Hicham Tebbikh - Academia.edu (original) (raw)

Papers by Hicham Tebbikh

Journal européen des …, 2002

Résumé/Abstract The stochastic coloured well-formed nets «SCWN's» are based on two underlyin... more Résumé/Abstract The stochastic coloured well-formed nets «SCWN's» are based on two underlying model. The first one is the well formed coloured nets witch provide both a concise and a structured representation and are well adapted for representing big systems ...

Journal of Electronic Imaging, Jun 3, 2013

ABSTRACT One of the most recent techniques developed for face recognition is diagonal principal c... more ABSTRACT One of the most recent techniques developed for face recognition is diagonal principal component analysis (DiaPCA). In contrast to two-dimensional PCA (2DPCA), DiaPCA is based on diagonal representation of face images. Therefore, the correlations between variations of rows and those of columns can be kept. This paper proposes a novel projection scheme called bilateral DiaPCA (BDiaPCA) which simultaneously performs left and right projections of original face images in a new subspace derived from diagonal representation. Therefore BDiaPCA reduces the dimension of the original image matrix in both column and row directions. Experiments were carried out under varying illumination and facial expression using three well-known face databases: Yale, PF01, and a subset of FERET containing 200 people. Comparisons with 2DPCA and DiaPCA show that BDiaPCA outperforms these methods and achieves higher recognition rates using the Yang, Frobenius, or assembled matrix distance metrics.

Dans ce travail, on etudie la structure et la commande des systemes polynomiaux. On montre que la... more Dans ce travail, on etudie la structure et la commande des systemes polynomiaux. On montre que la classe des systemes homogenes de degre p de r#n a m entrees et q sorties, notee c#p(n, m, q) admet une realisation minimale en x(0)=0 dans c#p(n, m, q) ou n n et on donne un algorithme de calcul en un nombre fini de pas. Puis, on etudie le probleme de classification par retour d'etat des systemes quadratiques de r#n a m entrees. On montre le lien entre les invariants par feedback et les singularites de l'application entree-sortie. Des formes canoniques sont calculees et un systeme complet d'invariants est donne lorsque m=n1. On traite ensuite le probleme d'equivalence locale par feedback d'un systeme polynomial avec un systeme quadratique. Dans la partie commande des systemes polynomiaux, on etudie les problemes de controlabilite et de stabilite de ces systemes. On rappelle les c. N. S. De controlabilite des systemes homogenes impairs, dues a jurdjevic-kupka, et on explique pourquoi ce resultat ne se generalise pas au cas pair. On montre heuristiquement par l'etude de la controlabilite des systemes quadratiques que les systemes pairs sont peu controlables. La stabilisation est ensuite etudiee. On montre que les systemes impairs sont toujours stabilisables et on donne les c. N. S. De stabilite locale des systemes pairs. Pour cela, on fait usage pour l'aspect local de la theorie des varietes invariantes et en particulier centrales. Pour l'aspect global, on utilise la structure homogene de ces systemes ainsi qu'un certain resultat de jurdjevic-quinn. Comme application de ces techniques on etudie la stabilite des systemes quadratiques quand m differe peu de n

Nucleation and Atmospheric Aerosols, 2008

Face recognition is a specialized image processing which has attracted a considerable attention i... more Face recognition is a specialized image processing which has attracted a considerable attention in computer vision. In this article, we develop a new facial recognition system from video sequences images dedicated to person identification whose face is partly occulted. This system is based on a hybrid image feature extraction technique called ACPDL2D (Rouabhia et al. 2007), it combines two-dimensional principal component analysis and two-dimensional linear discriminant analysis with neural network. We performed the feature extraction task on the eyes and the nose images separately then a Multi-Layers Perceptron classifier is used. Compared to the whole face, the results of simulation are in favor of the facial parts in terms of memory capacity and recognition (99.41% for the eyes part, 98.16% for the nose part and 97.25 % for the whole face).

ABSTRACT In this article, we are studying, in small dimension control properties of a specific cl... more ABSTRACT In this article, we are studying, in small dimension control properties of a specific class of nonlinear systems, which occurs for instance in the attitude control problem of a rigid spacecraft.

International Journal of Systems Science, Mar 22, 2016

HAL is a multidisciplinary open access archive for the deposit and dissemination of scientific re... more HAL is a multidisciplinary open access archive for the deposit and dissemination of scientific research documents, whether they are published or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. L'archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d'enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

CORIA, 2010

Cet article propose une nouvelle mesure de similarité pondérée basée sur des matrices pour la cla... more Cet article propose une nouvelle mesure de similarité pondérée basée sur des matrices pour la classification et la reconnaissance de visages. Le calcul de distances s'effectue entre deux matrices caractéristiques obtenues par deux méthodes bidimensionnelles à savoir l'Analyse en Composantes Principales (ACP2D) et l'Analyse Discriminante Linéaire (ADL2D). Les poids de pondération utilisés sont l'inverse des valeurs propres de la matrice de covariance des matrices de visages triés dans l'ordre décroissant. Quant à la stratégie de classification des images test, elle est l'affectation au plus proche voisin. Pour évaluer les performances de la distance proposée, des tests de simulations ont été menés sur la base de visages internationale ORL. Cette dernière a été partitionnée en deux sous bases de 200 images chacune, l'une pour l'apprentissage et l'autre pour le test. Les résultats montrent l'efficacité et la supériorité de la distance pondérée vis-à-vis les distances de Yang et Frobenius. ABSTRACT.This paper proposes a novel weighted matrix similarity measure for image face classification and recognition. The distance is measured between two feature matrices obtained by two-dimensional Principal Component Analysis (2DPCA) and two-dimensional Linear Discriminant Analysis (2DLDA). The weights are the inverse of the eigenvalues of the total scatter matrix of faces matrices sorted in decreasing order and the classification strategy adopted is the nearest neighbour algorithm. To test and evaluate the efficiency of the proposed similarity measure, experiments were carried out using the ORL image face database which was partitioned into two sets, one for training and the other for testing. The experimental results show the high performances of the weighted matrix distance metric over the Yang and the Frobenius distances.

This paper aims at presenting a hybrid FMS short-term scheduling problem. Scheduling problem, we ... more This paper aims at presenting a hybrid FMS short-term scheduling problem. Scheduling problem, we are to study in this paper, is first introduced, and than a survey of the HFMS scheduling litterature is presented. Based on the art state of general scheduling algorithms, we present the Meta-heuristics; we have decided to apply for a given example of HFMS. That is the study of three advanced meta hauristics, Taboo searchTS, Genetic algorithms GA and simulated annealing SA.The HFMS model based on hierarchical Petri nets, helps to represent static and dynamic behaviour of the HFMS. This model is also used to design scheduling solutions. Hierarchical Petri nets model represents the whole process, and can be regarded as being made up a set of single timed colored Petri nets models. Each single model represents one process that is composed of many operations and tasks (manufacturing, transport, handling). TS, GA and SA are applied on each sub-model to make ressourses allocation and to resolve conflicts on shared production ressourses. A comparison between the three Meta-heuristics is made in simulation phasis to choose the best one to use in different kinds of HFMS.

International Conference on Signal Processing, Feb 15, 2006

This article presents a new approach so-called 'diffusive pseudo invariant' with application on t... more This article presents a new approach so-called 'diffusive pseudo invariant' with application on the robust control of a DC motor with uncertain parameters. The fundamental property of this approach is to preserve as much as possible, in the uncertainty domain, the dynamic characteristics of the system. The controllers obtained are of fractional nature and can be achieved in a non-hereditary way by diffusive model.

Archives of Control Sciences, 2011

In this paper, a new similarity measure is developed for human face recognition, namely, weighted... more In this paper, a new similarity measure is developed for human face recognition, namely, weighted matrix distance. The key difference between this metric and the standard distances is the use of matrices and weights rather than the vectors only. The two feature matrices are obtained by two-dimensional principal component analysis (2DPCA). The weights are the inverse of the eigenvalues sorted in decreasing order of the covariance matrix of all training face matrices. Experiments are performed under illumination and facial expression variations using four face image databases: ORL, Yale, PF01 and a subset of FERET. The results demonstrate the effectiveness of the proposed weighted matrix distances in 2DPCA face recognition over the standard matrix distance metrics: Yang, Frobenius and assembled matrix distance (AMD).

HAL (Le Centre pour la Communication Scientifique Directe), May 21, 2007

Dans ce papier, nous décrivons le problème de propagation des ondes dans les milieux viscoélastiq... more Dans ce papier, nous décrivons le problème de propagation des ondes dans les milieux viscoélastiques. Nous avons mis au point l'approche du calcul fractionnaire pour des systèmes hyperboliques avec amortissement viscoélastique fractionnaire. Nous avons présenté un modèle diffusif avec entrée-sortie ainsi qu'une méthode constructive réalisant cette opération de manière non héréditaire. L'application des schémas de Lax-Wendroff et de Lax-Friedrichs, aux systèmes viscoélastiques considérés, a permis de passer d'une forte formulation à une faible formulation facile à simuler numériquement.

This paper presents the stabilization of a hinged flexible beam attached to a rigid arm. The appl... more This paper presents the stabilization of a hinged flexible beam attached to a rigid arm. The applied control at the articulation point consists of the absorbing waves with the goal of minimizing the reflection of the energy at this point. The stabilization of the beam involves fractional operators that can be achieved in the non-hereditary way using a diffusive model. The numerical simulation results clearly show the powerful and the effectiveness of the control using diffusive representation.

The Open Automation and Control Systems Journal, 2017

The complexity and size are increasingly important in industrial processes; they lead to the supe... more The complexity and size are increasingly important in industrial processes; they lead to the supervision of a growing number of variables whose knowledge is generally based on measurements and/or on processes models. Introduction: Different monitoring methods are generally based on output measurements and/or process modeling. Materials and Methods: With the lack of precision and the monitoring speed in most work due to the model or measurement noise, for this we aim to develop a new and more powerful detection model based on clocks, to ensure a rapid detection and determine fault duration to facilitate the fault identification.

dcc.ufla.br

Abstract. Skin color detection is often an effective means used to define a set of candidate area... more Abstract. Skin color detection is often an effective means used to define a set of candidate areas likely to contain faces, hands, or other human organ in a scene. This can be performed by using human skin color models or by threshold of the appropriate color space. A few papers ...

Archives of Control Sciences, 2011

In this paper, a new similarity measure is developed for human face recognition, namely, weighted... more In this paper, a new similarity measure is developed for human face recognition, namely, weighted matrix distance. The key difference between this metric and the standard distances is the use of matrices and weights rather than the vectors only. The two feature matrices are obtained by two-dimensional principal component analysis (2DPCA). The weights are the inverse of the eigenvalues sorted in decreasing order of the covariance matrix of all training face matrices. Experiments are performed under illumination and facial expression variations using four face image databases: ORL, Yale, PF01 and a subset of FERET. The results demonstrate the effectiveness of the proposed weighted matrix distances in 2DPCA face recognition over the standard matrix distance metrics: Yang, Frobenius and assembled matrix distance (AMD).

Int. Arab J. Inf. Technol., 2016

The Multiple Targets Tracking (MTT) problem is addressed in signal and image processing. When the... more The Multiple Targets Tracking (MTT) problem is addressed in signal and image processing. When the state and measurement models are linear, we can find several algorithms that yield good performances in MTT problem, among them, the Multiple Hypotheses Tracker (MHT) and the Joint Probabilistic Data Association Filter (JPDAF). However, if the state and measurement models are nonlinear, these algorithms break down. In this paper we propose a method based on particle filters bank, where the objective is to make a contribution for estimating the trajectories of several targets using only bearings measurements. The main idea of this algorithm is to combine the Multiple Model approach (MM) with Sequential Monte Carlo methods (SMC). The result from this combination is a Nonlinear Multiple Model Particle Filters algorithm (NMMPF) able to estimate the trajectories of multiple targets.

Archives of Control Sciences, 2011

The paper addresses multiple targets tracking problem encountered in number of situations in sign... more The paper addresses multiple targets tracking problem encountered in number of situations in signal and image processing. In this paper, we present an efficient filtering algorithm to perform accurate estimation in jump Markov nonlinear systems, which we aim to contribute in solving the problem of multiple targets tracking using bearings-only measurements. The idea of this algorithm consists of the combination between the multiple model approach and particle filtering methods, which give a nonlinear multiple model particle filters algorithm. This algorithm is used to estimate the trajectories of multiple targets assumed to be nonlinear, from their noisy bearings.

Journal européen des …, 2002

Résumé/Abstract The stochastic coloured well-formed nets «SCWN's» are based on two underlyin... more Résumé/Abstract The stochastic coloured well-formed nets «SCWN's» are based on two underlying model. The first one is the well formed coloured nets witch provide both a concise and a structured representation and are well adapted for representing big systems ...

Journal of Electronic Imaging, Jun 3, 2013

ABSTRACT One of the most recent techniques developed for face recognition is diagonal principal c... more ABSTRACT One of the most recent techniques developed for face recognition is diagonal principal component analysis (DiaPCA). In contrast to two-dimensional PCA (2DPCA), DiaPCA is based on diagonal representation of face images. Therefore, the correlations between variations of rows and those of columns can be kept. This paper proposes a novel projection scheme called bilateral DiaPCA (BDiaPCA) which simultaneously performs left and right projections of original face images in a new subspace derived from diagonal representation. Therefore BDiaPCA reduces the dimension of the original image matrix in both column and row directions. Experiments were carried out under varying illumination and facial expression using three well-known face databases: Yale, PF01, and a subset of FERET containing 200 people. Comparisons with 2DPCA and DiaPCA show that BDiaPCA outperforms these methods and achieves higher recognition rates using the Yang, Frobenius, or assembled matrix distance metrics.

Dans ce travail, on etudie la structure et la commande des systemes polynomiaux. On montre que la... more Dans ce travail, on etudie la structure et la commande des systemes polynomiaux. On montre que la classe des systemes homogenes de degre p de r#n a m entrees et q sorties, notee c#p(n, m, q) admet une realisation minimale en x(0)=0 dans c#p(n, m, q) ou n n et on donne un algorithme de calcul en un nombre fini de pas. Puis, on etudie le probleme de classification par retour d'etat des systemes quadratiques de r#n a m entrees. On montre le lien entre les invariants par feedback et les singularites de l'application entree-sortie. Des formes canoniques sont calculees et un systeme complet d'invariants est donne lorsque m=n1. On traite ensuite le probleme d'equivalence locale par feedback d'un systeme polynomial avec un systeme quadratique. Dans la partie commande des systemes polynomiaux, on etudie les problemes de controlabilite et de stabilite de ces systemes. On rappelle les c. N. S. De controlabilite des systemes homogenes impairs, dues a jurdjevic-kupka, et on explique pourquoi ce resultat ne se generalise pas au cas pair. On montre heuristiquement par l'etude de la controlabilite des systemes quadratiques que les systemes pairs sont peu controlables. La stabilisation est ensuite etudiee. On montre que les systemes impairs sont toujours stabilisables et on donne les c. N. S. De stabilite locale des systemes pairs. Pour cela, on fait usage pour l'aspect local de la theorie des varietes invariantes et en particulier centrales. Pour l'aspect global, on utilise la structure homogene de ces systemes ainsi qu'un certain resultat de jurdjevic-quinn. Comme application de ces techniques on etudie la stabilite des systemes quadratiques quand m differe peu de n

Nucleation and Atmospheric Aerosols, 2008

Face recognition is a specialized image processing which has attracted a considerable attention i... more Face recognition is a specialized image processing which has attracted a considerable attention in computer vision. In this article, we develop a new facial recognition system from video sequences images dedicated to person identification whose face is partly occulted. This system is based on a hybrid image feature extraction technique called ACPDL2D (Rouabhia et al. 2007), it combines two-dimensional principal component analysis and two-dimensional linear discriminant analysis with neural network. We performed the feature extraction task on the eyes and the nose images separately then a Multi-Layers Perceptron classifier is used. Compared to the whole face, the results of simulation are in favor of the facial parts in terms of memory capacity and recognition (99.41% for the eyes part, 98.16% for the nose part and 97.25 % for the whole face).

ABSTRACT In this article, we are studying, in small dimension control properties of a specific cl... more ABSTRACT In this article, we are studying, in small dimension control properties of a specific class of nonlinear systems, which occurs for instance in the attitude control problem of a rigid spacecraft.

International Journal of Systems Science, Mar 22, 2016

HAL is a multidisciplinary open access archive for the deposit and dissemination of scientific re... more HAL is a multidisciplinary open access archive for the deposit and dissemination of scientific research documents, whether they are published or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. L'archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d'enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

CORIA, 2010

Cet article propose une nouvelle mesure de similarité pondérée basée sur des matrices pour la cla... more Cet article propose une nouvelle mesure de similarité pondérée basée sur des matrices pour la classification et la reconnaissance de visages. Le calcul de distances s'effectue entre deux matrices caractéristiques obtenues par deux méthodes bidimensionnelles à savoir l'Analyse en Composantes Principales (ACP2D) et l'Analyse Discriminante Linéaire (ADL2D). Les poids de pondération utilisés sont l'inverse des valeurs propres de la matrice de covariance des matrices de visages triés dans l'ordre décroissant. Quant à la stratégie de classification des images test, elle est l'affectation au plus proche voisin. Pour évaluer les performances de la distance proposée, des tests de simulations ont été menés sur la base de visages internationale ORL. Cette dernière a été partitionnée en deux sous bases de 200 images chacune, l'une pour l'apprentissage et l'autre pour le test. Les résultats montrent l'efficacité et la supériorité de la distance pondérée vis-à-vis les distances de Yang et Frobenius. ABSTRACT.This paper proposes a novel weighted matrix similarity measure for image face classification and recognition. The distance is measured between two feature matrices obtained by two-dimensional Principal Component Analysis (2DPCA) and two-dimensional Linear Discriminant Analysis (2DLDA). The weights are the inverse of the eigenvalues of the total scatter matrix of faces matrices sorted in decreasing order and the classification strategy adopted is the nearest neighbour algorithm. To test and evaluate the efficiency of the proposed similarity measure, experiments were carried out using the ORL image face database which was partitioned into two sets, one for training and the other for testing. The experimental results show the high performances of the weighted matrix distance metric over the Yang and the Frobenius distances.

This paper aims at presenting a hybrid FMS short-term scheduling problem. Scheduling problem, we ... more This paper aims at presenting a hybrid FMS short-term scheduling problem. Scheduling problem, we are to study in this paper, is first introduced, and than a survey of the HFMS scheduling litterature is presented. Based on the art state of general scheduling algorithms, we present the Meta-heuristics; we have decided to apply for a given example of HFMS. That is the study of three advanced meta hauristics, Taboo searchTS, Genetic algorithms GA and simulated annealing SA.The HFMS model based on hierarchical Petri nets, helps to represent static and dynamic behaviour of the HFMS. This model is also used to design scheduling solutions. Hierarchical Petri nets model represents the whole process, and can be regarded as being made up a set of single timed colored Petri nets models. Each single model represents one process that is composed of many operations and tasks (manufacturing, transport, handling). TS, GA and SA are applied on each sub-model to make ressourses allocation and to resolve conflicts on shared production ressourses. A comparison between the three Meta-heuristics is made in simulation phasis to choose the best one to use in different kinds of HFMS.

International Conference on Signal Processing, Feb 15, 2006

This article presents a new approach so-called 'diffusive pseudo invariant' with application on t... more This article presents a new approach so-called 'diffusive pseudo invariant' with application on the robust control of a DC motor with uncertain parameters. The fundamental property of this approach is to preserve as much as possible, in the uncertainty domain, the dynamic characteristics of the system. The controllers obtained are of fractional nature and can be achieved in a non-hereditary way by diffusive model.

Archives of Control Sciences, 2011

In this paper, a new similarity measure is developed for human face recognition, namely, weighted... more In this paper, a new similarity measure is developed for human face recognition, namely, weighted matrix distance. The key difference between this metric and the standard distances is the use of matrices and weights rather than the vectors only. The two feature matrices are obtained by two-dimensional principal component analysis (2DPCA). The weights are the inverse of the eigenvalues sorted in decreasing order of the covariance matrix of all training face matrices. Experiments are performed under illumination and facial expression variations using four face image databases: ORL, Yale, PF01 and a subset of FERET. The results demonstrate the effectiveness of the proposed weighted matrix distances in 2DPCA face recognition over the standard matrix distance metrics: Yang, Frobenius and assembled matrix distance (AMD).

HAL (Le Centre pour la Communication Scientifique Directe), May 21, 2007

Dans ce papier, nous décrivons le problème de propagation des ondes dans les milieux viscoélastiq... more Dans ce papier, nous décrivons le problème de propagation des ondes dans les milieux viscoélastiques. Nous avons mis au point l'approche du calcul fractionnaire pour des systèmes hyperboliques avec amortissement viscoélastique fractionnaire. Nous avons présenté un modèle diffusif avec entrée-sortie ainsi qu'une méthode constructive réalisant cette opération de manière non héréditaire. L'application des schémas de Lax-Wendroff et de Lax-Friedrichs, aux systèmes viscoélastiques considérés, a permis de passer d'une forte formulation à une faible formulation facile à simuler numériquement.

This paper presents the stabilization of a hinged flexible beam attached to a rigid arm. The appl... more This paper presents the stabilization of a hinged flexible beam attached to a rigid arm. The applied control at the articulation point consists of the absorbing waves with the goal of minimizing the reflection of the energy at this point. The stabilization of the beam involves fractional operators that can be achieved in the non-hereditary way using a diffusive model. The numerical simulation results clearly show the powerful and the effectiveness of the control using diffusive representation.

The Open Automation and Control Systems Journal, 2017

The complexity and size are increasingly important in industrial processes; they lead to the supe... more The complexity and size are increasingly important in industrial processes; they lead to the supervision of a growing number of variables whose knowledge is generally based on measurements and/or on processes models. Introduction: Different monitoring methods are generally based on output measurements and/or process modeling. Materials and Methods: With the lack of precision and the monitoring speed in most work due to the model or measurement noise, for this we aim to develop a new and more powerful detection model based on clocks, to ensure a rapid detection and determine fault duration to facilitate the fault identification.

dcc.ufla.br

Abstract. Skin color detection is often an effective means used to define a set of candidate area... more Abstract. Skin color detection is often an effective means used to define a set of candidate areas likely to contain faces, hands, or other human organ in a scene. This can be performed by using human skin color models or by threshold of the appropriate color space. A few papers ...

Archives of Control Sciences, 2011

In this paper, a new similarity measure is developed for human face recognition, namely, weighted... more In this paper, a new similarity measure is developed for human face recognition, namely, weighted matrix distance. The key difference between this metric and the standard distances is the use of matrices and weights rather than the vectors only. The two feature matrices are obtained by two-dimensional principal component analysis (2DPCA). The weights are the inverse of the eigenvalues sorted in decreasing order of the covariance matrix of all training face matrices. Experiments are performed under illumination and facial expression variations using four face image databases: ORL, Yale, PF01 and a subset of FERET. The results demonstrate the effectiveness of the proposed weighted matrix distances in 2DPCA face recognition over the standard matrix distance metrics: Yang, Frobenius and assembled matrix distance (AMD).

Int. Arab J. Inf. Technol., 2016

The Multiple Targets Tracking (MTT) problem is addressed in signal and image processing. When the... more The Multiple Targets Tracking (MTT) problem is addressed in signal and image processing. When the state and measurement models are linear, we can find several algorithms that yield good performances in MTT problem, among them, the Multiple Hypotheses Tracker (MHT) and the Joint Probabilistic Data Association Filter (JPDAF). However, if the state and measurement models are nonlinear, these algorithms break down. In this paper we propose a method based on particle filters bank, where the objective is to make a contribution for estimating the trajectories of several targets using only bearings measurements. The main idea of this algorithm is to combine the Multiple Model approach (MM) with Sequential Monte Carlo methods (SMC). The result from this combination is a Nonlinear Multiple Model Particle Filters algorithm (NMMPF) able to estimate the trajectories of multiple targets.

Archives of Control Sciences, 2011

The paper addresses multiple targets tracking problem encountered in number of situations in sign... more The paper addresses multiple targets tracking problem encountered in number of situations in signal and image processing. In this paper, we present an efficient filtering algorithm to perform accurate estimation in jump Markov nonlinear systems, which we aim to contribute in solving the problem of multiple targets tracking using bearings-only measurements. The idea of this algorithm consists of the combination between the multiple model approach and particle filtering methods, which give a nonlinear multiple model particle filters algorithm. This algorithm is used to estimate the trajectories of multiple targets assumed to be nonlinear, from their noisy bearings.