SEARCH
0-9 A B C D E F G H I J K L M N O P Q R S T U V W X Y Z
Prev | Current Page 478 | Next

Yingshu Li, My T. Thai, and Weili Wu

"Wireless Sensor Networks and Applications"

Consider
a sensor network where each sensor node periodically generates temperature
samples. Viewed across the entire network, the collection of temperature
samples can be represented by a 2D array of temperature values indexed by
location and time. If all these values could be collected in a central location,
then it is clear that one could compute wavelet coe?±cients on the data to
e?±ciently answer spatial-temporal queries described above. However, the design
of the DIMENSIONS system makes the observation that these wavelet
coe?±cients can be computed and stored far more e?±ciently in a distributed
fashion. This capability rests on the two properties: resolution in the spatial
dimension corresponds to di?®erently sized regions of the sensor field, and that
lower resolution wavelet coe?±cients can be computed from higher-resolution
coe?±cients. DIMENSIONS constructs a storage hierarchy using data-centric
storage concepts, and stores successively lower resolution coe?±cients in the
higher levels of the hierarchy. Given a number of levels of resolution d, the
system e?®ectively divides the geographic region occupied by the sensors into
d levels recursively. The 0-level has only one region, that is, the area covered
by the sensor network. The i-level has 2i sub-regions. The d-level has
2d sub-regions. The size of each sub-region in i-level is four times the size of
a sub-region in i + 1-level. To construct the 1-level from the 0-level in the
hierarchical structure, the system first hashes the names of the datasets corresponding
to 0-level to a location in sub-region of 0-level.


Pages:
466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490