I. INTRODUCTION AND MOTIVATION
Mobile Augmented Reality (MAR) enhances the way users
perceive and interact with the world, by making it a part of the
user interface, so that accessing and understanding locationbased
information and services becomes easier. Geospatial
MAR could revolutionize the way we discover Points Of
Interest (POIs) and experience our surroundings, but this
vision is still far from fulfilled, because the information
contents supporting location-based resource discovery and
route planning are usually shallow. Most solutions model POIs
just with name and category labels, so supporting only exact
discovery. A possible retrieval cannot exploit more detailed
POI descriptions to filter out and rank resources according
to relevance. The problem is exacerbated when dealing with
accessibility information, in order to provide personal navigation
to users with disabilities. Typical cartographic corpuses
–both commercial and open ones like OpenStreetMap (OSM,
http://www.openstreetmap.org/)– lack systematic information
about accessibility Points Of Attention (POAs) [1] such as
entrances, stairs/elevators, sidewalks, curbs, obstacles and road
signs (with the exception of one-way signals), which are
fundamental for barrier-free route planning for people with
mobility impairments.
Semantic-based technologies can allow more articulated
descriptions of locations, POAs and POIs. The use of metadata
(annotations) endowed with formal machine-understandable
meaning enables advanced location-based resource discovery
through proper inferences. This paper proposes a smartphone
framework and prototypical system integrating a semanticbased
MAR discovery tool with a novel indoor/outdoor navigation
solution for people with motion disabilities, such
as wheelchair users. The proposal leverages Semantic Web
technologies and crowd-sourced OSM cartography to tag POIs
with semantic annotations and POAs with accessibility information.
Starting from an annotated user profile, the system
executes a matchmaking with semantic-enhanced OSM POIs
in a reference region around users location. Outcomes are
displayed as color-coded markers on the display adopted as
device camera viewfinder; they correspond to the real direction
and distance of each POI from the user. By touching a marker,
the user can see a logic-based explanation for the result, in
terms of missing and/or conflicting characteristics between
her profile and the POI. After selecting a POI as destination,
the routing engine computes the best path from the user’s
current position. The open source OsmAnd (http://osmand.net/)
navigation tool for the Android platform on locally stored
maps was enhanced by taking into account accessibility features
as well as multi-floor path planning within buildings.
By combining trilateration of Wi-Fi access points with the
IndoorAtlas (https://www.indooratlas.com/) technology based
on Earth Magnetic Field (EMF), real-time indoor positioning
is attained automatically, without user actions or environment
modifications.
The remainder of the paper is structured as follows. Section
II discusses relevant related work, while Section III describes
in detail the overall discovery and navigation framework. In
order to clarify the benefits of the proposal, Section IV reports
illustrative examples from a case study in the campus of
Techincal University of Bari in Italy. The paper is closed by
Section V.
I. INTRODUCTION AND MOTIVATIONMobile Augmented Reality (MAR) enhances the way usersperceive and interact with the world, by making it a part of theuser interface, so that accessing and understanding locationbasedinformation and services becomes easier. GeospatialMAR could revolutionize the way we discover Points OfInterest (POIs) and experience our surroundings, but thisvision is still far from fulfilled, because the informationcontents supporting location-based resource discovery androute planning are usually shallow. Most solutions model POIsjust with name and category labels, so supporting only exactdiscovery. A possible retrieval cannot exploit more detailedPOI descriptions to filter out and rank resources accordingto relevance. The problem is exacerbated when dealing withaccessibility information, in order to provide personal navigationto users with disabilities. Typical cartographic corpuses–both commercial and open ones like OpenStreetMap (OSM,http://www.openstreetmap.org/)– lack systematic informationabout accessibility Points Of Attention (POAs) [1] such asentrances, stairs/elevators, sidewalks, curbs, obstacles and roadsigns (with the exception of one-way signals), which arefundamental for barrier-free route planning for people withmobility impairments.Semantic-based technologies can allow more articulateddescriptions of locations, POAs and POIs. The use of metadata(annotations) endowed with formal machine-understandablemeaning enables advanced location-based resource discoverythrough proper inferences. This paper proposes a smartphoneframework and prototypical system integrating a semanticbasedMAR discovery tool with a novel indoor/outdoor navigationsolution for people with motion disabilities, suchas wheelchair users. The proposal leverages Semantic Webtechnologies and crowd-sourced OSM cartography to tag POIswith semantic annotations and POAs with accessibility information.Starting from an annotated user profile, the systemexecutes a matchmaking with semantic-enhanced OSM POIsin a reference region around users location. Outcomes aredisplayed as color-coded markers on the display adopted asdevice camera viewfinder; they correspond to the real directionand distance of each POI from the user. By touching a marker,the user can see a logic-based explanation for the result, interms of missing and/or conflicting characteristics betweenher profile and the POI. After selecting a POI as destination,the routing engine computes the best path from the user’scurrent position. The open source OsmAnd (http://osmand.net/)navigation tool for the Android platform on locally storedmaps was enhanced by taking into account accessibility featuresas well as multi-floor path planning within buildings.By combining trilateration of Wi-Fi access points with theIndoorAtlas (https://www.indooratlas.com/) technology basedon Earth Magnetic Field (EMF), real-time indoor positioningis attained automatically, without user actions or environmentmodifications.The remainder of the paper is structured as follows. SectionII discusses relevant related work, while Section III describesin detail the overall discovery and navigation framework. Inorder to clarify the benefits of the proposal, Section IV reportsillustrative examples from a case study in the campus ofTechincal University of Bari in Italy. The paper is closed bySection V.
การแปล กรุณารอสักครู่..
