Cerpa et al. [46] suggested collaborating
those multiple sensor nodes surrounding the object so that the data
provided could be more complete, reliable and accurate. Zhang et al. in [47]
proposed a dynamic convoy tree-based collaboration (DCTC) framework for
object tracking. Sensor nodes surrounding the moving object will construct
an initial convoy tree. The root will be the final recipient of reports from each
node and will process them using certain algorithms [48], [49] to generate a
consolidated sensing report that will be saved locally, waiting for query, or
sent to the sinks. An improvement on selecting a new root to replace the old
one as an object moves far away is introduced in [50].
4.2 Quorum-Based Method
A. Method Description
To further improve the performance of write and read operation, we can
also apply a quorum-based method to track moving objects. The basic idea
is that when any sensor node has detected e.g., a new tank, it updates this
event locally, and advertises this information by propagating an ADV in both
north and south directions until reaching the north and south boundaries of
the network, to form a write quorum; and when any soldier needs to access
some information (this node is called a querying node), it sends out certain
REQ in both east and west directions. Nodes on this east-west direction are
collected as a read quorum. The REQ is bound to meet corresponding ADV
at a rendezvous node if the DATA was advertised before.
Pages:
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410