Database Reference
In-Depth Information
In addition, as computer and scientific technologies have improved recently, small
size handheld mobile devices such as PDAs, mobile phones, and Auto PCs have been
used in various fields of mobile computing and Telexistence technologies more and
more. The need to utilize a variety of service applications such as car navigation,
MP3/WAV player, car maintenance program, and information center solution con-
necting to server, on these devices is increasing. However, an embedded hardware
system has limited resources that are not enough to handle the large amounts of data,
and analyze them. Thus, an embedded technique to resolve this problem is required.
To cope with this problem, we propose a decision theoretic fusion framework that
includes the multiple level-of-abstraction approach which combines multiple-level
association rules and a summary table as well as active interaction rule generation
algorithm for actionability in an embedded car navigation system. In addition, it
includes the sensory and data fusion level rule extraction algorithm to cope with si-
multaneous events occurring from multi-modal interfacing. This embedded system is
connected to the data mining server based on the web in order to extract and access
the rules and data. This is because the Web not only contains a huge amount of in-
formation, but also can provide a powerful infrastructure for communication and
information sharing [6][8]. With this data mining server, the proposed system can
provide an efficient data representative service as well as actionability to present
interactive methods without processing the raw data.
The proposed system is applied to command, control, communication, and intelli-
gent car navigation systems. This provides an efficient speech interactive agent (SIA)
rendering smooth car navigation by employing a conversational tool; embedded
automatic speech recognition, embedded text-to-speech, and distributed speech rec-
ognition modules, all the while enabling safe driving. The embedded car navigation
system is extended to provide a user-friendly service and interactive capability by
using the conversational tools. The system can reveal the status of the system and its
scheduled jobs by actively using the active interaction rule generation algorithm. This
is due to the fact that the driver has an access pattern about specific applications that
are frequently used. In addition, the information about traffic, weather, news, daily
schedules, and car management can provide valuable information to the driver as well
as decision-making advice on what action the proposed system should take. Using
such information, the speech interactive agent provides efficient interactive methods
to operate for the required events.
First, this system uses sensory fusion rules in order to combine multiple events
simultaneously occurring from multi-modal sensors such as push-to-talk, remote
controller, touch screen, mute, hands-free, external buttons, and application events
received from multimedia service applications in the embedded client system. Sec-
ond, the data fusion framework is provided by using the features extracted from
sensory fusion rules. At this time, user access patterns occurring by user driven
events operate a specific service application, and are mined and stored in databases on
an embedded system for certain periods. This feature provides the means to decide a
specific action. The proposed system can connect the Internet server using a CDMA
200 terminal to represent large amounts of information. However, an embedded sys-
tem has a small sized memory that has not enough space to store a lot of information.
To resolve this problem, the multiple level-of-abstraction approach for the multiple-
level association rules is applied.
Search WWH ::
Custom Search