(March 2009)
Mohammed F.A. Ahmed and Sergiy A. Vorobyov
License and Referencing
This code package is licensed
under the GPLv2 license. If you in any way use this code for research that
results in publications, please cite our original article M.F.A. Ahmed and S.A. Vorobyov,
"Sidelobe
control in collaborative beamforming via node selection," IEEE
Trans. Signal Processing, vol. 58, no. 12, pp. 6168–6180, Dec. 2010.
CBsidelobeControl is a Matlab implementation of the examples described in the paper M.F.A. Ahmed and S.A. Vorobyov,
"Sidelobe
control in collaborative beamforming via node selection," IEEE
Trans. Signal Processing, vol. 58, no. 12, pp. 6168–6180, Dec. 2010.
Fig 3: Beampattern: The intended BS/AP is located at
\varphi_0= 0^o and 4 unintended BSs/APs at directions \varphi_1=-140^o,
\varphi_2=-70^o, \varphi_3=70^o, and
\varphi_4=140^o: M = 512, N = 256, L = 32, \varphi_0=0^o, and \eta_thr = 10~dB.
Fig 4: Beampattern: Multi-link beampatterns with BSs/APs at directions
\varphi_0 = 0^o, \varphi_1 = -140^o, \varphi_2 =
-70^o, \varphi_3 = 70^o, and \varphi_4 = 140^o: M = 512, N = 256, L = 32, and \eta_thr = 10~dB.
Fig 5: Beampattern: The interference is limited in the range \phi \in [ 25^o \, 45^o ]: M = 512, N = 256, L = 32, \varphi_0=0^o,
and \eta_thr = 10~dB.
Fig 6: Beampattern: The unintended BSs/APs are at directions corresponding to
the peaks of the average beampattern: M = 512, N = 256, L = 32, \varphi_0=0^o,
and \eta_{\text{thr}} = 10~dB.
Fig 7: Average number of trials E{T} versus threshold
\eta_thr: M = 512, N = 256, \varphi_0=0^o, and
\varphi_1=65^o.
Fig 8: Average number of trials E{T} versus threshold
\eta_thr for different values of D: M = 512, N = 256,
L = 32, and \varphi_0=0^o.
Fig 9: The CCDF of the INR for different values of the threshold \eta_thr: M = 512, N = 256, L = 32, \varphi_0=0^o, and
\varphi_1=65^o.
Fig 10: The CCDF of the INR for different values of K: M = 512, N = 256, L =
32, \varphi_0=0^o, and \eta_thr = 10~dB.
Fig 11: The SNR of the single-link CB and the SINR of the multi-link CBs with
and without node selection for different values of K: M = 512, N = 256, L = 32,
and \varphi_0=0^o.
Fig 12: The transmission rate of the single- and multi-link CBs with and
without node selection for different values of K: M = 512, N = 256, L = 32, and
\varphi_0=0^o.
To simulate the examples and generate the figures of
the paper use the following m-files:
Fig3andFig4.m
Fig5.m
Fig6.m
Fig7.m
Fig8.m
Fig9.m
Fig10.m
Fig11andFig12.m
In examples 3,4, 5, and 6,
- The channel effect is neglected to focus on the beampattern.
- The plotted beampattern is normalized to the noise
power.
SN : Data structure for a wireless sensor
network (WSN).
WSN.R : Disk radius normalized to \lambda.
WSN.x :
x coordinates of the sensor node locations.
WSN.y :
y coordinates of the sensor node locations.
WSN.M : The total number of sensor nodes in the
coverage area of the transmitting source node.
WSN.N : The desired number of collaborative nodes to
be selected.
WSN.L : The size of a group of candidate sensor nodes.
WSN.APsDirections : Directions of the BSs/APs.
WSN.K : Number of the BSs/APs
WSN.TargetedAP : Index of the intended BS/AP in WSN.APsDirections
WSN.UnTargetedAP:Index of the unintended BS/AP in WSN.APsDirections
WSN.Index : Index of the collaborative nodes.
WSN.phi :
The angle at which the beampattern is calculated.
WSN.ChannelGain : The channel coeffiecents.
WSN.SNR: The power budget for all N collaborative nodes normalized to the noise
power \sigma_w^2.
WSN.INRthr: The INR threshold \eta_thr.
WSN.NoisePower: Noise power.
WSN.Wmax :
The transmitting of the collaborative nodes.
WSN.v: The variance of the Gaussian distribution
corresponding log-normal disdtibuted
channel coefficient.
WSN.m: The mean of the Gaussian distribution
corresponding log-normal disdtibuted
channel coefficient.
WSN.ma: The mean of the channel coefficient.
WSN.va: The variance of the channel coefficient.
WSN.mx: The mean of the array factor.
WSN.vx: The variance of the array factor.
WSN.NoOfTrials: The number of trials required for
node selection.
INR_sim: Simulation value of the INR.
CCDF: Analytical value of the CCDF.
CCDF_sim: Simulation value of the CCDF.
BP: The beampattern.
- GeneralWSN: Creats the data structure WSN and sets its initial data
values.
- UniformWSN: Generats a sensor nodes coordinates in 2D plan. The sensor nodes are
scattered over a disk of radius R. The coordinates WSN.x
and WSN.y follow a Uniform spatial distribution.
Input(s):
WSN.M
WSN.R
Output(s):
WSN.x
WSN.y
- BeamPattern: Finds the beampattern of the
collaborative sensor nodes.
Input(s):
WSN.Wmax
WSN.x
WSN.y
WSN.ChannelGain
WSN.Index
WSN.phi
WSN.APsDirections
Output(s):
BP
- BeamPatternUsingEquation: Finds the average
beampattern of the uniform distributed sensor nodes using equation (15) from
the paper H. Ochiai, P. Mitran,
H. V. Poor, and V. Tarokh, "Collaborative
beam-forming for distributed wireless ad hoc sensor networks," IEEE Trans.
Signal Processing, vol. 53, no. 11, pp. 4110-4124, Nov. 2005. Note that the
beampattern in this equation is normalized to 1.
Input(s):
WSN.phi
Output(s):
BPequation
- MyCCDF: Generate the CCDF.
- NodeSelection: Selects a subset of WSN.N
collaborative nodes from the WSN.M nodes in the coverage area of each source
node. The size of one group of nodes to be tested in
each trial is WSN.L.
Input(s):
WSN.SNR
WSN.INRthr
WSN.NoisePower
WSN.M
WSN.N
WSN.L
WSN.K
WSN.phi
WSN.APsDirections
WSN.TargetedAP
WSN.UnTargetedAP
WSN.ChannelGain
Output(s):
WSN.Index
WSN.NoOfTrials
Please report any bugs to Mohammed Ahmed
<mfahmed@ualberta.ca> or Sergiy A. Vorobyov <svor@ieee.org>.
SensSel.rar
Last updated: September 20, 2010.