Optical Linear Algebra Processors for Adaptive Signal Processing.
Abstract
This report summarizes effort concerning the use of optical linear algebra processors (OLAP's) for adaptive radar signal processing. This effort involved extensive algorithm study, significant software development and optical laboratory hardware and architectural work. Adaptive phased array radar (APAR) relies heavily on the solution of matrix vector equations and eigenvalue/eigenvector problems, both of which are suited for OLAP implementation. This report, presents a new method to estimate the location, center frequency and bandwidth data on noise sources using a new acousto-optic (AO) architecture. Initial test results are provided from computer simulations of the performance of several APAR algorithms implemented on an OLAP. The results indicate that one formulation of the steepest decent algorithm to be preferable for optical implementation. Details are provided about a laboratory OLAP near completion. This handles complex data (needed in APAR) and can provided high accuracy. Details are provided for encoding the data for OLAP and the associated error modeling.
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 01, 1985
- Accession Number
- ADA177118
Entities
People
- B. V. Vijaya Kumar
- David P. Casasent
Organizations
- Carnegie Mellon University