A New Efficient Algorithm for Approximation
Abstract
Classical radar signal processing techniques assume that the signal interference is Gaussian in nature. However, it has been shown that this interference or clutter is not always Gaussian. When non-Gaussian clutter exists, other signal processing techniques which are optimal, or more robust in non-Gaussian clutter may be more effective than the classical techniques. This requires determination of the clutter characteristics for each clutter region and then applying the appropriate signal processing technique to the data ideally in real time. In order to achieve real time it is necessary to determine this approximate Probability Density Function (PDF) using small sample data set sizes. However, until the development of the Ozturk Algorithm, there has not existed an efficient algorithm to determine an approximate PDF for a small clutter data sample set. The Ozturk Algorithm is a new statistical algorithm capable of approximating the PDF of a set of random data using on the order of 100 sample points, whereas classical techniques typically require thousands of samples. It consists of two parts, a goodness of fit test and the PDF Approximation. The goodness of fit test determines whether a sample data set is statistically consistent with a given PDF. The PDF Approximation selects the best approximate PDF from a variety of PDFs and is simply an extension of the goodness of fit test. This report describes the Ozturk Algorithm and shows an application of the algorithm to some temporal L-band radar clutter data.
Document Details
- Document Type
- Technical Report
- Publication Date
- Aug 01, 1997
- Accession Number
- ADA329961
Entities
People
- Lias K. Slaski
- Murali Rangaswamy
Organizations
- Rome Laboratory