This estimate is called the periodogram [7]. The power
spectral density, ?† Pxx(f), of a process of length L is given in (1):
?† Pxx(f) =
| XL(f) |2
fsL
, (1)
where the discrete-time Fourier transform, XL is given in (2):
XL(f) =
L??’1
X
n=0
xL[n]e
??’2??jfn
fs . (2)
8 Experimental Evaluation
In sensor networks, nodes typically have similar system configuration and enduser
applications. Accordingly, the nodes in our experiments were completely
433
10
0
10
1
10
2
10
3
0
0.2
0.4
0.6
0.8
1
Number of Packets Transmitted
Non??’Switching Nodes
Cherita Corbett, John Copeland, and Raheem Beyah
(a) CDF of number of packets transmitted
(b) CDF of duration at hotspot
(c) CDF of time elapsed to 1st rate switch
Fig. 4. Rate switching nodes.
434
10
0
10
1
10
2
10
3
10
4
10
5
0
0.2
0.4
0.6
0.8
1
Number of Packets Transmitted
0 20 40 60 80 100 120
0
0.2
0.4
0.6
0.8
1
Duration at Hotspot(min)
10
??’3
10
??’2
10
??’1
10
0
10
1
10
2
0
0.2
0.4
0.6
0.8
1
Time Elaspe to 1st Rate Switch (min)
Chapter 18 Unauthorized Sensor Node Identification
homogeneous, with the only variable being the RI. We again emulate a sensor
network using an 802.11 WLAN, due to its ease of deployment.
8.1 Experimental Setup
The testbed is comprised of three nodes. A 1GHz Toshiba laptop functions
as the source sensor node that generates wireless tra?±c. A Linksys 802.11b
wireless router serves as a relay node to forward messages from the source
to the sink node.
Pages:
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680