Geoscience Reference
In-Depth Information
addresses can be easily tied to their physical locations, we can determine where people and things
are located more easily than at any time in human history. The close coupling of ubicomp with GC
has contributed to an ambient spatially intelligent environment.
16.3.2 c
ouPling
u
BicoMP
with
gc: a
MBient
S
Patial
i
intelligence
and
d
ecentraliSed
S
Patial
c
oMPuting
The emergence of everyware in the metaverse is rapidly creating a new sentient environment with
growing capabilities of ambient spatial intelligence (AmSI) (ambientspatial.net). As stated earlier, it
is no longer a luxury to discuss GC in the context of ubicomp. It has become a necessity to design GC
in the AmSI environment, where GC is closely coupled with the ubicomp environment. Some of the
general principles have been laid out in the decentralised spatial computing (DeSC) framework of
Duckham (2013). The goal of DeSC is to (1) respond efficiently to queries about events, (2) support
better understanding of those events in real time and (3) improve human decision-making based on
information about spatial events.
According to Satoh (2005), ubicomp environments have several unique requirements as follows:
•
Mobility
: Not only entities, for example, physical objects and people, but also computing
devices can be moved from location to location. The location model in ubicomp is required
to be able to represent mobile computing devices and spaces as well as mobile entities.
Furthermore, it needs to be able to model mobile spaces, for example, cars, which may
contain entities and computing devices.
•
Heterogeneity
: A ubicomp environment consists of heterogeneous computing devices, for
example, embedded computers, handheld/wearable computers, sensor networks of various
kinds and public terminals. Location-based and personalised services must be executed
using computing devices whose capabilities can satisfy the requirements of the services
(Gartner and Ortag, 2011). GC is thus required to maintain the capabilities of computing
devices as well as their locations.
•
Availability
: Ubicomp devices may have limited memories and processors, so they cannot
support all the services that they need to provide. Software must be able to be deployed at
computing devices on demand. GC should be able to manage the (re)location of service-
provider software.
•
Absence of centralised databases
: Since ubicomp devices are organised in an ad hoc
and peer-to-peer manner, they cannot always access database servers to maintain loca-
tion models. The model should be available without database servers, enabling computing
devices to be organised without centralised management servers.
Satoh (2007) further developed and implemented a model for location-aware and user-aware services
in ubicomp environments. This model can be dynamically organised and implemented like a tree
based on geographical containment, such as user-room-floor-building, and each node in the tree
can be constructed as an executable software component. The model is unique to existing approaches
because it can be managed by multiple computers in an ad hoc manner and is also capable of pro-
viding a unified view of the locations of not only physical entities and spaces, including users and
objects, but also computing devices and services. A prototype implementation of this model was
developed on a Java-based mobile agent system. Figure 16.5 shows the overall design, which contains
four components: (1) Virtual component (VC) is a digital representation of a physical entity or space
in the physical world; (2) aura component is a virtual or semantic scope surrounding a physical entity
or computing device; (3) proxy component bridges the world model and computing device and main-
tains the subtree of the model or executes services located in the VC; and (4) service component is a
software module that defines application-specific services associated with physical entities or places.
