??? Threshold Decision Scheme: If Pt is larger than a pre-selected threshold
?µ 2 (0, 1), it is claimed that an event is correctly detected by Si.
161
Ren-Shiou Liu, Lifeng Sang, and Prasun Sinha
??? Optimal Threshold Decision Scheme: If k ?? 0.5N, it is claimed that an
event is correctly detected by Si.
Both the analysis and simulation results [5] show that the optimal threshold
decision is the best policy, which means each node accepts its own reading
if and only if more than half of its neighbors have the same reading. Though
this is an intuitive result, a decision can be made without calculating Pt which
makes the optimal threshold scheme more feasible.
3 Centralized edge determination
Centralized edge determination uses a designated sensor node with su?±cient
power or the sink to determine event boundaries based on the data collected
from the sensor field. Although the cost for the operation of collecting data
from all sensor nodes may be very high, sometimes it is still desirable to do
so, because boundaries can be detected more accurately and e?±ciently with a
???panorama??? of the sensor field of interest. Also, if the total number of sensors
is not too many, and all sensor nodes and the sink can hear each other, then
energy consumption can be kept at a low level as well.
With a knowledge of the values of event predicates and the positions of
all sensor nodes, a number of techniques can be applied in the centralized
edge determination approach.
Pages:
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280