Browsing by Author "Oukas, Nourredine"

Now showing 1 - 12 of 12

ArabAlg: A new Dataset for Arabic Speech Command Recognition for Machine Learning Applications
(University of Bahrain, 2024) Oukas, Nourredine; Haboussi, Samia; Maiza, Chafik; Benslimane, Nassim
Automatic Speech Recognition (ASR) systems have witnessed significant advancements in recent years, thanks to the emergence of deep learning techniques and the availability of large speech datasets in various languages. With the increasing demand for Arabic voice-enabled technologies, the availability of a high-quality and representative dataset for the Arabic language becomes crucial. This paper presents the development of a new dataset called ArabAlg, specifically designed for Arabic Speech Command Recognition (ASCR), to support the integration of Arabic voice recognition systems into smart devices in the Internet of Things (IoT). This research focuses on collecting and annotating a diverse range of Arabic speech commands, encompassing various domains and applications. The dataset construction process involves recording and preprocessing several utterances from native Arabic speakers. To ensure precision and reliability, quality control measures are implemented during data collection and annotation. The resulting dataset provides a valuable resource for training and evaluating ASCR systems tailored for Arabic speakers using Machine Learning and Deep Learning.
Arabic speech recognition using deep learning and common voice dataset
(IEEE, 2022) Oukas, Nourredine; Zerrouki, Taha; Haboussi, Samia; Djettou, Halima
Speech recognition is critical in creating a natural voice interface for human-to-human communication with modern digital life equipment. Smart homes, vehicles, autonomous devices in the Internet of Things, and others need to recognize various spoken languages. Meanwhile, the Arabic language has a shortage of speech recognition systems. This study comes to develop an Arabic speech-to-text tool for Arabic language. Our solution uses DeepSpeech model which is a deep learning approach and uses a data set from the Common Voice Mozilla project. The results showed a 24.3 percent Word Error Rate and a 17.6 percent character error rate. So, the proposed model reduces the Word Error Rate by 11.7% compared to Bakheet's Wav2Vec model
A colored petri net to model message differences in energy harvesting WSNs
(Springer, 2021) Oukas, Nourredine; Boulif, Menouar
This paper proposes a modelling for Energy Harvesting Wireless Sensor Networks (EHWSNs) by using Coloured Generalized Stochastic Petri Nets (CGSPN). Given that the transmission process consumes the most part of the energy, we consider it in more detail. Therefore, in contrast to the related works in the literature, the proposed formulation differentiates between messages since energy consumption can significantly differ according to the type of information to send. Furthermore, in order to get a more realistic model, we also consider that the sensor has a limited buffer capacity. We conduct some experiments by feeding the model with varying input parameters to show that it can predict the actual behaviour of the network
Evaluating autonomous-energy-harvesting device lifetime for the internet of medical things with a petri net formulation considering battery SoH
(2022) Oukas, Nourredine; Djouabri, Abderrezak; Boulif, Menouar
During charging-discharging operations, the batteries of the Internet of Things (IoT) devices are subject to a depletion that should be considered when predicting their lifetime. This paper proposes a new modeling for the IoT autonomous devices (AD) using Colored Generalized Stochastic Petri Nets (CGSPN). The ADs we consider are equipped with an energy harvesting system, and use a wireless link to connect with their neighbors. The CGSPN formulation models AD functionalities, and evaluates their impact on the battery lifetime by considering its state of health (SoH). The conducted analysis shows the ability of the proposed model to predict the ADs’ lifetime which is very critical for medical applications
A fluid approach to model and assess the energy level of autonomous devices in IoT with solar energy harvesting capability
(IEEE, 2022) Oukas, Nourredine; Djouabri, Abderrezak; Arab, Karima; Helal, Mira
This paper proposes new modeling of autonomous devices in Internet of Things (IoT) using extended Hybrid Petri nets (xHPN). This formulation uses the continuous concept of battery recharge and discharge instead quantification principle. We consider that the autonomous device is equipped with solar energy harvesting (SEH) system and deployed in diverse zones of Algeria with different photovoltaic panel orientations. To conserve energy, we adopt the famous dual sleeping mechanism. The conducted analysis proves that the proposed model is more suitable for the energy assessment of such devices
Generalized stochastic petri nets modelling for energy harvesting WSNs considering neighbors with different vicinity levels
(IEEE, 2020) Oukas, Nourredine; Boulif, Menouar
In this paper, we use Generalized Stochastic PetriNets (GSPN) formalism to model the communication betweenan SN and its neighbors in wireless sensor networks (WSN).This modelling considers several actual considerations such assensor vacations and retrial calls phenomenon. Furthermore,given that sensor nodes (SN) consume almost all their energy inthe transmission process that varies according to the distanceof the neighbors, our model considers different levels of vicinityfor communicating neighbors. Our study proves that ourmodelling, which provides a more realistic approach todescribe the actual behavior of the WSN, can identify the inputparameter scenario to have a network with a good compromisebetween longevity and performance
Modélisation des eh-wsns à capteurs non fiables
(Université M'Hamed Bougara Boumerdes : Faculté des Sciences, 2022) Oukas, Nourredine; Boulif, Menouar(Directeur de thèse)
Les réseaux de capteurs sans ?l (RCSFs) présentent un inconvénient majeur lié à la non ?abilité des capteurs. Ceci est dû à l’épuisement de l’énergie stockée dans leurs petites batteries qui est la cause principale de la majorité des pannes de ces réseaux. A cet effet, l’objectif principal de notre investigation est la détermination de solutions permettant aux RCSFs d’économiser de l’énergie. L’utilisation des énergies renouvelables de l’environnement pour alimenter les capteurs représente une des solutions ef?caces pour remédier à ce problème. Néanmoins, les RCSFs ont besoin d’utiliser d’autres stratégies de conservation d’énergie pour assurer la continuité et la qualité de service, surtout, dans le cas de déploiements à long-terme. En vue de cela, implanter un mécanisme de veille intelligent permet d’augmenter la durée de vie des batteries ; ce qui engendrera des effets positifs sur tout le réseau. L’évaluation des performances des RCSFs par des outils de simulations et/ou de modélisation est nécessaire pour prévoir le comportement du réseau avant son installation réelle. Ceci évitera de transgresser la contrainte du budget alloué. Parmi les outils de modélisation les plus performants, les Réseaux de Petri (RdP) avec leurs différentes extensions permettent à la fois de modéliser et d’évaluer les performances des RCSFs. Dans ce manuscrit, nous proposons une modélisation basée sur les RdPs Stochastiques Généralisés (RdPSGs) et RdPSGs Colorés, en tenant compte diverses circonstances et contraintes de déploiement réel. Ensuite, par une analyse expérimentale et une étude de cas, nous montrons comment utiliser les modèles proposés a?n de trouver le paramétrage qui permet au RCSF d’assurer les performances escomptées
A new generalized stochastic Petri net modeling for energy-harvesting-wireless sensor network assessment
(John Wiley and Sons Inc, 2023) Oukas, Nourredine; Boulif, Menouar; Eric, Campo; Adrien, van den Bossche
This paper proposes an energy-harvesting-aware model that aims to assess the performances of wireless sensor networks. Our model uses generalized stochastic Petri nets to define a sensor–neighbors relationship abstraction. The novelty of the proposed formulation is taking into account several real-life considerations such as battery-over breakdowns, unavailability of neighbors, retrial attempts, and sleeping mechanism in a single model. We use TimeNet tool to simulate the network behavior in order to evaluate its performance throughout different formulas after it had reached its steady state. Finally, we present a case study featuring the different solar energy recovery capabilities of the vast Algerian territory. The aim is to show with the presented model how to determine the kind of resources to be acquired in order to cope with the sensor deployment project requirements. The proposed model allows us to ensure that the battery energy level of sensors deployed in Algiers province for example is almost equal to 80% for 100 messages per day and (1 min/2 min) for (awakening time/sleeping time) ratio
A new petri nets for wsns to model the behaviour of solar-energy harvesting sensors with double sleeping strategy
(IEEE, 2022) Oukas, Nourredine; Boulif, Menouar; Hadiouche, Hadda; Bengharabi, Chaima
This paper introduces a new Generalized Stochastic Petri Nets modelling for sensor nodes (SNs) in wireless sensor networks (WSNs). All SNs are equipped with a solar Energy Harvesting (EH) system. Because there is no solar energy source at night, this formulation considers the case where SNs are intelligently configured to use a double sleeping strategy to get a trade-off between energy conservation and good performances. To achieve this aim, the SN joins the standby state most of the time in the night. In the day, the SN awakes most of the time to get good performances thanks to EH. From the proposed modelling, we derive performance formulas that allow determining the most suitable compromise between energy consumption reduction and rapidity of service
A new stochastic petri nets modeling for dual cluster heads configuration in Energy-Harvesting WSNs
(IEEE, 2021) Oukas, Nourredine; Boulif, Menouar
This paper proposes a new Stochastic Petri Nets modeling to describe the route taken by the packets to reach the base station from any sensor node in wireless sensor networks. This formulation examines the case where the network is structured into clusters. Each cluster contains two leaders: A Cluster Head and a Collector that cooperate to route the packets from the source to the endpoint. This configuration aims to conserve energy by balancing it through the network. Furthermore, given that a sensor node consumes the majority of its power in the communication process that is affected by the distance of the recipient, this formulation associates data gathering and data processing to the Collector whereas it associates the far sending task to the cluster head. From the proposed formulation, we derive performance formulas and we conduct some experimental analysis that allows to determine the most suitable compromise between energy consumption reduction and longevity of service
A novel fluid-based modeling approach using extended Hybrid Petri nets for power consumption monitoring in wireless autonomous IoT devices, with energy harvesting capability and triple sleeping strategy
(Springer Nature, 2024) Oukas, Nourredine; Boulif, Menouar; Arab, Karima
This paper presents a novel approach to model and monitor the energy dynamics of smart devices within the context of the Internet of Things (IoT). The proposed approach employs eXtended Hybrid Petri nets (xHPN) to emulate the behavior of interconnected smart devices forming a wireless network. The novelty of this study lies in the utilization of a fluidic representation to model the battery behavior of smart devices, allowing for the simulation of continuous energy consumption and replenishment via renewable energy harvesting to reflect real-world scenarios. Furthermore, in order to conserve energy, we introduce a new sleeping mechanism named the Triple Sleeping Strategy (TSS). By considering the mean battery charge and the mean sleeping percentage as evaluation metrics, the experimental study showcases the predictive capabilities of the developed model in simulating the performance of IoT networks prior to their actual deployment. Comparative analysis against recent works that use simple and double sleeping strategies, demonstrates the benefits of our approach, in terms of energy efficiency and device lifespan. For instance, when the device is configured with a 90 % sleeping percentage, TSS maintains a decent mean battery level for ten days, almost 8% higher than the double sleeping strategy. Furthermore, the presented case study demonstrates the ability of the proposed model to select appropriate parameters and configurations such as solar panel area and position, battery capacity, packet length, and deployment zone, to cope with the desired performance criteria.
Sensor performance evaluation for Long-Lasting EH-WSNs by GSPN formulation, considering seasonal sunshine levels and dual standby strategy
(Springer, 2023) Oukas, Nourredine; Boulif, Menouar
In this paper, we propose a generalized stochastic Petri net (GSPN) to model the sensor node (SN) interaction in solar energy harvesting wireless sensor networks. Our GSPN formulation models the energy stored in the SN battery by using the quantization principle. Furthermore, it considers that the sensors’ deployment territory is subject to different sunshine levels. In addition, each SN adopts the double sleeping strategy to cope with solar energy shortage in the nighttime. We conduct a comparative analysis between the proposed model and three others taken from the literature to identify which is best suited to predicting the input parameters that extend the network lifetime to fulfill its long-lasting duty with decent performances