Joint ference on Artificial Intelligence Los Angeles,
Con-CA (pp 432-435) San Francisco, Con-CA: Morgan
Kaufmann
Klein, W (1978) Wegauskuenfte Zeitschrift für
Literaturwissenschaft und Linguistik, 33, 9-57.
Klippel, A., Tappe, T., & Habel, C (2003) torial representations of routes: Chunking route segments during comprehension In C Freksa, W
Pic-Brauer, C Habel & K F Wender (Eds.), Spatial
Cognition III Routes and Navigation, Human Memory and Learning, Spatial Representa- tion and Spatial Learning (pp 11-33) Berlin:
Springer
Klippel, A., Tappe, T., Kulik, L., & Lee, P U (2005) Wayfinding choremes - A language for
modeling conceptual route knowledge Journal
of Visual Languages and Computing, 16(4),
311-329
Kray, C., Laakso, K., Elting, C., & Coors, V
(2003) Presenting route instructions on mobile
devices Paper presented at the IUI’03, January
12-15, 2003, miami, Florida, USA
Kuipers, B (1978) Modelling spatial knowledge
Cognitive Science, 2(2), 129-154.
Trang 14
Cognitively Ergonomic Route Directions
Kuipers, B (2000) The spatial semantic hierarchy
Artificial Intelligence, 119, 191-233.
Lovelace, K., Hegarty, M., & Montello, D R
(1999) Elements of good route directions in
familiar and unfamiliar environments In C
Freksa & D M Mark (Eds.), Spatial information
theory Cognitive and computational foundations
of geographic information science (pp 65-82)
Belin: Springer
Lynch, K (1960) The image of the city
Cam-bridge, MA: MIT Press
Mabrouk, M (2005) OpenGis Location Services
(OpenLS): Core Services OGC Implementation
Specification 05-016 Version 1.1 Open Gis
Con-sortium Inc.
Oviatt, S L (2003) Multimodal interfaces In J
Jacko & A Sears (Eds.), The Human-Computer
Interaction Handbook: Fundamentals, Evolving
Technologies and Emerging Applications (pp
286-304) Mahwah, NJ: Lawrence Erlbaum
Presson, C C., & Montello, D R (1988) Points
of reference in spatial cognition: Stalking the
elu-sive landmark British Journal of Developmental
Psychology, 6, 378-381.
Richter, K.-F., & Klippel, A (2005) A model for
context-specific route directions In C Freksa,
M Knauff & B Krieg-Brueckner (Eds.), Spatial
Cognition IV Reasoning, Action, and
Interac-tion: International Conference Spatial
Cogni-tion 2004, Frauenchiemsee, Germany, October
11-13, 2004, Revised Selected Papers (pp 58-78)
Berlin: Springer
Sorrows, M., & Hirtle, S C (1999) The nature
of landmarks for real and electronic spaces In C
Freksa & D M Mark (Eds.), Spatial information
theory Cognitive and computational foundations
of geographic information science (pp 37-50)
Berlin: Springer
Timpf, S (2002) Ontologies of wayfinding: A
traveler’s perspective Networks and Spatial
Environments, 2, 3-33.
Tom, A., & Denis, M (2003) Referring to mark or street iniformation in route directions: What difference does it make? In W Kuhn, M
land-Worboys & S Timpf (Eds.), Spatial information
theory Foundations of geogrpahic information science International conference, COSIT 2003, Kartause Ittingen, Switzerland, September 2003
(pp 362-374) Berlin: Springer
Tversky, B., & Lee, P U (1999) Pictorial and verbal tools for conveying routes In C Freksa &
D M Mark (Eds.), Spatial information theory
Cognitive and computational foundations of geographic information science (pp 51-64)
Berlin: Springer
Werner, S., Krieg-Brückner, B., & Herrmann,
T (2000) Modeling navigational knowledge by route graphs In C Freksa, W Brauer, C Habel
& K F Wender (Eds.), Spatial cognition II
Integrating abstract theories, empirical studies, formal methods, and practical applications (pp
mation systems that places a strong emphasis on cognitive aspects In the case of route directions the design aims for a lower cognitive load and enhanced location awareness at the same time
Granularity: Here, it refers to the detail in
route directions; from coarse levels for general planning to finer levels to provide context-specific information, for example at decision points
Landmark: Any entity in the environment
that sticks out from the background
Trang 15Cognitively Ergonomic Route Directions
OpenLS: Specification of an open platform
for location-based services defining their core
functionality (directory service, gateway service,
location utility service, presentation service, route
service)
Personalization: Adaptation of information
presentation and interaction with a device /
soft-ware to the needs and preferences of a specific,
individual user
Route Directions: A set of instructions that
allow a wayfinder in known or unknown
envi-ronments to follow a route from a start point to
a destination
Spatial Semantic Hierarchy (SSH): A
com-putational model defining acquisition and sentation of spatial knowledge on different levels
repre-of abstraction ranging from sensory information
to topological knowledge
Wayfinding: The cognitive conceptual
activ-ity of planning and finding ones way
Wayfinding Choremes: Mental
conceptu-alizations of functional wayfinding and route direction elements
Trang 16
Chapter XXX Multicast Over Location-Based
Budapest University of Technology and Economics, Hungary
Copyright © 2009, IGI Global, distributing in print or electronic forms without written permission of IGI Global is prohibited.
Abstr Act
This chapter details the potential found in combining to different technologies The two basically ferent technologies, LBSs in mobile communication and the well-elaborated multicast technology are merged in the multicast via LBS solutions As this chapter demonstrates, this emerging new area has a lot of possibilities, which have not been completely utilized.
Currently, an important area of mobile
commu-nication is ad-hoc computer networks, where
mobile devices need base stations however they form an overlay without any Internet-related infrastructure, which is a virtual computer net-work among them In this case, the selective,
Trang 17Multicast Over Location-Based Services
bAckground
The positioning technologies in the LBS solutions are based on the various distances of the commu-nication mobile from the different base stations With advances in automatic position sensing and wireless connectivity, the application range of mobile LBSs is rapidly developing, particularly
in the area of geographic, tourist and local travel information systems (Ibach et al., 2005) Such systems can offer maps and other area-related information The LBS solutions give the capability
to deliver location-aware content to subscribers on the basis of the positioning capability of the wire-less infrastructure The LBS solutions can push location-dependent data to mobile users according
to their interest or the user can pull the required information by sending a request to a server that provides location-dependent information.LBSs process information with respect to the location of one or several persons, also referred to
as targets before presenting it to the user In recent
years, LBSs have become increasingly important and have helped accelerate the development to-wards ubiquitous computing environments Tradi-tional LBSs map targets to locations (e.g., Where
is person X located?), i.e., they find the position
of a specific person or group of people This type
of LBS is denoted as Tracking Services.
There are a lot of location positioning ods and technologies, such as the satellite-based
meth-Global Positioning System (GPS) that is widely
applied (Hofmann-Wellenhof et al., 1997) The location determination methods that do not use the GPS can be classified into three categories:
Proximity, Triangulation (lateration), and Scene analysis or pattern recognition (Hightower &
Borriello, 2001) Signal strength is frequently applied to determine proximity As a proximity measurement, if a signal is received at several known locations, it is possible to intersect the coverage areas of that signal to calculate a location area If one knows the angle of bearing (relative
location-related communication has not been
solved completely
Traditional Location-Based Services (LBSs)
determine the current location of a given person
or a given group of people in order to process
location-dependent information This use does
not cover the full range that is conceivable for
these services This article introduces so-called
Zone Services as a new sub-category of LBSs In
contrast to traditional LBSs, Zone Services
col-lect information about persons currently located
in a given geographic area For these services,
new considerations regarding data collection,
privacy, and efficiency have to be made Hence, it
has to be determined what techniques or
mecha-nisms common in traditional LBSs or in other
areas like databases or mobile communication
systems can be reused and what concepts have
to be developed
One of the various communication models
among software entities is the one-to-many data
dissemination, called multicast The multicast
communication over mobile ad-hoc networks
has increasing importance (Hosszú, 2005) The
article described the fundamental concepts and
solutions on the area of LBSs and the possible
multicasting over the LBS systems This kind
of communication is in fact a special case of the
multicast communication model, called geocast,
where the sender disseminates data to a subset
of the multicast group members that are in a
specific geographical area This chapter shows
that this special kind of multicast utilizes the
advantages of LBSs, since multicast is based on
location-aware information that is available in
location-based solutions
The two basically different technologies, LBSs
in mobile communication and the well-elaborated
multicast technology are merged in the multicast
via LBS solutions As the chapter demonstrates,
this emerging new area has a lot of possibilities,
which has not been completely utilized
Trang 18
Multicast Over Location-Based Services
the target device, then the target location can be
accurately calculated Similarly, if somebody
knows the range from three known positions
to a target, then the location of the target object
can be determined A GPS receiver uses range
measurements to multiple satellites to calculate
its position The location determination methods
can be server-based or client-based according
to the place of computation (Hightower &
Bor-riello, 2001)
LBSs utilize their ability of
location-aware-ness to simplify user interactions With advances
in wireless connectivity, the application range of
mobile LBSs is rapidly developing, particularly in
the field of tourist information systems - telematic,
geographic, and logistic information systems
However, current LBS solutions are
incompat-ible with each other since manufacturer-specific
protocols and interfaces are applied to aggregate
various components for positioning, networking,
or payment services In many cases, these
com-ponents form a rigid system If such a system has
to be adapted to another technology, e.g., moving
from GPS based positioning to in-house IEEE
802.11a-based Wireless Local-Area Network
(WLAN) or Bluetooth based positioning, it has
to be completely redesigned (Haartsen, 1998)
As such, the ability of interoperation of
differ-ent resources under changeable interconnection
conditions becomes crucial for the end-to-end
availability of the services in mobile environments
(Ibach, & Horbank, 2004)
Chen et al (2004) introduces an enabling
in-frastructure, which is a middleware, in order to
support LBSs This solution is based on a Location
Operating Reference Model (LORE) that solves
many problems of constructing LBSs, including
location modeling, positioning, tracking,
dependent query processing and smart
location-based message notification Another interesting
solution is the mobile yellow page service
An interesting development is the Compose
project, which aims to overcome the drawbacks
of the current solutions by pursuing a service
integrated approach that encompasses pre-trip and on-trip services where on-trip services could be split into in-car and last-mile services
(Bocci, 2005) The pre-trip service means the
3D navigation of the users in a city
environ-ment, and the on-trip service means the in-car and the last-mile services together The in-car
service is a composition of an LBS and a satellite
broadcasting/multicasting method In this case,
the user has wireless-link access by Personal
Digital Assistant (PDA) to broadcast or multicast
The last-mile service helps the mobile user with
PDAs to receive guidance during the final part
of the journey
The article focuses on the multicast solutions over the current LBS solutions This kind of com-munication is in fact a special case of the multicast communication model, called geocast, where the sender disseminates the data to a subset of the multicast group members that are in a specific geographical area
MuLt Ic Ast Ing
The models of multicast communication differ in the realization of the multiplication function in the intermediate nodes In the case of Datalink-Level the intermediate nodes are switches, on the Network-Level they are routers and on the Application-Level the fork points are applica-tions on hosts
The Datalink-Level based multicast is not ible enough for new applications, which is why it
flex-has no practical importance The Network-Level
Multicast (NLM), known as IP-multicast, is well
elaborated and sophisticated routing protocols are developed for it However, it has not yet been
widely deployed since routing among the
Autono-mous Systems (AS) has not been solved perfectly
The application level solution gives less efficiency compared to the IP-multicast, however, its deploy-ment depends on the application itself and it has
no influence on the operation of the routers That
Trang 19Multicast Over Location-Based Services
is why the Application-Layer Multicast (ALM)
is currently increasing in importance
There are a lot of various protocols and
imple-mentations of the ALM, some of which are
suit-able for communication over wireless networks,
which enhance the importance of the ALM The
reason for this is that in the case of mobile devices
the importance of ad-hoc networks is increasing
Ad-hoc is a network that does not need any
infra-structure Such networks are Bluetooth (Haartsen,
1998) and Mobile Ad Hoc NETwork (MANET),
which comprise a set of wireless devices that can
move around freely and communicate in relaying
packets on behalf of one another (Mohapatra et
al., 2004)
In computer networking, there is a weaker
definition of this ad-hoc network Ad-hoc is a
computer network that does not need a routing
infrastructure It means that the mobile devices
that use base stations can create ad-hoc computer
networks In such situations, the usage of
Applica-tion-Level Networking (ALN) technology is more
practical than IP-Multicast In order to support this
group communication, various multicast routing
protocols are developed for the mobile
environ-ment The multicast routing protocols for ad-hoc
networks differ in terms of state maintenance,
route topology and other attributes
The simplest ad-hoc multicast routing methods
are ooding and tree-based routing Flooding
is very simple, which offers the lowest control
overhead at the expense of generating high data
traffic This situation is similar to the traditional
IP-Multicast routing However, in a wireless
ad-hoc environment, the tree-based routing
fundamentally differs from a wired IP-Multicast
situation, where tree-based multicast routing
algorithms are obviously the most efficient ones,
such as in the Multicast Open Shortest Path First
(MOSPF) routing protocol (Moy, 1994) Though
tree-based routing generates optimally small
data traffic on the overlay in the wireless ad-hoc
network, the tree maintenance and updates need
a lot of control traffic That is why the simplest methods are not scalable for large groups
A more sophisticated ad-hoc multicast
rout-ing protocol is the Core-Assisted Mesh
Proto-col (CAMP), which belongs to the mesh-based
multicast routing protocols (Garcia-Luna-Aceves
& Madruga, 1999) It uses a shared mesh to support multicast routing in a dynamic ad-hoc environment This method uses cores to limit the control traffic needed to create multicast meshes Unlike the core-based multicast routing protocol
as the traditional Protocol Independent
Multi-cast-Sparse Mode (PIM-SM) multicast routing
protocol (Deering et al., 1996), CAMP does not require that all traffic flow through the core nodes CAMP uses a receiver-initiated method for routers
to join a multicast group If a node wishes to join the group, it uses a standard procedure to announce its membership When none of its neighbors are mesh members, the node either sends a join request toward a core or attempt to reach a group member using an expanding-ring search process Any mesh member can respond to the join request with a
join Acknowledgement (ACK) that propagates
back to the request originator
Compared to the mesh-based routing protocols, which exploit variable topology, the so-called gossip-based multicast routing protocols exploit randomness in communication and mobility Such multicast routing protocols apply gossip
as a form of randomly controlled flooding to solve the problems of network news dissemina-tion This method involves member nodes to talk periodically to a random subset of other members After each round of talk, the gossipers can recover their missed multicast packets from each other (Mohapatra et al., 2004) Compared
to the deterministic approaches, this probabilistic method will better work in a highly dynamic ad hoc network because it operates independently of network topology and its random nature fits the typical characteristics of the network
Trang 20
Multicast Over Location-Based Services
t he Loc At Ion-A wAre
MuLt Ic Ast
The geocasting can be combined with flooding
Such methods are called forwarding zone
meth-ods, which constrain the flooding region The
forwarding zone is a geographic area that extends
from the source node to cover the geocast zone
The source node defines a forwarding zone in the
header of the geocast data packet Upon receiving
a geocast packet, other machines will forward
it only if their location is inside the forwarding
zone The Location-Based Multicast (LBM) is
an example for such geocasting-limited flooding
(Ko & Vaidya, 2002)
An interesting type of ad-hoc multicasting is
the geocasting The host that wishes to deliver
packets to every node in a certain geographical
area can use such a method In this case, the
po-sition of each node with regard to the specified
geocast region implicitly defines group
member-ship Every node is required to know its own
geographical location For this purpose they can
use the GPS The geocasting routing method does
not require any explicit join or leave actions The
members of the group tend to be clustered both
geographically and topologically The
geocast-ing method of routgeocast-ing exploits the knowledge
of location
LbM geoc Ast Ing And Ip
MuLt Ic Ast Ing
Using LBM in a network where routers are in
fixed locations and their directly connected
hosts are within a short distance, the location of
these hosts can be approximated with the
loca-tion of their router These requirements are met
by most of the GSM (Global System for Mobile
Communications), UMTS (Universal Mobile
Telecommunication System), WIFI (Wi-Fi
Cer-tified), WIMAX (Worldwide Interoperability
for Microwave Access) and Ethernet networks,
therefore a novel IP layer routing mechanism can
be introduced
This new method is a simple kind of the
geo-casting-limited flooding, extending the normal
Multicast Routing Information Base (MRIB) with the geological location of the neighbor rout-ers Every router should know its own location, and a routing protocol should be used to spread location information between routers The new IP protocol is similar to the User Datagram Protocol (UDP) protocol, but it extends it with a source location and a radius parameter The source loca-tion parameter is automatically assigned by the first router When a router receives a packet with empty source location, it assigns its own location
to it The radius parameter is assigned by the plication itself, or it can be an administratively defined parameter
ap-This method requires changes in routing operational systems, but offers an easy way to start geocasting services on an existing IP in-frastructure without using additional positioning devices (e.g., GPS receiver) on every sender and receiver The real advantages of the method are that geocasting services can be offered for all existing mobile phones without any additional device or infrastructure
f uture t rends
The multicast communication over mobile hoc networks has increasing importance The article has described the fundamental concepts and solutions It especially focused on the area of LBSs and the possible multicasting over them It was shown that a special kind of the multicast,
ad-called geocast communication model utilizes
the advantages of LBSs, since it is based on the location-aware information made available in the location-based solutions
There are two known issues of this IP level geocasting The first problem is the scalability, the flooding type of message transfer is less robust as
Trang 21Multicast Over Location-Based Services
compared to multicast tree based protocols, but
this method is more efficient in a smaller
envi-ronment than using tree allocation overhead of
multicast protocols The second issue is that the
source must be connected directly to the router
that is physically in the center position in order to
become source of a session The proposed
geocast-ing-limited flooding protocol should be extended
to handle those situations where the source of a
session and the target geological location are in
different places
c onc Lus Ion
The two basically different technologies, the
Location-Based Services in the mobile
commu-nication world and the well-elaborated multicast
communication technology of the computer
networking are jointed in the multicast over LBS
solutions As it was described, this emerging new
area has a lot of possibilities, which have not been
completely utilized
As a conclusion it can be stated that despite
the earlier predicted slower development rate of
the LBS solutions, nowadays the technical
possi-bilities and the consumers’ demands have already
met The geospatial property of LBSs provides
technical conditions to apply a specialized type
of the multicast technology, called geocasting,
which gives an efficient and user group targeted
solution for one-to-many communication
r eferences
Bocci, L (2005) Compose Project Web Site,
Retrieved from http://www.newapplication
it/compose
Chen, Y., Chen, Y Y., Rao, F Y., Yu, X L., Li,
Y., & Liu, D (2004) LORE: An Infrastructure to
Support Location-aware services IBM Journal
Deering, S E., Estrin, D., Farinacci, D., Jacobson, V., Liu, C-G., & Wei, L (1996) The PIM architec-
ture for wide-area multicast routing IEEE/ACM
Trans on Networking, 4(2), 153-162.
Garcia-Luna-Aceves, J J., & Madruga, E L (1999,
August) The Core-Assisted Mesh Protocol IEEE
Journal of Selected Areas in Communications,
1380-1394
Haartsen, J (1998) The universal radio interface
for ad hoc, wireless connectivity Ericsson
Re-view,3 Retrieved 2004 from http://www.ericsson.
com/reviewHightower, J., & Borriello, G (2001 Aug.) Loca-
tion Systems for Ubiquitous Computing IEEE
Computer, 57-65.
Hofmann-Wellenhof, B., Lichtenegger, H., &
Collins, J (1997) Global Positioning System:
Theory and Practice Fourth Edition,
Springer-Verlag Wien, New York, NY
Hosszú, G (2005) Reliability Issues of the ticast-based Mediacommunication In Pagani, M
Mul-(Ed.), Encyclopedia of Multimedia Technology
and Networking (pp 875-881) Hershey, PA: Idea
Group Reference
Ibach, P., Horbank, M (2004) Highly-Available
Location-based Services in Mobile Environments
Paper presented at the International Service ability Symposium 2004, Munich, Germany, May 13-14
Avail-Ibach, P., Tamm, G., & Horbank, M (2005) namic Value Webs in Mobile Environments Using
Dy-Adaptive Location-Based Services In
Proceed-ings of the 38th Hawaii International Conference
on System Sciences (9 pages) IEEE.
Ko, Y-B., & Vaidya, N H (2002) Flooding-Based Geocasting Protocols for Mobile Ad Hoc Net-
works Proceeding of the Mobile Networks and
Applications Kluwer Academic, 7(6), 471-480.
Trang 22
Multicast Over Location-Based Services
LocatioNet, (2006) LocatioNet and Ericsson
En-ter Into Global Distribution Agreement Retrieved
from http://www.locationet.com
Mohapatra, P., Gui, C., & Li, J (2004) Group
Communications in Mobile Ad Hoc Networks
Computer, 37(2), 52-59.
Moy, J (March 1994) Multicast extensions to
OSPF Network Working Group RFC 1584.
Uppuluri, P., Jabisetti, N., Joshi, U., & Lee, Y
(2005) P2P Grid: Service Oriented Framework for
Distributed Resource Management In
Proceed-ings of the IEEE International Conference on Web
Services, July 11-15, Orlando, FL, USA.
Weiss, D (2006) Zone Services — A New
Ap-proach of Location-based Services, Retrieved
from http://www.pervasive2006.org/
key t er Ms
Ad-hoc Computer Network: Mobile devices
that require base stations can create the ad-hoc
computer network if they do not need routing
infrastructure
Application-Layer Multicast (ALM): A
novel multicast technology, which does not require
any additional protocol in the network routers,
since it uses the traditional unicast (one-to-one) IP
transmission Its other name: Application-Level
Multicast (ALM).
Application-Level Network (ALN): The
applications, which are running in the hosts, can
create a virtual network from their logical
con-nections This is also called overlay network The
operations of such software entities are not able
to understand without knowing their logical
rela-tions In most cases these ALN software entities
use the P2P model (see below), not the client/
server (see below) for the communication.
Autonomous System (AS): A network with
common administration; it is a basic building element of the Internet Each AS is independent from the others
Client/Server Model: It is a communicating
model, where one hardware or software entity (server) has more functionalities than the other entity (the client), whereas the client is responsible
to initiate and close the communication session towards the server Usually the server provides services that the client can request from the server
Its alternative is the P2P model (see below).
Geocast: One-to-many communications
among communicating entities, where an entity
in the root of the multicast distribution tree sends data to that certain subset of the entities in the multicast dissemination tree, which are in a spe-cific geographical area
IP-Multicast: Network-level multicast
tech-nology, which uses the special class-D IP-address range It requires multicast routing protocols in the
network routers Its other name: Network-Level
Multicast (NLM).
Multicast Routing Protocol: In order to
forward the multicast packets, the routers have
to create multicast routing tables using multicast routing protocols
Multicast Tree: A virtual graph, which
gives the paths of sending multicast data from the source (the root of the tree) to the nodes of
the tree Its other name: Dissemination tree or
Trang 23
Chapter XXXI Routing
to fit many different application areas, including shortest path problems, vehicle routing problems, and the traveling salesman problem, among many others There are also a range of optimal and heuristic solution procedures for solving instances of those problems Research is ongoing to expand the types of routing problems that can be solved, and the environments within which they can be applied.
Routing is the act of selecting a course of travel
This process is undertaken by nearly every active
person every day The route from home to school
or work is chosen by commuters The selection
of stops one will make for shopping and other
commercial activities and the paths between
those stops is a routing activity Package delivery
that packages are delivered within specified time windows School buses are assigned routes that will pick up and deliver children in an efficient manner Less tangible objects such as telephone calls or data packets are routed across informa-tion networks Routing is the most fundamental logistical operation for virtually all transportation and communications applications
As in the examples above, routing is frequently
Trang 24
Routing
Its importance to geoinformatics, however, lies
in the nature of routing as a general problem
Transportation, communications, or utility
sys-tems can all be modeled as networks—connected
sets of edges and vertices—and the properties of
networks can be examined in the context of the
mathematical discipline of graph theory Routing
procedures can be performed on any network
dataset, regardless of the intended application
This chapter will discuss the formulation of
rout-ing problems includrout-ing shortest path problems,
and will review in detail general vehicle routing
problems and the traveling salesman problem
Solution procedures for routing problems are
discussed and future trends in routing research
are outlined
bAckground
Generally, a routing procedure is based on an
objective—or goal—for the route, and a set of
constraints regarding the route’s properties By
far the most common objective for routing
prob-lems is to minimize cost Cost can be measured
in many different ways, but is frequently defined
as some function of distance, time, or difficulty
in traversing the network Thus the problem of
locating the least cost or shortest path between
two points across a network is the most common
routing problem It is also a problem for which
there are several extremely efficient algorithms
that can determine the optimal solution The most
widely cited algorithm that solves the least cost
path problem on directed graphs with
non-nega-tive weights was developed by Edsgar Dijkstra
(1959), and an even more efficient version of
this algorithm—the two-tree algorithm—exists
(Dantzig, 1960) Alternative algorithms have been
presented that will solve this problem where
nega-tive weights may exist (Bellman, 1958), where
all the shortest paths from each node to every
other node are determined (Dantzig, 1966; Floyd,
1962), and where not only the shortest path but
also the 2nd, 3rd, 4th, or kth shortest path must be found (Evans & Minieka, 1992)
network des Ign prob LeMs
The shortest path problem is just one of a class
of related routing problems that can be described
as network design problems Network design problems require that some combination of the elements of a network (edges and vertices) be cho-sen in order to provide a route (or routes) through the network This group includes the minimal spanning tree problem, the Steiner tree problem, the Traveling Salesman Problem, and the vehicle routing problem, among many others (Magnanti
& Wong, 1984) The modeling of these problems frequently takes the form of integer programming models Such models define an objective and a set
of constraints Solution procedures are applied that require decisions to be made that generate a route that optimizes the objective while respecting the constraints Given the limited space in this forum, the following sections will focus on the modeling
of two significant routing problems in an effort
to demonstrate the characteristics of the general class Vehicle Routing Problems are presented in order to discuss the range of possible objectives for routing problems, and the Traveling Salesman Problem is presented to demonstrate the formula-tion of the objectives and constraints
Vehicle Routing Problems
Vehicle Routing Problems (VRPs) are those that seek to find a route or routes across a network for the delivery of goods or for the provision of transport services From their earliest incarnations VRPs have been formulated as distance or cost minimization problems (Clarke & Wright, 1964; Dantzig & Ramser, 1959) This overwhelming bias persists to this day Nine out of ten research articles regarding route design in the context of transit routing written between 1967 and 1998 and
Trang 25reviewed by Chien and Yang (2000) employed
a total cost minimization objective When the
route is intended as a physical transport route,
the cost objective is nearly always formulated
as a generalized measure of operator costs (List,
1990), user costs (Dubois et al., 1979; Silman et
al., 1974), or both operator and user costs (Ceder,
2001; Chien et al., 2001; Lampkin & Saalmans,
1967; Newell, 1979; Wang & Po, 2001)
The few exceptions include a model that
maxi-mizes consumer surplus (Hasselström, 1981), a
model that seeks to maximize the number of public
transport passengers (van Nes et al., 1988), a model
that seeks equity among users (Bowerman et al.,
1995), a model that seeks to minimize transfers
while encouraging route directness and demand
coverage (Zhao & Gan, 2003), and a model that
seeks to maximize the service provided to the
population with access to the route (Curtin &
Biba, 2006) VRPs for transport services can
be designed to either determine single optimal
routes, or a system of routes (Ceder & Wilson,
1986; Chakroborty & Dwivedi, 2002; List, 1990;
Silman et al., 1974)
A substantial subset of the literature posits
that routing problems are not captured well by
any single optimization objective, but rather
multiple objectives should be considered (Current
& Marsh, 1993) Among the proposed
multi-ob-jective models are those that tradeoff maximal
covering of demand against minimizing cost
(Current & Pirkul, 1994; Current et al., 1984,
1985; Current & Schilling, 1989), those that seek
to both minimize cost and maximize
accessibil-ity in terms of distance traveled (Current et al.,
1987; Current & Schilling, 1994), and those that
tradeoff access with service efficiency (Murray,
2003; Murray & Wu, 2003)
Regardless of the objective that is deemed
ap-propriate for a routing application, the problem
will frequently be posited in the form of a
struc-tured mathematical model In the next section
the Traveling Salesman Problem is presented to
t he t raveling salesman problem
The Traveling Salesman Problem (TSP) is ably the most prominent problem in combinatorial optimization The simple way in which the prob-lem is defined in combination with its notorious difficulty has stimulated many efforts to find an efficient solution procedure The TSP is a classic routing problem in which a hypothetical salesman must find the most efficient sequence of destina-tions in his territory, stopping only once at each, and ending up at the initial starting location The TSP has its origins in the Knight’s Tour problem first formally identified by L Euler and A T Vandermonde in the mid-1700s In the 1800s, the problem was identified as an element of graph theory and was studied by the Irish mathemati-cian, Sir William Rowan Hamilton The problem was named the Hamiltonian cycle problem in his honor (Hoffman A J & Wolfe P., 1985) The first known mention of the TSP under that name appeared in a German manual published
argu-in 1832, and this was followed by four applied appearances of the problem in the late 1800s and early 20th century (Cook, 2001) The mathemati-cian and economist Karl Menger publicized the TSP in the 1920s in Vienna (Applegate D., 1998), then introduced it in the United States at Harvard University as a visiting lecturer, where the prob-lem was discussed with Hassler Whitney who at that time was doing his Ph.D research in graph theory In 1932, the problem was introduced at Princeton University by Whitney, where A W Tucker and Merrill Flood discussed the problem
in the context of Flood’s school-bus routing study
in New Jersey (Schrijver, 2004) Flood went on to popularize the TSP at the RAND Corporation in Santa Monica, California in late 1940s In 1956 Flood mentioned a number of connections of the TSP with the Hamiltonian paths and cycles in graphs (Flood, 1956) Since that time the TSP has been considered one of the classic models in combinatorial optimization, and is used as a test
Trang 26
Routing
There are many mathematical formulations
for the TSP, with a variety of constraints that
enforce the requirements of the problem Since
this is not the appropriate forum for reviewing
all of the potential formulations, the formulation
attributed to Vajda (Vajda, 1961) has been chosen
in order to demonstrate how such a formulation is
specified The following notation is used:
n = the number of cities to be visited;
i and j = indices of cities that can take integer
values from 1 to n
t = the time period, or step in the route between
the cities
x ijt = 1 if the edge of the network from i to j is used
in step t of the route, and 0 otherwise
d ij = the distance or cost from city i to city j
The objective function is to minimize the
sum of all costs (distances) of all of the selected
elements of the tour:
For all cities, there is just one other city which is
being reached from it, at some time, hence
all for
For all cities, there is some other city from which
it is being reached, at some time, hence
= 1 for all j
∑∑
x
When a city is reached at time t, it must be left
at time t + 1, in order to exclude disconnected
subtours that would otherwise meet all of the above constraints These subtour elimination constraints are formulated as:
t.
and j all for
In addition to the above constraints the decision variables are constrained to be integer values in the range of 0 to 1: 0 ≤ xijt ≤ 1
Like any routing problem structured as an integer program, in order to solve the TSP a pro-cedure must be employed that allows decisions
to be made regarding the values of the decision variables The choice of a solution procedure depends in part on the difficulty of the routing problem and the size of the problem instance be-ing solved The following section describes the combinatorial complexity of routing problems and the solution procedures that can be used to solve them
rout Ing prob LeMs
The TSP and most VRPs are considered to be in a class of problems that are highly combinatorially
complex There are, for example, (n – 1)! possible
tours for the TSP Therefore, as the number of
cities to visit, n, grows, the number of possible
tours grows very rapidly So rapidly, in fact, that even small instances of these problems cannot
be solved by enumeration (the inspection of all possible combinations)
If this is the case these integer programming problems may be solved optimally using a ver-sion of the simplex method to generate fractional optimal solutions from the linear programming relaxation of the integer program, followed by a branch and bound search procedure to identify integer optimal solutions A variety of reformu-lation techniques, preprocessing routines, and