les Thèses Soutenues à l'USTO MB
vous cherchez une thèse?
A partir de cette page vous pouvez :
Retourner au premier écran avec les dernières notices... | Votre compte |
Détail de l'auteur
Auteur Abdelhadi LOTFI
Documents disponibles écrits par cet auteur
Faire une suggestion Affiner la rechercheDétection de cibles radar par les réseaux de neurones probabilistes en milien bruité / Abdelhadi LOTFI
Titre : Détection de cibles radar par les réseaux de neurones probabilistes en milien bruité Type de document : texte imprimé Auteurs : Abdelhadi LOTFI, Auteur Année de publication : 2005 Importance : 63 p. Langues : Anglais (eng) Catégories : Informatique:RFIA Mots-clés : Les réseauxde neurons, detection,PNN,radar,RBF,MLP,gaussian Résumé : Detection of signal in noise is sometimes a difficult task , this difficulty grows bigger when the noise ‘s distribution is not Gaussian , in this work , probabilistic neural networks (PNN) are applied to radar target detection in non –gaussian noise PNN are a special class oDetection of signal in noise is sometimes a difficult task , this difficulty grows bigger when the noise ‘s distribution is not Gaussian , in this work , probabilistic neural networks (PNN) are applied to radar target detection in non –gaussian noise PNN are a special class of RBFneural networks and hence belong to MLPnetworks , however , unlike MLPs , PNNs benefit from solid mathematical basis , they are used to estimate probability density functions (pdf ) from the solid mathematical basis , they are used to estimate probality density functions (pdf ) from the input data space , tow PNNs were presented , namely gram –charlier and generalized probabilistic neural networks (GCNN and GPNN) this includes the mathematical basis , the architecture and the training algorithm for each network
Application to radar target detection presented in this work showed that probabilistic neural networks outperform other neural networks for almost all cases of noise , the performance was evaluated in terms of probability of detection versus signal to noise ratio and falso alarms (false positives ) generated by each detector , moreover , the architecture of PNNs allows a parallel computing algorithm with can reduce significantly the time of traiing and tests A computing Grid can be used successfully here
Directeur de thèse : BENYETTOU,A Détection de cibles radar par les réseaux de neurones probabilistes en milien bruité [texte imprimé] / Abdelhadi LOTFI, Auteur . - 2005 . - 63 p.
Langues : Anglais (eng)
Catégories : Informatique:RFIA Mots-clés : Les réseauxde neurons, detection,PNN,radar,RBF,MLP,gaussian Résumé : Detection of signal in noise is sometimes a difficult task , this difficulty grows bigger when the noise ‘s distribution is not Gaussian , in this work , probabilistic neural networks (PNN) are applied to radar target detection in non –gaussian noise PNN are a special class oDetection of signal in noise is sometimes a difficult task , this difficulty grows bigger when the noise ‘s distribution is not Gaussian , in this work , probabilistic neural networks (PNN) are applied to radar target detection in non –gaussian noise PNN are a special class of RBFneural networks and hence belong to MLPnetworks , however , unlike MLPs , PNNs benefit from solid mathematical basis , they are used to estimate probability density functions (pdf ) from the solid mathematical basis , they are used to estimate probality density functions (pdf ) from the input data space , tow PNNs were presented , namely gram –charlier and generalized probabilistic neural networks (GCNN and GPNN) this includes the mathematical basis , the architecture and the training algorithm for each network
Application to radar target detection presented in this work showed that probabilistic neural networks outperform other neural networks for almost all cases of noise , the performance was evaluated in terms of probability of detection versus signal to noise ratio and falso alarms (false positives ) generated by each detector , moreover , the architecture of PNNs allows a parallel computing algorithm with can reduce significantly the time of traiing and tests A computing Grid can be used successfully here
Directeur de thèse : BENYETTOU,A Exemplaires
Code-barres Cote Support Localisation Section Disponibilité 4308 02-08-148 version papier Bibliothèque Centrale Mémoire de Magister Exclu du prêt
BUC USTOMB'Thèses
Service Thèse de la BUC met à votre votre disposition L'ensemble des thèses de doctorat et mémoires de magister soutenues à l'USTO MB
Adresse
BUC USTOMB'ThèsesBibliothèque centrale USTOMB
BP 1505 EL M'Naouer USTO ORAN
Algérie
(213)041627180
contact