Hyperbolic Transforms in Array Processing
Abstract
The subject of the research is detection and estimation employing an array of sensors. Of particular concern is efficient and numerically reliable computational strategies for implementing prevalent detection/estimation procedures. A number of important array processing problems lead to a differencing of matrix outer products. This leads to potential ill-conditioning when implemented explicitly. The avoidance of outer products, at the expense of extra operations, has long been a crusade of sorts in the numerical analysis community (Golub). One can do without outer products by means of orthogonal, or for complex data unitary, transforms in the usual case where a sum of outer products arise. (Typical transforms that have been found to be particularly useful are Givens, Jacobi, and Householder transforms). The idea is to transform the data into sparse form while preserving its pertinent statistics (usually the sample covariance matrix). This research concerns generalizing this 'trick' to the case of a difference of outer products by means of hyperbolic, rather than orthogonal transforms.
Document Details
- Document Type
- Technical Report
- Publication Date
- Feb 28, 1991
- Accession Number
- ADA247061
Entities
People
- Allan O. Steinhardt
Organizations
- Cornell University College of Engineering