Improved Neural Network Based CFAR for Non Homogeneus Background and Multiple Target Situations

GÁLVEZ NELIDA; COUSSEAU JUAN E; PASCIARONI JOSE; AGAMENNONI OSVALDO

LATIN AMERICAN APPLIED RESEARCH

PLAPIQUI(UNS-CONICET)

Lugar: Bahia Blanca; Año: 2012 vol. 42 p. 343 - 350

0327-0793

The Neural Network Cell Average - Order Statistics Constant False Alarm Rate (NNCAOS CFAR) detector is presented in this work. NNCAOS CFAR is a combined detection methodology which uses the effectiveness of neural networks to search for non homogeneities like clutter banks and multiple targets within the radar return. In addition, the methodology proposed applies a convenient cell average (CA) or order statistics (OS) CFAR detector according to the context situation. Exhaustive analysis and comparisons show that NNCAOS CFAR has better performance than CA CFAR, OS CFAR and even CANN CFAR detectors (the latter, a previously proposed neural network based detector). Furthermore, it is verified that the new proposal presents a robust operation when maintaining a constant probability of false alarm under different radar return situations.