Title :
Quadratic Variational Framework for Signal Design on the 2-Sphere
Author :
Wei, Liying ; Kennedy, Rodney A. ; Lamahewa, Tharaka A.
Author_Institution :
Coll. of Eng. & Comput. Sci., Australian Nat. Univ., Canberra, ACT, Australia
Abstract :
This paper introduces a quadratic variational framework for solving a broad class of signal design problems on the 2-sphere. The functional, to be extremized, combines energy concentration measures using a weighting function in the spatial domain, multiplicative weights in the spectral domain, and a total energy constraint. This leads to two formulations of the signal design problem on the 2-sphere, one a Fredholm integral equation in the spatial domain and the other an infinite matrix equation in the spectral domain. The framework is illustrated by deriving the key equations for the two classical spatio-spectral concentration problems on the 2-sphere, and for an isotropic filter design that maximizes the filtered energy. In addition, using the proposed framework, we formulate a joint 3-D beamforming application which achieves optimal directivity and spatial resolution simultaneously.
Keywords :
Fredholm integral equations; array signal processing; matrix algebra; signal resolution; variational techniques; 2-sphere; Fredholm integral equation; energy concentration measure; energy constraint; infinite matrix equation; isotropic filter design; joint 3D beamforming; multiplicative weight; optimal directivity; quadratic variational framework; signal design; spatial domain; spatial resolution; spatio-spectral concentration problem; spectral domain; weighting function; Energy measurement; Harmonic analysis; Optimization; Signal design; Spectral analysis; Time frequency analysis; Weight measurement; Directional derivative; isotropic convolution; uncertainty principle; unit sphere;
Journal_Title :
Signal Processing, IEEE Transactions on
DOI :
10.1109/TSP.2011.2162506