AUGUST 2004 (VOL. 37, NO. 8) P. 4
0018-9162/04/$31.00 © 2004 IEEE
Published by the IEEE Computer Society
Published by the IEEE Computer Society
|MODELING COMPLEX SPOKEN DIALOG, PP. 32-40|
|SENSOR NETWORK APPLICATIONS, PP. 50-61|
|WISENET: AN ULTRALOW-POWER WIRELESS SENSOR NETWORK SOLUTION, PP. 62-70|
|THE FLOCK: MOTE SENSORS SING IN UNDERGRADUATE CURRICULUM, PP. 72-78|
PDFs Require Adobe Acrobat
MODELING COMPLEX SPOKEN DIALOG, PP. 32-40
Hans Dybkjær and Laila Dybkjær
Commercial spoken dialog systems, which handle increasingly advanced tasks, must deal with three development challenges: how to model the dialog when the user can select from many tasks, how to communicate with customers about SDS design, and how to develop code for a large SDS when the design will likely require updates throughout development.
To develop a commercial phone-based SDS that supplies information to Danish employees about their holiday allowance, the authors had to meet all these challenges. Key to this system was their use of the Conceptual Dialog Language, which expresses patterns specific to the dialog model while still providing a clear picture of it to domain experts and supporting model updates.
CDL captures the specifics of an SDS application while remaining flexible enough to support a range of modeling styles and requirements. To facilitate effective communication with domain experts, modelers can translate the dialog model in CDL to an HTML document, or they can compile it to a programming language for executable code.
SENSOR NETWORK APPLICATIONS, PP. 50-61
Kirk Martinez, Jane K. Hart, and Royan Ong
Sean M. Brennan, Angela M. Mielke, David C. Torney, and Arthur B. Maccabe
Miklos Maroti, Gyula Simon, Akos Ledeczi, and Janos Sztipanovits
Sensor networks can be used to monitor our environment, objects in that environment, and the interactions of objects with each other and their encompassing environment. Examples include environmental and habitat monitoring, structural monitoring and condition-based equipment maintenance, and disaster management and emergency response. Researchers working on three diverse projects have developed other novel applications for sensor network technology.
In "Environmental Sensor Networks," Kirk Martinez and his coauthors from the University of Southampton describe their GlacsWeb project, which involves ongoing research in subglacial bed deformation. They also discuss the challenges encountered in extracting data gathered by sensor nodes deployed in remote locations.
In "Radiation Detection with Distributed Sensor Networks," Sean M. Brennan and coauthors discuss a project being developed at Los Alamos National Laboratory in cooperation with the University of New Mexico to provide a distributed sensor network for detecting vehicles transporting radioactive isotopes that could potentially be detonated over a densely populated area.
Finally, in "Shooter Localization in Urban Terrain," Akos Ledeczi and coauthors describe PinPtr, a prototype system that provides a novel approach for detecting and locating a sniper in a challenging environment such as complex urban terrain.
These three projects demonstrate the promise of sensor network technology for monitoring our environment and our safety in that environment.
WISENET: AN ULTRALOW-POWER WIRELESS SENSOR NETWORK SOLUTION, PP. 62-70
Christian C. Enz, Amre El-Hoiydi, Jean-Dominique Decotignie, and Vincent Peiris
Since wireless sensor networks often are deployed in regions that are difficult to access, the nodes should not require maintenance. They must be energetically autonomous, using batteries that do not need to be replaced or recharged. In many application scenarios, the targeted node lifetime typically ranges from two to five years, imposing drastic constraints on power consumption. With a single 1.5-V AA alkaline battery, the average power consumption ranges from 100 to 10 microwatts, for a node lifetime ranging between two and seven years.
To conserve power, the nodes must sleep most of the time. Thus, the Swiss Center for Electronics and Microtechnology developed WiseNET to optimize power consumption. An ultralow-power platform for the implementation of wireless sensor networks, WiseNET achieves low-power operation through a careful codesign approach. The system combines a complex system-on-chip sensor node with a dedicated duty-cycled radio and WiseMAC, a low-power MAC protocol designed for low-duty-cycle wireless sensor networks. The WiseNET solution consumes about 100 times less power than comparable alternatives available today.
THE FLOCK: MOTE SENSORS SING IN UNDERGRADUATE CURRICULUM, PP. 72-78
Bruce Hemingway, Waylon Brunette, Tom Anderl, and Gaetano Borriello
Educational excellence requires exposing students to the current edge of research. To ensure that student projects follow the same trajectory that industry is traveling, educators must continually introduce emerging techniques, practices, and applications into the curriculum.
While many graduate-level classes focus on sensor networks, they do not serve as a good template for the undergraduate curriculum because they assume a much greater breadth of knowledge on the students' part as well as greater maturity to absorb new topics on their own.
The "Flock of Birds" project in the Department of Computer Science and Engineering at the University of Washington integrates the theory and practice of wireless sensor networks into the mainstream curriculum early enough to form a basis for all students' understanding of embedded computing—not just a short-lived application exercise for some of their capstone design projects.