Detection And Opportunistic Spectrum Access In Sensor Networks
Abstract
This thesis examines target detection problems in Radar Sensor
Networks (RSN) and opportunistic spectrum access problem in
Cognitive Sensor Networks (CSN). First, studies on the Space-Time
Adaptive Processing (STAP) and radar waveform design are provided.
Investigation into the target detection performance gain of RSN when
STAP and radar waveform design are combined in RSN is then
performed. Studies in this thesis show that detection performance of
RSN using our proposal is superior to that of a single radar system
using STAP only. To further studies on target detection, the
multi-target detection problem in RSN is also examined. Signal,
interference, and noise at radar sensors are modeled and analyzed.
At the clusterhead of RSN, a Maximum Likelihood Multi-Target
Detection algorithm is proposed to estimate the possible number of
targets in a surveillance area. Achieved results show that detection
performance of RSN is much better than that of a single radar system
in terms of the miss-detection probability and the root mean square
error.
Besides detection in RSN, this thesis studies an opportunistic
spectrum access problem and proposes a spectrum access scheme in
CSN. The spectrum access scheme is built using Fuzzy Logic System
(FLS); and spectrum access decision is based on: (1) spectrum
utilization efficiency of the secondary user (SU); (2) its degree of
mobility; and (3) its average distance to primary users (PU). The
output of the FLS provides the probabilities of accessing spectrum
band for SUs and the SU with the highest probability will be
assigned the available spectrum. Studies also show that our scheme
performs much better than random access approach.