Given a tra?±c capture T seconds long, we will have N=T/s samples.
The maximum frequency that can be represented is 1/2s hertz.
431
Cherita Corbett, John Copeland, and Raheem Beyah
(a) Number of rate switches per node
(b) CDF of rate switches for all clients
(c) CDF of rate switches excluding non-switching nodes that
sent less than eight packets
Fig. 2. Rate switching nodes.
432
0 10 20 30 40 50 60
10
0
10
1
10
2
10
3
10
4
Node #
Number of Rate Switches
All Nodes
10
0
10
1
10
2
10
3
10
4
0
0.2
0.4
0.6
0.8
1
Number of Rate Switches
All Nodes
10
0
10
1
10
2
10
3
10
4
0
0.2
0.4
0.6
0.8
1
Number of Rate Switches
Nodes Transmitting More Than 8 Packets
Chapter 18 Unauthorized Sensor Node Identification
Fig. 3. CDF of number of packets transmitted by non-switching nodes.
7.2 Signal Processing
Signal processing has been previously applied to network research to detect
and classify Denial of Service attacks [9], to map the flow of wireless tra?±c, and
to extract information about protocol behavior in encrypted wireless tra?±c
[8]. We apply signal processing to wireless streams to analyze the timing of
packet arrival, particularly while the rate switching algorithm is invoked to
distinguish between RIs with presumably di?®erent algorithms.
Given a signal x(t) constructed using the process discussed above, we
estimate the power spectral density (PSD) by computing the discrete-time
Fourier transform of the samples of the process and taking the magnitude
squared of the result.
Pages:
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680