156
Chapter 6 Boundary Detection for Sensor Networks
2.4 Random bisection approach
The localized edge detection schemes we have discussed so far are based on
the 0??“1 binary reading (the event predicate) of each sensor. However, this
decision making mechanism may have the following drawbacks:
??? 0/1 event predicates do not tell any spatial information on deployed sensors.
??? 0/1 event predicates are the preprocessed results of the real readings. An
edge detection approach based on a second round approximation may not
be accurate.
??? 0/1 event predicates may o?®er misleading information due to faulty sensors.
Thus, instead of using the binary reading, the random bisection approach
uses the actual reading of an interested factor or phenomenon, such as temperature,
light, sound, chemical densities, and so on. Furthermore, the random
bisection approach divides a sensor??™s neighborhood equally into two sectors
and extends a faulty sensor detection algorithm to identify edge sensors [2].
Faulty sensor detection
Sensors located in the same region are spatially correlated. A sensor??™s reading
without the support of its neighbors would indicate a faulty sensor. To determine
whether a sensor Si is faulty, its reading is compared with those of its
neighbors. Let N(Si) denote the set of sensors including Si and its k neighbors
Si1, Si2, ..., Sik within the range of a radius R that is centered at Si. Assume
x(i)
1 , x(i)
2 , .
Pages:
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273