For Full-Text PDF, please login, if you are a member of IEICE,|
or go to Pay Per View on menu list, if you are a nonmember of IEICE.
Realistic Analysis of Energy Efficiency in Multihop Wireless Sensor Networks
IEICE TRANSACTIONS on Information and Systems
Publication Date: 2014/12/01
Online ISSN: 1745-1361
Type of Manuscript: Special Section PAPER (Special Section on Parallel and Distributed Computing and Networking)
Category: Wireless Network
energy efficiency, multihop wireless sensor networks, optimization, realistic analysis, connectivity, nonlinear programming,
Full Text: PDF>>
As one of the most widely investigated studies in wireless sensor networks (WSNs), multihop networking is increasingly developed and applied for achieving energy efficient communications and enhancing transmission reliability. To accurately and realistically analyze the performance metric (energy efficiency), firstly we provide a measurement of the energy dissipation for each state and establish a practical energy consumption model for a WSN. According to the analytical model of connectivity, Gaussian approximation approaches to experimental connection probability are expressed for optimization problem on energy efficiency. Moreover, for integrating experimental results with theories, we propose the methodology in multihop wireless sensor networks to maximize efficiency by nonlinear programming, considering energy consumptions and the total quantity of sensing data to base station. Furthermore, we present evaluations adapting to various wireless sensor networks quantitatively with respect to energy efficiency and network configuration, in view of connectivity, the length of data, maximum number of hops and total number of nodes. As the consequence, the realistic analysis can be used in practical applications, especially on self-organization sensor networks. The analysis also shows correlations between the efficiency and maximum number of hops, that is the multihop systems with several hops can accommodate enough devices in ordinary applications. In this paper, our contribution distinguished from others is that our model and analysis are extended from experiments. Therefore, the results of analysis and proposal can be conveniently applied to actual networks.