الفهرس 
Chapter 1Only 14 pages are availabe for public view
 |  | from | 16 |  |  |
| | | from | 16 | | |
Abstract
With the advances in inexpensive sensor technology and wireless communica-
tions, development of large-scale Wireless Sensor Networks(WSNs) have become cost-
e ective and their viability has attracted attention from a wide range of civilian,
natural and military applications. Energy consumption and prolonging network life-
time are a primary challenges in many studies on WSNs. Thus, it is necessary that
energy-ecient protocols are designed to maximize the network lifetime. To ad-
dress these challenges, we propose an ecient technique called Ecient Compressive
Sensing based Technique(ECST). ECST utilizes compressive sensing (CS) theory as
in-network compression technique to maximize the network lifetime and improve the
performance of WSNs.We provide performance metrics to analyze the performance of
ECST approach and show by simulation results that ECST technique gives good per-
formance in terms of reducing the energy consumption and maximizing the network
lifetime. On the other hand, to reconstruct the original data from the CS compressed
data that received by the base station and also in order to improve the reconstruction
performance, we present reconstruction greedy algorithm called Adaptive Iterative
Forward-Backward Greedy Algorithm (AFB). AFB belongs to the general category
of Two Stages-type algorithms where it consists of consecutive forward and back-
ward stages. During the forward stage, AFB depends on solving the least square
problem to select columns from the measurement matrix. Furthermore, AFB uses a
simple backtracking step to detect the previous chosen columns accurately and then
removes the false columns at each time. The reconstruction quality of AFB algo-
rithm is demonstrated by both computer-generated signals and real data gathered by
a WSN located in the Intel Berkeley Research lab. The simulation results show that
AFB outperforms Forward-Backward Pursuit, Subspace Pursuit, Orthogonal Match-
ing Pursuit(OMP)and Regularized OMP in terms of reducing reconstruction error.