Some Statistical Aspects of Signal Processing

Abstract

This talk introduces my joint work with engineers at Air Force Research Laboratory, Signal Processing Branch. Basic signal processing system and electromagnetic data feature will be described briefly. Traditional signal detection algorithms assume homogeneous covariance matrix for all the secondary data. We deal with nonhomogeneity by a partitioning procedure, which screens out dissimilar secondary data prior to any detection. The partitioning statistic and its distribution will be used to derive P(CP), the probability of a correct partition. The procedure parameters can be found from P(CP) and be used to perform simulation studies and in data analysis. Simulation results and data analyses on US Air Force Multi-Channel Airborne Radar Measurement collected by Westinghouse will be presented to illustrate our theory. Overview: 1. Preparation and An Example 2. Complex Multivariate Normal Distribution 3. Radar Signal Detection Algorithms 4. Nonhomogeneity And A Partitioning Procedure 5. Simulation Results 6. A measured Data Example