Unbalanced Expander Based Compressive Data Gathering in Clustered Wireless Sensor Networks

Abstract

Conventional compressive sensing-based data gathering (CS-DG) algorithms require a large number of sensors for each compressive sensing measurement, thereby resulting in high energy consumption in clustered wireless sensor networks (WSNs). To solve this problem, we propose a novel energy-efficient CS-DG algorithm, which exploits the better reconstruction accuracy of the adjacency matrix of an unbalanced expander graph. In the proposed CS-DG algorithm, each measurement is the sum of a few sensory data, which are jointly determined by random sampling and random walks. Through theoretical analysis, we prove that the constructed M × N sparse binary sensing matrix is the adjacency matrix of a (k, ε) unbalanced expander graph when M = O (k log N/k) and t = O (N <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">c</sub> /(kq)) for WSNs with Nc clusters, where 0 ≤ q ≤ 1 and N <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">c</sub> > k. Simulation results show our proposed CS-DG has better performance than existing algorithms in terms of reconstruction accuracy and energy consumption. When hybrid energy-efficient distributed clustering algorithm is used, to achieve the same reconstruction accuracy, our proposed CS-DG can save energy by at least 27.8%.