Liste des publications
A Survey on Lightweight CNN-Based Object Detection Algorithms for Platforms with Limited Computational Resources
Autonomous drones must be able to identify the existence of one or more objects of interest in a complex environment with high accuracy and speed to fly around safely. Most existing object detection techniques, based on traditional machine learning algorithms, can't offer acceptable performance in complicated environments. Deep Convolutional Neural Networks (CNNs) provide us such ability with high performance. Today, deep CNN-based object detection algorithms are more and more used in Artificial Intelligence (AI) applications. However, it still very difficult to deploy large CNNs architectures on small devices with limited hardware resources, because they consist of millions of parameters, which make them computationally very exhausting. Lightweight CNN architectures are proposed as a solution to make the deployment of deep neural networks on small devices feasible. This paper focuses on reviewing recent used lightweight CNN architectures that can be implemented on embedded targets to improve the object detection performance for small devices-based systems, like drones. We need to select fast and lightweight CNN models to use them on drone platforms. The purpose of this reviewing is to choose the most accurate and fastest algorithm to implement it on our drones. Voir les détails
Mots clés : Computer vision, Deep Learning, Object Detection, Convolutional Neural Network, lightweight CNN
A Copper Oxide (CuO) Thin Films Deposited by Spray Pyrolysis Method
A simple and low-cost procedure (spray pyrolysis) was used to elaborate a copper oxide thin films on ordinary glass substrates. A copper nitrate was used and dissolved in two different solutions (Water, Methanol) S1 and S2 respectively in order to obtain an equal concentration; CS1,S2=0.5M. The spray pyrolysis deposition made at fixed temperature T=500°C and different volumes of S1 or S2 on the glass substrates. The X-ray diffraction (XRD), scanning electron microscopy (SEM) and UV-vis spectrophotometry were used to determinate the structural, morphological and optical properties of CuO thin films. The X-ray diffraction patterns confirm the presence of the polycrystalline phase of CuO as monoclinic crystal structure with preferential orientation along (110), (002), (111), (200) and (020). Their optical band gaps ranged from 3.95 to 4.02eV for thin films made with S1, and from 1.6 to 1.95eV for thin films made with S2 with a high absorbency in the visible region, which is in agreement with the values of the literature. Voir les détails
Mots clés : Copper oxide, Thin films, Spray pyrolysis, Band Gaps, XRD, SEM.
Improved Cu2O/AZO Heterojunction by Inserting a Thin ZnO Interlayer Grown by Pulsed Laser Deposition
Cu2O/ZnO:Al (AZO) and Cu2O/ZnO/AZO heterojunctions have been deposited on glass substrates by a unique three-step pulsed laser deposition process. The structural, optical, and electrical properties of the oxide films were investigated before their implementation in the final device. X-ray diffraction analysis indicated that the materials were highly crystallized along the c-axis. All films were highly transparent in the visible region with enhanced electrical properties. Atomic force and scanning electron microscopies showed that the insertion of a ZnO layer between the Cu2O and AZO films in the heterojunction enhanced the average grain size and surface roughness. The heterojunctions exhibited remarkable diode behavior and good rectifying character with low leakage current under reverse bias. The presence of the ZnO interlayer film significantly reduced the parasitic and leakage currents across the barrier, improved the quality of the heterostructure, made the energy band between AZO and Cu2O layers smoother, and eliminated the possibility of interface recombination, leading to much longer electron lifetime. Voir les détails
Mots clés : Heterojunctions, ZnO, Cu2O, AZO, pulsed laser deposition, solar cells
Optimal pump excitation frequency for improvement of damage detection by nonlinear vibro acoustic modulation method in a multiple scattering sample
We present a method to systematically optimize nonlinear damage detection in multiple scattering media by the nonlinear Vibro-Acoustic Modulation (VAM) technique. The latter consists here of exciting a medium simultaneously with a high frequency ultrasonic sinusoidal burst and with a low frequency continuous sinusoidal wave. Modulation of the high frequency (probe) by the low frequency (pump) is made possible by the presence of nonlinear scatterers, i.e. cracks, defects. A signal processing technique consisting of a closed loop system drives the automatic adaptation of the pumping frequency, yielding to the optimization of the nonlinear modulation (NM) of the output probing coda signal without a priori information on the medium and the scatterers. The correlation coefficient between a reference output probe signal without the pumping wave and an output modulated probe signal with a pumping wave was considered as our cost function. A multiple scattering solid beam where nonlinear scatterers can be controllably added or removed is designed and tested. The first step of this study is an empirical search of the correlation coefficient dependency on the pumping frequency to verify the performances of the proposed method. Then the implemented optimization algorithm based on genetic algorithm (GA) is used to find automatically the optimal pumping frequency. The obtained optimization results show a good agreement with the empirical study. Moreover, the genetic algorithm allowed to find the optimal pump frequency adapted to each configuration of nonlinear scatterers. This relatively fast search of the optimal nonlinear response could be extended to nonlinear scatterer imaging applications using the information on the resonant modes spatial shapes together with the associated optimal response. Voir les détails
Mots clés : Optimal command, Genetic algorithm, Nonlinear- Vibro Acoustic Modulation, Crack detection
Optimal pump excitation frequency for improvement of damage detection by nonlinear vibro acoustic modulation method in a multiple scattering sample.
We present a method to systematically optimize nonlinear damage detection in multiple scattering media by the nonlinear Vibro-Acoustic Modulation (VAM) technique. The latter consists here of exciting a medium simultaneously with a high frequency ultrasonic sinusoidal burst and with a low frequency continuous sinusoidal wave. Modulation of the high frequency (probe) by the low frequency (pump) is made possible by the presence of nonlinear scatterers, i.e. cracks, defects. A signal processing technique consisting of a closed loop system drives the automatic adaptation of the pumping frequency, yielding to the optimization of the nonlinear modulation (NM) of the output probing coda signal without a priori information on the medium and the scatterers. The correlation coefficient between a reference output probe signal without the pumping wave and an output modulated probe signal with a pumping wave was considered as our cost function. A multiple scattering solid beam where nonlinear scatterers can be controllably added or removed is designed and tested. The first step of this study is an empirical search of the correlation coefficient dependency on the pumping frequency to verify the performances of the proposed method. Then the implemented optimization algorithm based on genetic algorithm (GA) is used to find automatically the optimal pumping frequency. The obtained optimization results show a good agreement with the empirical study. Moreover, the genetic algorithm allowed to find the optimal pump frequency adapted to each configuration of nonlinear scatterers. This relatively fast search of the optimal nonlinear response could be extended to nonlinear scatterer imaging applications using the information on the resonant modes spatial shapes together with the associated optimal response. Voir les détails
Mots clés : Optimal command, Genetic algorithm, Nonlinear Vibro Acoustic Modulation, Crack detection
Early stages of copper microparticles electrodeposition on polypyrrole film
In this work, we studied the electrodeposition of copper (Cu) microparticles on polypyrrole (PPy) ?lms using cyclic voltammetry and chronoamperometry techniques. The initial stages of Cu deposition were investigated by performing current transients. Models based on Scharifker and Hills calculations were established to determine the nucleation and growth type. The results suggest that the deposition of Cu proceeds via an instantaneous nucleation followed by three-dimensional (3D) di?usion-limited growth. The values of the number density of active sites N∞ and di?usion coe?cient D were also determined. Voir les détails
Mots clés : composites, copper, electrodeposition, polymer, thin film
IDENTIFICATION OF THERMAL AND MICROSTRUCTURAL PROPERTIES OF HOT ROLLING SCALE
A significant amount of scale is produced during casting of ingots and processing of hot-rolled products. In manufacturing steel, during the various rolling operations, the amount of scale produced is approximately 0.1% of the annual production of the rolling mills. The quality of the thin sheet during rolling is affected by the behavior of the iron oxide layers formed on their surfaces. For this reason, acids and oils are used for the descaling of slabs and billets by means of pressurized water. The calamine, contaminated by these various acids and used oils, is rejected and stored involuntarily on important areas and pollutes soil and groundwater. Micrographic observations as well as X-ray diffraction analysis have shown that calamine consists mainly of iron oxides. Hematite and magnetite become the main components for oxidation times greater than 1 hour. Characterization tests have shown that calamine is dense (ρ = 4.8 g/cm3), its particle size is variable depending on the degree of oxidation (from 0.5 to 10 mm). Simultaneous thermal analysis showed that an increase in mass of the calamine sample with a release of heat. Studies are underway for the physico-chemical characterization of the soils of the storage areas. Voir les détails
Mots clés : Scale, Risks, soil, thermal analysis, granulometry, X-ray diffraction
Diagnosis Method for GTO Open Switch FaultApplied to Reconfigurable Three-Level 48-PulseSTATCOM
In the recent years, several research worksare focusing on the use of STATCOM in electrical net-works because it is used to regulate the voltage, to im-prove the dynamic stability of the power system be-sides allowing better management of the power ?ow.All these positive tasks have guaranteed an importantposition of STATCOM within a family of Flexible Al-ternating Current Transmission System (FACTS). Inthis paper study, the control and operation of a threelevels 48-pulse GTO based STATCOM is implementedwith series connected transformers. The system may,unfortunately, be prone to GTO switch faults and there-fore may a?ect reactive power transiting. In this pa-per, a new diagnostic approach is proposed based onthe Single-Sided Amplitude Spectrum (SSAS) methodof the three-leg converter currents for detection and lo-calization of open-circuit faults. The integration of theSTATCOM recon?gurable fault tolerant to the systemis also considered to ensure service continuity. Sev-eral results are presented and discussed in this paperto illustrate the performance of the STATCOM fault-tolerant diagnostic. Voir les détails
Mots clés : Detection, diagnosis, FACTS, GTO, open switch fault, recon?guration, SSAS, STATCOM
CONTRIBUTION À LA RÉCONFIGURATION D’ONDULEUR ÀTOLÉRANCE DE PANNE APPLIQUÉE AU PARC ÉOLIEN CONNECTÉAU RÉSEAU ÉLECTRIQUE
Cet article propose une nouvelle technique de diagnostic basée sur les vecteurs de Park associés aux coordonnées polaires pour ladétection et la localisation des défauts de circuit- ouvert (CO), ainsi que l’intégration de l’onduleur reconfigurable à tolérance depanne afin de permettre la continuité de service du système du parc éolien. Le modèle utilisé est celui d'une éolienne à vitessevariable couplée à une génératrice asynchrone à double alimentation (GADA) connecté au réseau électrique via unconvertisseur à tolérance de panne pour améliorer les performances après l’apparition du défaut. Ce convertisseur est utilisé àbase d’IGBT pour l’obtention d’une précision acceptable avec des fréquences de commutation élevées. Ce modèle est bienadapté pour l'observation des harmoniques et la performance dynamique du système de contrôle sur des périodes de tempsrelativement courtes (généralement des centaines de millisecondes à une seconde). Cette méthode permet d'extraire l'énergiemaximale du vent lors d’une faible vitesse du vent en optimisant la vitesse de la turbine tout en minimisant les contraintesmécaniques sur la turbine pendant les rafales de vent. Pour illustrer l’amélioration de diagnostic des défauts en circuit ouvert duconvertisseur à tolérance de panne, plusieurs résultats sont présentés et discutés dans cet article. Voir les détails
Mots clés : Génératrice asynchrone à double alimentation (GADA) Turbine, Onduleur, diagnostic, détection, localisation, Reconfiguration, défaut, Circuit- ouvert, Insulated gate bipolar transistor (IGBT).
Rolling Bearing Fault Diagnosis Based on an Improved Denoising Method Using the Complete Ensemble Empirical Mode Decomposition and the Optimized Thresholding Operation
Vibration signals are widely used in monitoring and diagnosing of rolling bearing faults. These signals are usually noisy and masked by other sources, which may therefore result in loss of information about the faults. This paper proposes an improved denoising method in order to enhance the sensitivity of kurtosis and the envelope spectrum for early detection of rolling bearing faults. The proposed method is based on a complete ensemble empirical mode decomposition with an adaptive noise (CEEMDAN) associated with an optimized thresholding operation. First, the CEEMDAN is applied to the vibration signals to obtain a series of functions called the intrinsic mode functions (IMFs). Second, an approach based on the energy content of each mode and the white noise characteristic is proposed to determine the trip point in order to select the relevant modes. By comparing the average energy of all the unselected IMFs with the energy of each selected IMFs, the singular IMFs are selected. Third, an optimized thresholding operation is applied to the singular IMFs. Finally, the kurtosis and the envelope spectrum are used to test the effectiveness of the proposed method. Different experimental data of the Case Western Reserve University Bearing Data Center are used to validate the effectiveness of the proposed method. The obtained experimental results illustrate well the merits of the proposed method for the diagnosis and detection of rolling bearing faults compared to those of the conventional method. Voir les détails
Mots clés : Vibration analysis, bearing Fault diagnosis, CEEMDAN, Denoising, thresholding operation, envelope, Kurtosis