Author(s): ABDOU Latifa
Journal: Journal of Electrical and Electronics Engineering
ISSN 1844-6035
Volume: 4;
Issue: 2;
Start page: 11;
Date: 2011;
VIEW PDF
DOWNLOAD PDF 
Original page
Keywords: distributed system | OS-CFAR sensor | dependent sensors | evolutionary strategy | genetic algorithm
ABSTRACT
This paper proposes an original method forthreshold optimization in distributed ordered statisticsconstant false alarm rate (OS-CFAR) by using anappropriate Evolution Strategy (ES). Two fusion rules;“AND” and “OR” were considered in the case wherethe signals are dependant from sensor to sensor. Likethe Genetic Algorithms (GAs), the ESs are heuristicoptimization techniques proposed in literature foroptimization of real parameters. In contrast with GAs,the ESs work on the so-called phenotype space,without any mathematic representation of parameters.We proposed an ES technique by which a selfadaptationmutation is used. The results showed thatthe proposed ES method is a flexible technique, whenapplied to problems of optimization in CFAR systems,even if the observations are dependent from sensor tosensor. Also the same technique proved its capability toimprove the performance of the system, in somesituations, when compared to the GA.
Journal: Journal of Electrical and Electronics Engineering
ISSN 1844-6035
Volume: 4;
Issue: 2;
Start page: 11;
Date: 2011;
VIEW PDF
DOWNLOAD PDF Original pageKeywords: distributed system | OS-CFAR sensor | dependent sensors | evolutionary strategy | genetic algorithm
ABSTRACT
This paper proposes an original method forthreshold optimization in distributed ordered statisticsconstant false alarm rate (OS-CFAR) by using anappropriate Evolution Strategy (ES). Two fusion rules;“AND” and “OR” were considered in the case wherethe signals are dependant from sensor to sensor. Likethe Genetic Algorithms (GAs), the ESs are heuristicoptimization techniques proposed in literature foroptimization of real parameters. In contrast with GAs,the ESs work on the so-called phenotype space,without any mathematic representation of parameters.We proposed an ES technique by which a selfadaptationmutation is used. The results showed thatthe proposed ES method is a flexible technique, whenapplied to problems of optimization in CFAR systems,even if the observations are dependent from sensor tosensor. Also the same technique proved its capability toimprove the performance of the system, in somesituations, when compared to the GA.