On Enhancing Network Dynamic Adaptability for Compressive Sensing in WSNs

Abstract

Both energy efficiency and network dynamic adaptability are crucial research issues in wireless sensor networks (WSNs). Compressive sensing has been widely adopted in WSNs for efficient data gathering. Random walk is a typical dynamic routing mechanism. Integrating compressive sensing with random walk offers the opportunity to achieve both efficiency and dynamic adaptability. However, there are still many problems in the existing schemes, such as semi-dynamic routing (which contains both dynamic and static routing), non-uniform sampling and relying on global coordinate information. In order to address these problems, we propose a dual random walk based compressive sensing. First, we design a dual random walk, which is a directed random walk but does not depend on the coordinate information of each node, and it also achieves a uniform sampling. Then, we construct a dynamic compressive sensing data gathering scheme based on the dual random walk, which effectively enhances the network dynamic adaptability. Both theoretical analysis and experimental evaluation are conducted. The experimental evaluation results demonstrate that it outperforms the most closely related work.