Robust Recursive RSSD Based Source Localization in Gaussian Mixture Channels

Hannan Lohrasbipeydeh | T. Aaron Gulliver

---

URL :   http://research.shahed.ac.ir/WSR/WebPages/Report/PaperView.aspx?PaperID=147982
Date :  2020/11/11
Publish in :    IEEE Communications Letters
DOI :  https://doi.org/10.1109/LCOMM.2020.3007191
Link :  http://dx.doi.org/10.1109/LCOMM.2020.3007191
Keywords :Gaussian mixture noise, received signal strength difference (RSSD), robust recursive least squares (RRLS

Abstract :

---

Signal strength based localization methods are of great interest due to their low cost and simple implementation. In this letter, a received signal strength difference (RSSD) based approach is presented to localize a source with unknown transmit power. A robust two stage estimator is proposed. First, an RSSD based optimization problem is formulated in the presence of Gaussian mixture measurement noise based on Huber’s minimax model. This yields a nonlinear problem which can be approximated by Taylor series expansion. Then, a robust recursive least squares (RRLS) method is developed to provide a robust solution of the approximated problem. The corresponding Cramér-Rao lower bound (CRLB) for correlated Gaussian mixture noise is also derived as a performance benchmark. Results are presented which show that the RRLS method attains the CRLB for a sufficiently large signal to noise ratio (SNR).