Title :
An H∞ approach to multi-source tracking
Author :
Ratnarajah, Tharm
Author_Institution :
Commun. Res. Lab., McMaster Univ., Hamilton, Ont.
Abstract :
In this paper, a novel H∞ approach is proposed for tracking of polarized co-channel sources using an array of tripole antennas. The proposed approach partitions the observation data matrix into two sub-matrices that are used, in conjunction with a new state-space model, to provide an H∞-type recursive estimation of a linear combiner. The linear combiner then provides estimates of the noise and signal subspaces, from which the directions of the incident signals can be estimated and tracked. The proposed technique is also capable of handling the tracking of appearing/disappearing sources during the observation interval and, furthermore, can accommodate array modeling uncertainties. The difficult problem of tracking the crossing sources can be successfully handled by using diversely polarized array
Keywords :
H∞ optimisation; antenna arrays; array signal processing; direction-of-arrival estimation; matrix algebra; noise; recursive estimation; state-space methods; target tracking; H∞ approach; H∞-type recursive estimation; appearing sources; array modeling uncertainties; crossing sources; direction; disappearing sources; diversely polarized array; incident signals; linear combiner; multi-source tracking; noise subspace; observation data matrix; partitioning; polarized co-channel sources; signal subspace; state-space model; sub-matrices; tripole antenna array; Additive white noise; Azimuth; Bismuth; Covariance matrix; Eigenvalues and eigenfunctions; Equations; Estimation error; Polarization; Sensor arrays; Vectors;
Conference_Titel :
Acoustics, Speech and Signal Processing, 1998. Proceedings of the 1998 IEEE International Conference on
Conference_Location :
Seattle, WA
Print_ISBN :
0-7803-4428-6
DOI :
10.1109/ICASSP.1998.681585