2. Performance of ACQUIRE.
392
0 2 4 6 8 10 12 14 16 18 20
0
5
10
15
20
25
30
35
40
Look??’ahead Parameter (d)
Expected Search Cost
c = 0.0003
c = 0.0017
c = 0.0057
c= 0.0117
10
??’3
10
??’2
10
??’1
0
2
4
6
8
10
12
14
Update to Query Ratio (c)
Look??’ahead Parameter (d)
Chapter 16 Modeling Data Gathering in Wireless Sensor Networks
of a single source. On the other extreme, when ?± = 1, there is no correlation
at all, with a joint entropy that is equal to the sum of the entropies of each
source.
To illustrate the tradeo?®s involved in joint routing and compression, we
consider a simple network scenario illustrated in Figure 3. The n sources are
equally spaced, each having D additional intermediate nodes between them
and the sink. The way the information is routed from each source to the sink
is as follows.
First the set of sources is divided into clusters of s nodes. Within each
cluster, the data is routed sequentially from node to node, with compression
at each successive step. Thus the H1 bits of data from the first sensor move
to the second sensor, where they are compressed jointly with the data at that
node (we assume an idealized entropy coding that achieves the maximum
compression possible); then H2 bits are transmitted from the second sensor
to the third, and so on till the last node within the cluster. Then the jointly
compressed Hs bits of data from each cluster are routed to the sink along the
shortest path.
Pages:
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624