Robust statistical feature based aircraft identification

Abstract

The statistical feature based (StaF) classifier is presented for robust high range resolution (HRR) radar aircraft identification (ID). HRR signature peak features are selected "on the fly" with no a priori assumptions about the number or location of the features. Features extracted depends on the information content of the observed signature making the number, location, and amplitude of features random variables. A primary goal for this research is to increase classifier robustness by maintaining high known target ID while minimizing unknown target errors. Results are presented demonstrating that the StaF classifier can significantly reduce errors associated with unknown targets while maintaining a high probability of correct classification.