Improved Maximum Likelihood Estimation of Target Position in Wireless Sensor Networks using Particle Swarm Optimization

Estimation of target position from multi-frame binary data provided by a wireless sensor network (WSN) can be done by optimizing a complex multimodal likelihood function. Deterministic quasi Newton-Raphson (QNR) schemes with line search are typically used for optimization in maximum likelihood estim...

Full description

Saved in:
Bibliographic Details
Published inThird International Conference on Information Technology: New Generations (ITNG'06) pp. 274 - 279
Main Authors Noel, M.M., Joshi, P.P., Jannett, T.C.
Format Conference Proceeding
LanguageEnglish
Published IEEE 2006
Subjects
Bandwidth
Estimation error
Intelligent networks
Maximum likelihood estimation
Multimodal sensors
Optimization methods
Particle swarm optimization
Sensor phenomena and characterization
Stochastic processes
Wireless sensor networks
Online AccessGet full text

Cover

More Information
Summary:Estimation of target position from multi-frame binary data provided by a wireless sensor network (WSN) can be done by optimizing a complex multimodal likelihood function. Deterministic quasi Newton-Raphson (QNR) schemes with line search are typically used for optimization in maximum likelihood estimation. However, these methods often find a local minimum, which leads to large estimation errors. This paper presents an approach that employs particle swarm optimization (PSO) techniques for global optimization of the likelihood function. Simulation results comparing the performance of a maximum likelihood target position estimation scheme employing QNR and PSO algorithms are presented. It is seen that the PSO algorithm provides significantly higher position estimation accuracy throughout the sensor field
ISBN:0769524974
9780769524979
DOI:10.1109/ITNG.2006.72