Low complexity improvement on linear least-squares localization

In this paper, a low complexity way to improve the Linear Least-Squares (LLS) method is introduced. The n-dimensional (n-D) positioning problem is first reduced to 1-D and then solved iteratively. Compared to the classic Gauss-Newton method, the n×n matrix inversion/factorization in each iteration i...

Full description

Saved in:
Bibliographic Details
Published in2008 11th IEEE Singapore International Conference on Communication Systems pp. 124 - 128
Main Authors Junlin Yan, Tiberius, C., Bellusci, G., Janssen, G.
Format Conference Proceeding
LanguageEnglish
Published IEEE 01.11.2008
Subjects
Computational complexity
Computational modeling
Difference equations
Geodesy
Least squares methods
Logistics
Mobile communication
Newton method
Recursive estimation
Wireless communication
Online AccessGet full text

Cover

More Information
Summary:In this paper, a low complexity way to improve the Linear Least-Squares (LLS) method is introduced. The n-dimensional (n-D) positioning problem is first reduced to 1-D and then solved iteratively. Compared to the classic Gauss-Newton method, the n×n matrix inversion/factorization in each iteration is reduced to the inversion of a scalar. Simulations are performed to compare the Gauss-Newton, the LLS and the improved LLS method versus the Cramer-Rao Lower Bound (CRLB). The Mean Squared Error (MSE) of the obtained estimator is very close to that of the Gauss-Newton method, while the computational complexity is kept at almost the same level of the LLS approach.
ISBN:1424424232
9781424424238
DOI:10.1109/ICCS.2008.4737156