Dynamic Channel Selection
Algorithm for Cognitive Radios
Circuits and Systems for Communications, 2008. ICCSC 2008. 4th IEEE International Conference on
26-28 May 2008 Page(s):275 - 278
Digital Object Identifier 10.1109/ICCSC.2008.64
Advisor: Wen-Hsing Kuo
Presenter: Che-Wei Chang
Network Model
GV , E
V : is the set of the cognitive users
N V : and E is the set of links between the active cognitive users
Su
The set of neighbors of active cognitive user u, u V
d u
The degree of cognitive user u .
D max d u (1)
uV
2
Dynamic Spectrum Sharing Algorithm(1/2)
For i u (here u 1 ), Pi ,u v is expressed by
Pi ,uv 1 p
Sv
(2)
While for i u , Pi ,u v is given by
Pi ,uv p1 p
Sv
(3)
Then the average probability of success over a frame for transmission u v
is given by 1 N
Pu v
N
P
i 1 i ,u v
1 N 1 p
1 p Sv (4)
N
3
Dynamic Spectrum Sharing Algorithm(2/2)
The system throughput for each available channel is written as
1 N 1 p
P
1
1 p Sv (5)
N uV N
~
For analysis more tractable, the system throughput P is approximated by P
1 N 1 p
V N 1 p avg
~ 1 (6)
P
S
N u
S avg denotes the average degree in the CR network, which is given by
Savg 1 / N uV Su (7)
4
Simulation Results(1/4)
Number of Average Number of Average
Number of Number of
active CR degree in CR p active CR degree in CR p
channels channels
users network users network
1 5 2 0.16 1 5 3 0.06
1 10 2 0.26 1 10 3 0.16
1 15 2 0.28 1 15 3 0.2
1 20 2 0.3 1 20 3 0.22
1 25 2 0.3 1 25 3 0.22
1 30 2 0.32 1 30 3 0.22
Table 1. Table 3. Table 2.
Degree = 2 Number of Average
Number of
active CR degree in CR p
channels
users network
1 5 5 0
1 10 5 0.08
1 15 5 0.1
Degree = 3
1 20 5 0.12
1 25 5 0.14
1 30 5 0.14 5
Simulation Results(2/4)
Fig.2. System throughput comparison versus the number of active CR users for K=1
6
Simulation Results(3/4)
Fig.3. The system throughput in terms of the number of active CR users for K=2
7
Simulation Results(4/4)
Fig.4. The system throughput versus the number of available channels for N=5
8
Thanks for your attention!