Distributed Data Gathering in Energy-Efficient Multi-Hop Wireless Sensor Networks

Markus Leinonen

Centre for Wireless Communications (CWC),
University of Oulu,
Finland

Abstract:

Wireless sensor networks (WSNs) consisting of multiple battery-powered sensors have been frequently proposed for different monitoring and measuring purposes under varying operating environments. Major requirements for the WSN design include high energy efficiency, light infrastructure and autonomous, distributed operation of the nodes. In order to meet the requirements, we will propose two distinct approaches.

The first approach tackles the problem via distributed cross-layer transmission optimization for data gathering multi-hop WSNs. Thus, we propose a novel algorithm which jointly optimizes the resource allocation and routing via consensus optimization and the alternating direction method of multipliers (ADMM). Based on the numerical results, the developed algorithm can improve the energy efficiency by significantly reducing the amount of necessary communication overhead as compared to the state of the art methods.

The second approach focuses on the data compression by exploiting spatial and temporal correlation structures of WSN signal ensembles. The signal dependency will be utilized by a joint data acquisition and reconstruction method, compressed sensing (CS), which allows accurate recovery of length-N signal from only M < N linear measurements. Accordingly, we will study the capabilities of various CS methods to reduce the number of transmissions in correlated data gathering multi-hop WSNs. Furthermore, we propose a novel method based on sliding window processing, where the sink can instantaneously reconstruct current WSN data samples by periodically collecting CS measurements from the sensors. The distinct correlation patterns present in each signal dimension will be jointly exploited by Kronecker sparsifying bases. Moreover, the method will use previously decoded estimates to progressively improve the accuracy of signal estimates. Therefore, the method can also control the trade-off between the decoding delay and complexity. By the numerical experiments, the proposed method can recover the data samples from aggregated linear projections with higher reconstruction accuracy, yet with lower decoding delay and complexity, as compared to the state of the art methods.

Biography:

Markus Leinonen (S'11) received his B.Sc. (Tech.) and M.Sc. (Tech.) degree in electrical engineering from University of Oulu, Oulu, Finland in 2010 and 2011, respectively. In 2010, he joined Centre for Wireless Communications (CWC) at University of Oulu, where he is currently working toward Dr.Sc. (Tech.) degree. His main research interests are in distributed transmission optimization via cross-layer design and compressive data gathering methods in energy-efficient multi-hop wireless sensor networks.