Mohamed Chtourou

• Prof. HDR, PhD, Eng.
• Principal Investigator at University of Sfax

This paper examines the formulation and implementation of a neuro-controller for the excitation system of synchronous generators in a Single-Machine Infinite Bus (SMIB) power system. The SMIB model is employed as a fundamental model of a power system, thereby facilitating the assessment and comparison of disparate control strategies with the object...

The stability of the power system is important for ensuring the power grid's reliable operation. Traditional stability analysis and stabilization methods have limitations in handling the complexity and nonlinearity of modern power systems. The combination of artificial neural network (ANN)based approaches holds particular promise for power system s...

The real world is nonlinear and in the control application field, this aspect needs to be resolved to build models so we need to refer to nonlinear system modeling techniques. Neuro-fuzzy systems and modular neural networks (NNs) are among the best modeling approaches for nonlinear systems. The combined features of both approaches provide better mo...

Recent years have seen advances in deep learning, including in the field of traffic management. Detecting distant objects that occupy a small number of pixels in the input image is one of the major challenges in computer vision for several reasons, including limited resolution. The challenges of detecting the rotation of objects may be attributed t...

This paper presents the latest advances in machine learning techniques and highlights deep learning (DL) methods in recent studies. This technology has recently received great attention as it can solve complex problems. This paper focuses on covering one of the deep learning algorithms (deep neural network) and learning about its types such as conv...

This paper presents an advanced control method for the stabilization of Electric power systems. This method is a decentralized control strategy based on a set of neural controllers. Essentially, the largescale power system is decomposed into a set of subsystems in which each one is constituted by a single machine connected to a variable bus. For ea...

In this paper, we propose an apple ripeness detection system based on deep learning methods. Three deep learning models, namely, Mask R-CNN, YOLOv5 and YOLOx, representing two-stage and one-stage object detectors, are employed to conduct apple fruit ripeness detection. Digital images related to three apple ripeness stages are collected from real sc...

The efficient operation of the solar photovoltaic (PV) system is influenced by the accurate determination of its model parameters. The appropriate selection of these parameters is a challenging task due to the multi-modality and non-linearity of the problem. Given less sensitivity to initial solutions and the powerful search capability, many metahe...

The environmental impact and scarcity of fossil fuels have led to a tremendous increase in the demand and use of renewable energy resources. The use of photovoltaic (PV) energy is due to several benefits, such as low maintenance and operating costs. The modeling of the PV systems process requires the identification of PV cells parameters, which can...

This paper proposes an innovative identification approach of nonlinear stochastic systems using Hammerstein–Wiener (HW) model with output-error autoregressive (OEA) noise. Two fuzzy systems are suggested for the identification of the input and output nonlinear blocks of a proposed model from given input-output data measurements. In this work, the n...

In this study, a fuzzy classification approach based on colour features has been investigated to estimate the ripeness of apple fruits according to three maturity stages; unripe, turning‐ripe and ripe. The K nearest neighbour algorithm was applied in order to segment the fruit image into four regions namely background, green area, yellow area and r...

In this paper, we propose the Particle Swarm Optimization (PSO) based approach of mobile robot path planning. It is the PSO based Canonical Force Field approach (PSO-CF²). The PSO optimization technique is applied here in the design of the shortest trajectory to the goal with the best collision avoidance. The PSO searches for the best combination o...

This paper deals with the control of nonlinear systems where the multimodel approach has been used to build global controller. In multimodel approaches, two problems could be generally encountered: (1) how to find the required number of models and (2) what are their locations in an operating space. The developed method integrates gap metric, margin...

In this paper we are interested to adaptive fuzzy controller for nonlinear SISO systems in presence of parametric uncertainties. The plant model structure is represented by a fuzzy system. The essential idea of the on-line parametric estimation of the plant model is based on a comparison of the measured state with the estimate one. The design of ad...

This paper focuses on the indirect adaptive fuzzy control of Single Input Single Output (SISO) non-linear systems with unknown linearities. The proposed adaptive fuzzy controller is based on feedback linearisation. Its parameters are updated online according to some adaptive laws such as tracking error-based method, composite tracking and modelling...

In this paper, we have investigated the application of a hybrid algorithm to indirect adaptive control of nonlinear system using a multilayer perceptron. The proposed hybrid algorithm is employed to reduce the learning speed of neural controller. The overall control scheme includes two neural networks. The first is used to identify the nonlinear sy...

In this paper, based on the combination of particle swarm optimization (PSO) algorithm and neural network (NN), a new adaptive speed control method for a permanent magnet synchronous motor (PMSM) is proposed. Firstly, PSO algorithm is adopted to get the best set of weights of neural network controller (NNC) for accelerating the convergent speed and...

In this paper, based on the combination of particle swarm optimization (PSO) algorithm and neural network (NN), a new adaptive speed control method for a permanent magnet synchronous motor (PMSM) is proposed. Firstly, PSO algorithm is adopted to get the best set of weights of neural network controller (NNC) for accelerating the convergent speed and...

Abstract—Fruit ripeness estimation is an important process that affects its quality and subsequently its marketing. Automatic ripeness evaluation through computer vision system has been an innovative topic interesting many researchers as it provides efficient solution to the slow speed, time consumption and high cost associated with the manual asse...

In this paper, based on the combination of particle swarm optimization (PSO) algorithm and neural network (NN), an adaptive neural network internal model control (NNIMC) is designed for a permanent magnet synchronous motor (PMSM). Firstly, in order to accelerate the convergent speed and to prevent problems of trapping in local minimum, PSO algorith...

The present paper deals with a new fuzzy strategy of model-based predictive control for nonlinear systems that are assumed to be interconnected (INS). Each subsystem, supposed nonlinear (NSS), is represented by a Takagi–Sugeno fuzzy model where fuzzy rule conclusions are linear state space models. The proposed control strategy consists of the assoc...

The robustness of a visual servoing task depends mainly on the efficiency of visual selections captured from a sensor at each robot’s position. A task function could be described as a regulation of the values sent via the control law to the camera velocities. In this paper we propose a new approach that does not depend on matching and tracking resu...

Purpose The purpose of this paper is to use the internal model control to deal with nonlinear stable systems affected by parametric uncertainties. Design/methodology/approach The dynamics of a considered system are approximated by a Takagi-Sugeno fuzzy model. The parameters of the fuzzy rules premises are determined manually. However, the paramete...

This work deals with robust inverse neural control strategy for a class of single-input single-output (SISO) discrete-time nonlinear system affected by parametric uncertainties. According to the control scheme, in the first step, a direct neural model (DNM) is used to learn the behavior of the system, then, an inverse neural model (INM) is synthesi...

This paper presents a constructive training algorithm applied to face recognition and facial expression recognition. The multi layer perceptron (MLP) neural network is formed by a single hidden layer using a predefined number of neurons and a small number of training patterns. During the learning, the hidden neuron number is incremented when the me...

In this paper, a new validity criterion determining the optimal decomposition and the minimum linear model bank was employed to design a controller for a nonlinear system. Based on the gap metric, the proposed validity criterion quantifies the compactness degree for a model-base. For a decomposition giving a low degree of compactness, the optimal l...

This paper deals with the control of nonlinear systems using multimodel approach. The main idea of this work consists on the association of the gap metric and the stability margin tools to reduce the number of models constituting the multimodel bank. In fact, the self-organisation map (SOM) algorithm is used, firstly, to develop a preliminary multi...

This paper proposes fuzzy model predictive control (FMPC) strategies for nonlinear interconnected systems based mainly on a system decomposition approach. First, the Takagi-Sugeno (TS) fuzzy model is formulated in such a way to describe the behavior of the nonlinear system. Based on that description, a fuzzy model predictive control is determined....

This paper proposes a new approach to achieve real-time robotic control using image-based visual servoing in the presence of constraints. The proposed method deals with visibility constraints and occultation avoidance using potential fields. Our approach aims to improve the real-time performance of the visual servoing scheme. In fact, the proposed...

This paper proposes a new approach to enhance real-world intensity-based visual servoing. The main goal is to direct successfully a robotic task without going through the entire image. An explicit two-dimensional map is used as a new global descriptor index. When sweeping the conspicuity of interest regions, a stratified random process recovers the...

Choosing the training algorithm and determining the architecture of artificial neural networks are very important issues with large application. There are no general methods which permit the estimation of the adequate neural networks size. In order to achieve this goal, a pruning algorithm based on the relevancy index of hidden neurons outputs is d...

Choosing the training algorithm and determining the architecture of artificial neural networks are very important issues with large application. There are no general methods which permit the estimation of the adequate neural networks size. In order to achieve this goal, a pruning algorithm based on the relevancy index of hidden neurons outputs is d...

The Canonical Force Field (CF2) method is an approach of mobile robot path planning. The variations of CF2 parameters P, c, k, Q and ρ0 are however vital to its performance. In this paper, we used the multi-objective particle swarm optimization (PSO) approach to optimize these parameters. The computation of the optimal parameters is restarted in ea...

In this paper, a novel adaptive tuning method of PID neural network (PIDNN) controller for nonlinear process is proposed. The presented method utilizes an improved gradient descent method to adjust PIDNN parameters where the margin stability will be employed to get high tracking performance and robustness with regard to external load disturbance an...

This paper proposes a novel gap metric based fuzzy decomposition approach resulting in a reduced model bank that provides enough information to design controllers. It requires, first, the determination of the model base. For this, the number of initial models is obtained via fuzzy c-means (FCM) algorithm. Then, a gap metric based method which aims...

This paper proposes a new idea for modeling complex systems: it is a modified version of the Multi-Network Neural Model ("MNNM") to solve the problem of nonlinear systems modeling. In fact, the Multi-Network Neural Model was carried out using an algorithm consisting on training simultaneously all local neural networks, then, the interpolation of th...

This paper describes a novel approach for mobile robot path planning in known dynamic environments. It is the Particle Swarm Optimization Dynamic Variable Speed Force Field (PSO-DVSF2) based on the Force Field (F2) approach proposed by D. Wang et al [5]. The basic concept of PSO-DVSF2 is to generate a continually changing parameterised Force Field...

Purpose – The purpose of this paper is to deal with the stabilization of the continuous Takagi Sugeno (TS) fuzzy models using their discretized forms based on the decay rate performance approach. Design/methodology/approach – This approach is structured as follows: first, a discrete model is obtained from the discretization of the continuous TS fu...

The 12th International Multi-Conference on Systems, Signals & Devices : 2015

This paper deals with the control of the continuous Takagi Sugeno (TS) fuzzy models using discrete control approach. This approach is structured as follows: first, a discrete model is obtained from the discretization of the continuous TS fuzzy model. In this case, the Euler discretization is used for order of approximation superior to one. Second t...

This paper deals with fuzzy modeling of nonlinear systems affected by bounded uncertainties. The proposed model is composed of two parts: a linear uncertain part and a nonlinear part. The linear uncertain part is obtained by system linearization around some operating points. Nonlinear part parameters are estimated through the use of the descent gra...

The PSO based-Canonical Force Field (PSO-CF2) method is a novel approach of mobile robot path planning with collision avoidance. The choice of CF2 parameters is however vital to its performance. In this paper, we propose the multi-objective Particle Swarm Optimization Canonical Force Field (PSO-CF2) technique to search for the best combination of t...

This paper concerns a robust pole assignment for the control of uncertain discrete-time nonlinear systems. A composed model of two parts is used to describe the dynamic of the considered system. The first part is linear affected by bounded uncertainties. It is obtained by the nominal system linearization around some operating points. The second par...

Fuzzy and predictive control methods are two modern control strategies that have been accepted by the industry to describe and solve complex problems. The present paper introduces a fuzzy control technique, which belongs to the popular family of control algorithms, called model predictive control (MPC). This method is based on the use of a Takagi-S...

This paper focuses on the study of modified constructive training algorithm for Multi Layer Perceptron 'MLP' which is applied to face recognition applications. In general, constructive learning begins with a minimal structure, and increases the network by adding hidden neurons until a satisfactory solution is found. The contribution of this paper i...

This paper presents new approach illustrating robust visual servoing based on global visual features: random distribution of limited set of pixels luminance. Our approach aims to improve the real-time performance of the visual servoing scheme. In fact, the use of our new features reduces the computation time of the visual servoing task and removes...

This paper deals with fuzzy modeling and robust control of nonlinear systems affected by bounded uncertainties. The proposed fuzzy model is composed of two parts: a linear uncertain part and a nonlinear one. The linear uncertain part is obtained by the nominal system linearization around some operating points. The nonlinear part is approximated by...

This study presents a modified constructive training algorithm for multilayer perceptron (MLP) which is applied to face recognition problem. An incremental training procedure has been employed where the training patterns are learned incrementally. This algorithm starts with a small number of training patterns and a single hidden-layer using an init...

In this paper, a hybrid method is proposed to control a nonlinear dynamic system using feedforward neural network. This learning procedure uses different learning algorithm separately. The weights connecting the input and hidden layers are firstly adjusted by a self organized learning procedure, whereas the weights between hidden and output layers...

This paper presents a constructive training algorithm for Multi Layer Perceptron (MLP) applied to facial expression recognition applications. The developed algorithm is composed by a single hidden-layer using a given number of neurons and a small number of training patterns. When the Mean Square Error MSE on the Training Data TD is not reduced to a...

Since the beginning of the fuzzy control theory, results have been obtained independently for continuous and discrete models. It is still quite difficult to use non quadratic Lyapunov functions for the continuous case, while this is much easier for the discrete case. This approach tries to put a bridge between the continuous and discrete cases for...

This paper proposes a Multi-Network Neural Model (“MNNM”) to deal with complex systems modeling. Indeed, the training of this architecture was performed using a parallel algorithm consisting on training simultaneously all local neural networks. The obtained results show the effectiveness of the “MNNM” compared to the Single-Network Neural Model (“S...

This paper presents robust visual servoing approach based on global descriptor. Our work aims to improve realtime performance of the visual servoing scheme. Indeed, the use of our new descriptor reduces the computation time of the visual servoing task. The error-dynamics considered in all visual servoing schemes were, usually, a first-order dynamic...

This work presents two methods of selection of neural models for identification of dynamic systems. Initially, a strategy of selection based on statistical tests, which relates to training and generalisation performances of a neural model is analysed. In the second time, a new constructive approach of neural model selection, which the training begi...

This chapter proposes a newapproach to achieve real-time visual servoing tasks. Our contribution consists in the definition of new global visual features as a random distribution of limited set of pixels luminance. The new method, based on a random process, reduces the computation time of the visual servoing scheme and removes matching and tracking...

Training and topology design of artificial neural networks are important issues with large application. This paper deals with an improved algorithm for feed forward neural networks (FNN)s training. The association of an incremental approach and the Lyapunov stability theory accomplishes both good generalization and stable training process. The algo...

This paper introduces a neural network optimization procedure allowing the generation of multilayer perceptron (MLP) network topologies with few connections, low complexity and high classification performance for phoneme’s recognition. An efficient constructive algorithm with incremental training using a new proposed Frame by Frame Neural Networks...

Since a few years, LMIs conditions associated to the control of continuous Takagi Sugeno (TS) fuzzy models have used non quadratic Lyapunov functions. Indeed they are much more general than classical quadratic functions. However, there are requirements about the derivative of the membership functions appearing in the LMIs. Whereas, this problem doe...

A new learning algorithm suited for training multilayered neural networks that we have named hybrid is hereby introduced. With this algorithm the weights of the hidden layer are adjusted using the Kohonen algorithm. While the weights of the output layer are trained using a gradient descent method with adaptive learning parameter based Lyapunov func...

This paper presents a new approach for fuzzy rule base reduction using similarity concepts and interpolation techniques. The algorithm consists on: First, measure similarity between rules for the best choice of which of them will be deleted. This operation is done without modification of membership functions. Second, if a new input data is presente...

This paper investigates the applicability of the constructive approach proposed by D. Liu et al. [IEEE Trans. Circuits Syst. 49, No. 12, 1876–1879 (2002; doi:10.1109/TCSI.2002.805733)] to wavelet neural networks (WNN). In fact, two incremental training algorithms will be presented. The first one, known as one pattern at a time (OPAT) approach, is t...

The control of variable speed wind turbines is a complex problem since they are considered as nonlinear and time varying systems. In general, classical control techniques do not take into consideration the stochastic and dynamical aspect of the wind and they are not very robust. In order to address these weaknesses, neural approaches are proposed:...

A novel neural network architecture, is proposed and shown to be useful in approximating the unknown nonlinearities of dynamical systems. In the variable structure neural network, the number of basis functions can be either increased or decreased with time according to specified design strategies so that the network will not overfit or underfit the...

Similar to neural networks, the generalization improvement of wavelet neural networks is also an important issue since a given network may have good approximation accuracy, but could not perform well on unseen data. Generally, to improve generalization different techniques could be used including regularization. In this paper, two newly regularizat...

This paper proposes a novel method to improve mobile robot robustness with respect to kinematic modeling errors during visual servoing task. Instead of using first-order error-dynamics, as it is usually done, we use the second-order error-dynamics leading to a new control law. The main aim of this approach is to guarantee a robust visual servoing s...

Topology design of artificial neural networks (ANNs) is a complex problem. This paper presents a study of some approaches which derived from a pruning technique (OBS). In the first step, we explicit the corresponding algorithms used to determine the adequate number of neurons and weights for neural structure. In the second step, a comparative study...

Multimedia design such as video decoders are typically composed of several communicating tasks. Each task is characterized by its workload variation. The target device of this kind of application contains several processing unit. This calls for a dynamic management of hardware units to improve the QOS of the application and to optimally allocate re...

In this work, we use the approach based on neural observer in order to introduce the diagnosis of a non-linear system. The synthesis of such a trained specific observer using the back-propagation algorithm leads to an estimation study then a determination of fault diagnosis and isolation of single actuator fault based on residual generation. The ro...

In this paper we introduce a new robust visual scheme intended to 2D visual servoing robotic tasks. The main object is to direct the robot to its desired position. To be able to carry out such a task robustly the tough and major step is primarily the image processing procedure. We should find good selections of visual data in order to be correctly...

Modeling complex systems is known as complex problem. Indeed , various studies have been used in order to facilitate the task of modeling in term of time and quality. The multi-model approach was recently developed to deal with the problems of complex systems. So, in this paper we propose a neural approach based on multi-models for modeling nonline...

In this paper, an internal multiple model control (IMMC) based on linear model's library is introduced. This approach supposes the definition of a set of local linear models. However, it remains beset with several difficulties such as the determination of the local models base. A new approach that combines fuzzy c-means (FCM) clustering algorithm a...

Purpose – The purpose of this paper is to deal with the stabilization of the continuous-time Takagi-Sugeno (TS) fuzzy models by using their discretized models. Design/methodology/approach – In this case, a discrete model is obtained from the discretization of the continuous TS fuzzy model. The gains obtained from a non-parallel distributed compens...

This paper deals with clustering of data using different techniques. It firstly presents popular unsupervised clustering algorithms, such as neuro-fuzzy and Fuzzy C-Means (FCM) techniques. Then a hybrid clustering algorithm, starting with the neuro-fuzzy approach and ending with the FCM one, is presented. Secondly, Agglomerative Clustering (AC) and...

Face recognition is one of the most efficient applications of computer authentication and pattern recognition. Therefore it attracts significant attention of researchers. In the past decades, many feature extraction algorithms have been proposed. In this paper Gabor features and Zernike moment were used to extract features from human face images fo...

In this work we are interested in direct fuzzy adaptive control. The continuous SISO non-linear system is presented by the Takgi-Sugeno type fuzzy state model. The strategy of reference model direct adaptive control theory is presented then the adaptive fuzzy adjustment algorithm is presented. The control law includes two terms. The first is respon...

This paper proposes a new hybrid approach for recurrent neural networks (RNN). The basic idea of this approach is to train an input layer by unsupervised learning and an output layer by supervised learning. In this method, the Kohonen algorithm is used for unsupervised learning, and dynamic gradient descent method is used for supervised learning. T...

works, we use the approach based on observers such as the Luenberger observer and the sliding mode observer in order to introduce the diagnosis of nonlinear systems. The robustness of the proposed observers is tested through a physical example. The obtained results show that for non linear systems the performances of sliding mode observer observer...

In this paper we present a system that aims at detecting and tracking moving objects from aerial video streams. These sequences are obtained from a camera mounted on an aerial vehicle which flies over roads and highways. In order to compensate the motion introduced by the dynamic behavior of the camera, we have to estimate geometrical transformatio...

In this paper we present an overview of recent robust approaches which aim 2D visual servoing systems. Since Visual servoing has been considered as a very effective technique for positioning and tracking functionalities, its main goal was to find good selections of visual data in order to be injected in a closed-loop. To guarantee the visual servoi...

Face recognition is becoming a difficult process because of the generally similar shapes of faces and because of the numerous variations between images of the same face. A face recognition system aims at recognizing a face in a manner that is as independent as possible of these image variations. Such variations make face recognition, on the basis o...

In this work, we use the approach based on observers such as the neural observer in order to introduce the diagnosis of nonlinear systems. There are different techniques for training the neural networks. Among these techniques, we quote the backpropagation technique, the backpropagation technique with momentum and the hybrid one which is a mixture...

A novel neural network architecture, is proposed and shown to be useful in approximating the unknown nonlinearities of dynamical systems. In the variable structure neural network, the number of basis functions can be either increased or decreased this is according to specified design strategies so that the network will not overfit or underfit the d...

This paper presents the stability analysis of parameter identification of fuzzy dynamic model. The fuzzy dynamic model is employed to represent discrete time nonlinear dynamic systems. Once the structure is fixed, the parameter identification is accomplished on-line by applying the gradient method. The stability of this algorithm is discussed by us...

The objective of our system is to detect vehicles from aerial sequences. Theses sequences are taken from a camera mounted on UAV which flies over roads and highways. Our approach is to firstly compensate the motion introduced by the dynamic behaviour of the camera. This leads us to a problem of image registration. The moving regions (vehicles) are...

This paper presents new global visual features: random distribution of limited set of pixels luminance. Our approach aims to improve the real-time performance of visual servoing applications. In fact, using these new features, we reduce the computation time of the visual servoing scheme. Our method is based on a random process which ensures efficie...

This paper presents a novel hybrid algorithm for feedforward neural networks, called a self organizing map-based initialization for hybrid training based on a two stage learning approach. First stage, a structure learning scheme which includes adding hidden neurons is used to determine the network size. Second stage, a FN (fuzzy neighborhood)-based...