Abstract:
Location based solutions for smartphones and other smart hand-held devices
have been signi cantly increased. Geo location is one of the key contexts
which can be easily captured with the current localization or geo positioning
technologies. Most recent geo-localized Points of Interest (POI) aware systems
perform much intelligent decisions and proactive actions by identifying nearby
places and the nature of the surrounding. For achieving that proactiveness,
Location Based Service (LBS) approaches utilize continuous feed of Global
Positioning System (GPS) which consumes more energy, makes a signi cant
battery drain and generates heat resulting in a severe reduction of operation
time.
Objective of this research is to introduce enhanced power utilization mechanisms
for POI aware systems by implementing intelligent location extraction
methods along with Application Programming Interface (API) level optimizations
as well.
In the relevant research literature mobile device power optimization has
been discussed and many solutions have been introduced and those have been
discussed and referred during the research work.
Applicable use cases which can be integrated with power management
mechanisms have been identi ed to address the above mentioned problem as
the rst step. GPS and WiFi based hybrid positioning system has been identi
ed as the main supportive GPS adaptation. Then intelligent GPS sampling
mechanisms and intelligent communication with the location based service
provider have been studied and classi ed based on the state di erentiation of
the applications.
In the implementation phase a prototype called \DealTella" has been created.
Activity recognition has been implemented for intelligent decision making
in location sampling. GPS adaptation using Wi-Fi trace based reversed
location extraction is the most important power utilization adaptation introduced
during the research work.
A considerable percentage of energy saving could be achieved by enabling
all the mechanisms explained under the implementation section along with
enabling intelligent sampling. Proposed implementation has been tested under
three main scenarios while enabling better battery consumption strategies.
Accuracy has been measured against the battery consumption and recommendations
have been provided based on results.
Further as part of the research work, a prototype has been developed just
to prove the concept and it will be enhanced and released as a marketable and
production quality application.
Modern leading operating systems invest more on optimizing battery consumption
natively. Since modern smart applications are heavy process oriented
for providing the best and most context related user experience. Those applications
consume more and more energy for achieving that proactiveness and
to feed the intelligence into applications. Still there exist a lot of research
opportunities in the context and some of the extensions have been proposed
to be carried out in a future phase.
