Parameter Estimation in Wireless Sensor Networks with Normally Distributed Sensor Gains
Zhenxing Luo, Department of Electrical and Computer Engineering, The University of Alabama at Birmingham, Birmingham, AL, USA.
Manuscript received on January 01, 2013. | Revised Manuscript received on January 02, 2013. | Manuscript published on January 05, 2013. | PP: 46-48 | Volume-2, Issue-6, January 2013. | Retrieval Number: F1087112612/2013©BEIESP
Open Access | Ethics and Policies | Cite
© The Authors. Published By: Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Wireless sensor networks (WSN) have attracted significant attention recently. The distributed estimation problem is an important research topic in WSNs. In the distributed estimation problem, the fusion center estimates an unknown parameter based on information gathered from sensors. Usually, it is assumed that sensors have identical gains. However, this may not be true due to manufacture errors or environmental influence. In this paper, we assume sensor gains follow normal distribution and present a maximum likelihood estimation (MLE) approach for distributed estimation in WSNs with normally distributed sensor gains. Moreover, the Cramer-Rao lower bound (CRLB) corresponding to this MLE approach is also derived. Simulation results showed that the root square mean (RMS) estimation errors given by this MLE approach were close to the CRLB if the variance of the sensor gains is small. If the variance of the sensor gains was large, the RMS estimation errors were not close to the CRLB.
Keywords: Distributed estimation, maximum likelihood estimation, Gaussian distribution, wireless sensor networks.