Advances in Computing Systems and Applications

Cyber-Physical Systems refer to the tight integration and coordination between computational and physical resources. Modeling their behavior requires handling continuous and discrete behaviors. The definition of the associated models refers to continuous and discrete systems theories. In this talk, we address the problem of designing correct software to control cyber-physical systems. We recall the necessary basic concepts allowing a designer to model such systems. We also give an overview of the different formal approaches supporting the formal verification of these hybrid systems and we highlight the results obtained using these techniques. In particular, we focus on the use of proof and refinement based methods.

One of the most promising technologies to enable the future Internet of Things is the Time Slotted Channel Hopping (TSCH) mode, which reveals the robustness of the IEEE 802.15.4 standard by providing high reliability, low latency, and energy efficiency. TSCH has received a lot of attention from the researchers’ community. In fact, the TSCH specification had never provided how to build and maintain a schedule; therefore, several researchers devised new scheduling algorithms. This paper focuses on scheduling in TSCH networks. It starts with a gentle introduction to the IEEE 802.15.4 standard and the TSCH mode. It then surveys the State-of-The-Art scheduling algorithms, where algorithms are classified and compared. Each of the algorithms are then presented along with its pros and cons. Finally, weaknesses that need to be addressed are identified.

Cloud computing plays an important role in the improvement of the Information Technology (IT) industry. However, the cloud Quality of service (QoS) level is considered the biggest challenge facing cloud providers and a major concern for enterprises today. That is why, resource allocation is the optimum solution towards this end, which could be done by the best use of task scheduling techniques. In this paper, we provide a literature analysis of the resource allocation in Cloud computing. As well, we propose a classification of tasks scheduling approaches used in the cloud. Furthermore, our work helps researchers develop existing approaches or suggest a new method in this area.

Nowadays, large amounts of heterogeneous data are generated continuously from social media, connected devices, and digital process, etc. To extract useful information and insight from these huge data, new management, and processing systems that support the high Volume, high Variety, and high Velocity of data are required. Cloud Computing provides suitable infrastructures to maintain big data processing and storage. However, outsourcing storage and computation of data to Cloud servers make security a big concern. This paper studies the security issues of big data processing in Cloud environment and schemes that are proposed to ensure its privacy, integrity, and availability in order to invasion future research directions.

Resource discovery in the Internet of Things (IoT) provides information about the available sensors, and actuators as well as the services and resources they provide. This task, crucial for IoT applications evolution, allows self-configuration and context awareness. When combined with semantic technologies, resource discovery increases the interoperability and backs the reasoning process. However, introducing semantics into the IoT incurs challenges regarding resources and power consumption as well as the compactness of representations and resource lookup efficiency. In this paper, we discuss such challenges and propose an architecture for including semantics in a scalable IoT system. Next, we present a mechanism aiming to increase semantic resource discovery efficiency in heterogeneous IoT systems. The proposed mechanism is based on the construction of a concept directory to accelerate semantic-based resource lookup. By using such a directory, we take benefits of semantic representations to better match user requests, without lowering discovery task efficiency. The feasibility and effectiveness of our solution was demonstrated by experimental evaluations.

The IoT development area has drawn the attention of nowadays researchers, some of them made assumptions regarding the use of clustering in their key management schemes. For example, in CL-EKM (Certificateless Effective Key Management) protocol, cluster-heads are assumed to be with high-processing capabilities and deployed within a grid topology. In fact, this is only possible in a controlled environment. In a hostile environment, such as battlefields, this assumption cannot be satisfied. In this work, an enhancement of the CL-EKM scheme has been proposed by introducing a distributed clustering algorithm. The performance of the implemented and enhanced system proved our assumptions.

Internet of Things’ (IoT) services represent a great interest research topic for both academic and industrial communities due to their large application domains. New dynamic and automatic development approaches are created in order to improve their effectiveness based on service composition process. Indeed, they allow the aggregation of smart object’s services to meet complex requirements from various application areas. This paper provides an architecture that describes a framework for service composition in IoT. Moreover, it proposes a rule-based reasoner where several facts and rules (knowledge base) are inferred to satisfy user queries. This approach is designed and implemented using SysML and Prolog platform respectively. The use-cases scenario and extensive tests show clearly the interest, feasibility, and suitability of the rule-based system for service composition in the IoT.

The operation of the wireless sensor networks (WSNs) with limited energy resources is an important challenge for researchers. In fact, many schemes have been proposed to save the wasted energy in a WSN or to equip its sensor nodes with an energy harvesting system. In this paper, we propose a preamble sampling scheme based MAC protocol for energy harvesting WSNs (PS-EHWSN). PS-EHWSN leverages the advantage of transmitter-initiated schemes and uses the low power listening (LPL) technique with short preamble messages where each sensor node in the network can determine its next sleep period to reduce the duty-cycle by the use of its residual energy, which can increase over time. Moreover, our proposition promotes even energy consumption of the sensor nodes with the energy harvesting capability. PS-EHWSN protocol was simulated using OMNeT++/MiXiM. The simulation results show that our proposed protocol exceeds some existing MAC protocols such as the BMAC.

The Elliptic Curve Diffie-Hellman (ECDH) is the basic protocol used for key agreement on Elliptic Curves (EC) and it is analogue to the standard Diffie-Hellman key exchange. The security of ECDH relies on the difficulty of the Elliptic Curve Discrete Logarithm Problem (ECDLP), however this protocol is vulnerable to man in the middle attack. In this paper, we first analyze the Ahirwal and Ahke encryption scheme which is based on ECDH key exchange and then we propose a new key agreement method to secure it from man in the middle attack.

Information systems are strongly impacted by new information technologies. New challenges emerge in the scientific and technological research caused by the craze around data and more particularly around big data. The data poses many challenges related to its exploitation. Its scope goes beyond traditional structures, such as databases or data warehouses. Furthermore, it is mostly unstructured which results in a need for new approaches to explore it. The different existing processes need to be redesigned. Their evolution poses new scientific obstacles as soon as they are projected in a big data framework. The Data Lakes represent today an emerging concept how to organize around the data to rethink innovation cycles within companies. On the other hand, this concept opens up new issues of investigation and promises real challenges, which will allow information technologies to evolve towards new perspectives in the professional world, and calls on new skills to develop and to capitalize on future users. What is it really about?

Text auto-completion speeds up user input on the desktop PC by proposing a list of alternative words to select from after typing few characters. This useful feature is even more important on mobile devices because their human-computer-interface is restricted. We find it very helpful to extend auto-completion to the Holy Quran text, by allowing the user to select from alternative Ayat () where the characters he typed appear, instead of typing the Aya () he wants to cite manually or copying it from another source. To achieve this goal, we have implemented Quran text auto-completion on the Android platform by extending an open-source software keyboard application. As this feature needs to search quickly and repetitively the whole Quran text, we tested a number of string search methods, including the Boyer Moore algorithm, to determine the fastest search solution to use. The result is an application that proposes a fast and smooth user experience that is available freely in the play store under the name ().

This paper presents the software implementation of Weil, Tate, Ate and Optimal Ate pairings in Jacobean coordinates, over Barreto-Naehrig curve, on Virtex-5 using the MicroBlaze software processor and the ZedBoard Zynq-7000 platform using ARM hardcore processor. The most pairing functions are constructed on the same model, one execution of the Miller’s algorithm plus a final exponentiation, which can be programed with addition chain method. Our flexible system can be performed for any curve parameters.

Context-based adaptive binary arithmetic coding (CABAC) is specified as the single operation mode for entropy coding in the newest standard High Efficiency Video Coding (HEVC). While it provides high coding efficiency, the data dependencies in H.265/HEVC CABAC make it challenging to parallelize and thus, limit its throughput. This paper proposes a multi-bins CABAC decoder architecture adaptive to HEVC syntax elements with small FSM (finite state machine) for the control of SE order. In order to reduce the critical path delay, we exploit different techniques of optimization such as a speculative decoding, logic balancing techniques and our proposed technique RLpsLZpStateIdx LUT. The parallel implementation can process 1,34 bin/cycle when operate at 134,23 MHz and improved high throughput of 179.86 Mbin/s with an optimized path delay compared to the serial process. The architecture is coded using VHDL ISE language, simulated and synthesized using Xilinx tools with virtex4 xc4vsx25-12ff668 board.

In practice, a large amount of Software Product Lines (SPLs) are developed using a bottom-up process. In this case, an SPL is synthesized from similar product variants that are developed for SPL using ad hoc reuse techniques such as copy-paste-modify. In this paper, we present an approach for migrating existing product variants into an SPL. This approach is applied on models and it takes as input a set of models that abstract the product variants. The result of the approach is a software product line represented by the SPL model and the variability model. SPL model is the result of merging input product models. The variability model is a Feature Model (FM) allowing the specification of the variability on the SPL model. We propose to construct the SPL in an incremental way. After an initialization step, the set of input products are integrated in the SPL one after another. To integrate a new product, we first compare the input product model with the SPL model in order to identify the variability, and then we update both the SPL model and the variability model. The approach is implemented and evaluated on a case study.

Social networks are considered today as the most popular interactive media where people can communicate, share information and express opinions without any limitation. The interest of the scientific community towards social contents has increased due to their importance in various fields such as marketing, sociology and politics. Several research areas related to social networks have emerged namely, community detection, sentiment analysis and topic detection. In this paper, we propose a domain-level topic detection approach for improving sentiment analysis in Arabic social content. The proposed approach is based on a supervised learning technique on Arabic collected data. Training dataset is mainly composed of Arabic press articles, while the test dataset is represented by posts and comments extracted from Arabic Facebook pages. Experimental evaluation showed that the proposed approach achieves good performances with precision values between 75.36% and 97.89%.

Over the last decade, social media have dominated our lives. The exploding number of data produced by these platforms triggered a wave of research works that mainly focus on the storage and analysis of this data. In this paper, we propose an original information warehouse architecture for the storage and analysis of social media information. A multidimensional model is defined and the information is extracted, transformed and loaded in the warehouse using ETL (Extract, Transform, Load). The described framework is implemented for Twitter and a data mining analysis is performed on the collected tweets using a clustering algorithm to uncover most discussed topics. The preliminary results are satisfactory and the proposed paradigm can be applied for various information sources such as newspapers and scientific papers.

The high way of presentation of entities or activities and relations which enterprises needed for their higher success and the powerful descriptive and representation method because of their complexity needed more and more in our days. In this paper, an approach for Business Process Modelling (BPM) using Hypergraphs representation based on Finite State Machines (FSM) will be presented.The main goal motives this work is the using of hypergraphs as a flexible mathematical structure describing Information System (IS) from various viewpoints to express the best graphical way for Business Process (BP).The model proposed use FSM as an automated formal model by defining concepts and components of the hypergraph elements based on different FSM patterns (simple, complex) to get representations can be analyzed by either using more traditional tools as logic and inference rules or by a set of tools belonging to data science later.

The surface electromyogram (SEMG) is an electrophysiology signal that can be used in many fields such as biomechanical engineering, medicine and sport. It’s applied to analyze and study the human movement. The SEMG recordings can be contaminated by spurious background spikes, quiescent baseline. These artifacts and noises produce false muscle activation detection. The muscle activation detection during movement depends on several parameters such as the beginning and the end of an activity, the nerve conduction velocity, the on-off interval, etc. In this paper, we conduct a study to detect the activation interval from the biceps brachial muscle using discrete wavelet transform (DWT) for SEMG signal denoising based on thresholding method. We compare our method with the integrated profile method. The results show that our method can effectively reduce the detection error.

The talk will highlight the general idea of Virtual and Augmented Reality (V&AR), from design specifications to evaluation methods. After a short introduction into the history of V&AR, we will have a look at various types and concepts of V&AR known today. We then discuss some experimental cases to highlight the need of specification of all V&AR parts, such as technical and hardware constraints but also those concerning tasks modeling and users needs. We will present some criteria we consider when evaluating a V&AR experiences success. Finally, some Algerian V&AR applications and projects will also be presented in order to show national developments in this area.

Textures such as grass, trees, mountains, buildings and others occupy large spaces of our visual environment. Numerous researches have been devoted to automatically analyze and characterize textures, where static textures found in single images were the first to be studied. Subsequently, this notion was extended to temporal dimension, known as dynamic texture representing variable properties in time such as flames, swaying trees, moving clouds, crowds in public places and even shadows, etc. Lately, temporal texture research is gaining a lot of attention, due to its importance as an effective component in the interpretation of video content.This paper presents a research survey that focuses on a very captivating subject: Dynamic texture analysis, characterization and recognition. Its motivation is to give an overview of the most up-to-date analysis approaches that have been proposed to characterize then recognize temporal textures in different fields.

This paper describes an approach that combines the navigation data given by a Doppler Velocity Logs (DVL), the MTi Motion Reference Unit (MRU) and a Mechanically Scanned Imaging Sonar (MSIS) as a principal sensor to efficiently solve underwater Simultaneous Localization and Mapping (SLAM) problem in structured environments such as marine platforms, harbors, or dams, etc. The MSIS has been chosen of its capacity to produce a rich representation of the environment. In recent years, to solve the SLAM Autonomous Underwater Vehicle (AUV) problem, very few solutions have been proposed. Our contribution has introduced a method based on the Nonlinear H-infinity filter $$(NH\infty )$$(NH∞) to solve the SLAM-AUV problem. In this work, the $$NH\infty $$NH∞-SLAM algorithm is implemented to construct a map in partially structured environments and localize the AUV within this map. The data-set used in this paper are taken from the experiments carried out in a marina located in the Costa Brava (Spain) with the Ictineu AUV which is necessary to test different SLAM algorithms. The validation of the proposed algorithm through simulation in offline is presented and compared to the EKF-SLAM algorithm. The $$NH\infty $$NH∞-SLAM algorithm provides an accurate estimate than EKF-SLAM and good results were obtained.

Many efforts are devoted to develop binary descriptors due to their low complexity, and flexibility in case of embedded systems. Almost all works on binary descriptor conception didn’t exploit all information of a given patch; they just involved pixels intensities into binary test process. This kind of solution lack efficiency on patch description. In this paper, we propose to design a new descriptor based on 3D polynomial interpolation by used pixels intensities. We must take into account geometric positions of pixels. We suggest to divide the patch into equal grid cells (sub patches). Each sub patch undergoes a dimension augmentation. It becomes a 3-dimensional vector by considering intensities values as the third dimension. Based on 3D polynomial interpolation, we approximate the point cloud by a surface. This step is followed by a binary tests between all coefficients of polynomials situated in neighborhoods. Our method shows a considerable discrimination in case of high similarity. The results of our approach are evaluated on a well-known benchmark dataset exhibit a considerable robustness and reliability in front of severe changes. A computation costing is reported in the end of results section.

This paper presents a new hybrid approach for 3D face verification based on tensor representation in the presence of illuminations, expressions and occlusion variations. Depth face images are divided into sub-region and the Multi-Scale Local Binarised Statistical Image Features (MSBSIF) histogram are extracted from each sub-region and arranged as a third order tensor. Furthermore, to reduce the dimensionality of this tensor data, we use a novel hybrid approach based on two steps of dimensionality reduction multilinear and non-linear. Firstly, Multilinear Principal Component Analysis (MPCA) is used. MPCA projects the input tensor in a new lower subspace in which the dimension of each tensor mode is reduced. After that, the non-linear Exponential Discriminant Analysis (EDA) is used to discriminate the faces of different persons. Finally, the matching is performed using distance measurement. The proposed approach (MPCA+EDA) has been tested on the challenging face database Bosporus 3D and the experimental results show that our method achieves a high verification performance compared with the state of the art.

This paper focuses on the evaluation of noisy image fusion for medical images obtained from different modalities. In general, medical images suffer from poor contrast and are corrupted by blur and noise due to the imperfection of image capturing devices. In order to improve the visual and quantitative quality of the fused image, we compare two algorithms with other fusion techniques. The first algorithm is based on Dual Tree Complex Wavelet Transform (DTCWT) while the second is based on Scale Mixing Complex Wavelet Transform (SMCWT). The tested algorithms are using different fusion rules in each one, which leads to a perfect reconstruction of the output (fused image), this combination will create a new method which exploits the advantages of each method separately. DTCWT presents a good directionality since it considers the edge information in six directions and provide approximate shift invariant as well as SM-CWT, the goal of PCA is to extract the most significant features (wavelet coefficients in our case) to improve the spatial resolution. We compared the tested methods visually and quantitatively to recent fusion methods presented in the literature over several sets of medical images at multiple levels of noise. Further, the tested fusion algorithms have been tested up to the important level of Gaussian, salt & pepper and speckle noise (350 test). For the quantitative quality, we used several well-known fusion metrics. The results show that the tested methods outperform each method individually and other algorithms proposed in the literature.

The main aim behind this study relates to comparing a variety of motion detection methods for a mono static camera and it endeavors to identify the best method for different environments and complex backgrounds whether in indoor or outdoor scenes. For this reason, we used the CDnet 2012 video dataset as a benchmark that comprises numerous challenging problems, ranging from basic simple scenes to complex scenes affected by shadows and dynamic backgrounds. Eleven ranging from simple to complex motion detection methods are tested and several performance metrics are used to achieve a precise evaluation of the results. This study sets its objective to enable any user to identify the best method that suits his/her need.

Due to the important development in Hardware technologies, sensors are everywhere. So, many real-world applications in different domains (such as Defense, Industry, Transport, Energy, Surveillance, Climate and Weather, Healthcare) use multi-sensors to collect tremendous amounts of data about their states and environments. Such data are inherently uncertain, erroneous and noisy on the one hand, and voluminous, distributed and continuous on the other hand. One of the major challenges the Governments, Industry, Companies and Organizations have to face today is how to manage and make sense of Big sensor data for the purpose of decision making. Recent advancements in science and technologies (like Computational Intelligence and Machine Learning) are opening the road to more advanced analytics techniques that can allow for the sensor data characteristics and extract useful insights. This allows building solutions that provide fast time responses and less resources consuming. In this talk, we show how techniques stemming from the recent Computational Intelligence field can contribute to the above solutions to manage and handle Sensor data. Some examples from the aeronautic/space domain are used to motivate our propositions.

A number of conventional clustering algorithms suffer from poor scalability, especially for data lake. Thus many modified clustering algorithms have been proposed to speed up these conventional algorithms based on the employment of data sampling techniques. However, these representations require the number of clusters to proceed to centroid selection for final data clustering. To address this limitation, this paper develops a two-phase clustering-based methodology. In the first phase, rather than attempting to construct a random sampling, we define a novel approach that computes plausible sample points, uses them as centroids for the final clusters. To speedup our clustering algorithm in the second phase we propose a parallelization scheme in conjunction with a Spark parallel processing implementation. Computational experiments reveal that the Global sampling method is more effective in terms of both quality and stability compared to the most popular K-means algorithm for the same parameter settings.

The Bees Algorithm (BA) is one of the most recent swarm-based meta-heuristic algorithms that mimic the natural foraging behavior of honey bees in order to solve optimization problems and find the optimal solution. Clustering analysis, used in various science fields and applications, is an important tool and a descriptive process attempting to identify similar classes of objects based on the values of their attributes. To solve clustering problems there are diverse ways, including machine learning techniques, statistics, and metaheuristic methods. In this work, an improved Bees Algorithm with memory scheme (BAMS), which is a modified version of the BA algorithm, is used for data clustering. In the BAMS algorithm, a simple memory scheme is introduced to prevent visiting sites which are close to previously visited sites and to avoid visiting sites with the same fitness or worse. Four real-life data sets are applied to validate the proposed algorithm, and results of this study are compared to BA and others state-of-the-art methods. The experimental results show that the proposed algorithm outperforms other methods.

We address load balancing and data locality problems in Hadoop. These two problems limit its performance, especially, during a reduce phase where the partitioning function assigns the keys to the reducers based on a hash function. We propose in this paper a new approach to assign the keys based on the reducers’ processing capability in order to ensure a good load balancing. In addition, our proposed approach called RTSBL takes into consideration the data locality during the partition. Our experiments prove that RTSBL achieves to up 87% improvements in the load balancing and 3$$\times $$× improvements of the data locality during the reduce phase in the standard Hadoop.

One of the important issues in range partitioning schemes is data skew. Tuples distribution across nodes may be skewed (some nodes have many tuples, while others may have fewer tuples). Processing skewed data not only slows down the response time, but also generates hot nodes. In such a situation, data may need to be moved from the most-loaded partitions to the least-loaded ones in order to achieve storage balancing requirements. Early works from the State-of-The-Art focused on achieving load balancing. However, today’s works focus on reducing the load balancing cost. This latter involves reducing the cost of maintaining partition statistics. In this context, we propose to improve one of the best load balancing work, that is the one of Ganesan et al., to reduce the cost of maintaining the statistics of load balancing. We introduce the concept of fuzzy system image. Both nodes and clients have approximate information about the load distribution. They can nevertheless locate any data with almost the same efficiency as using exact partition statistics. Furthermore, maintaining load distribution statistics do not require exchanging additional messages as opposed to the cost of efficient solutions from the State-of-The-Art (which requires at least $$\mathbb {O}(\log {n})$$O(logn) messages).

In this work, we developed and experimentally validated a novel model for external clustering validation to deal with huge data sets using Conditional Entropy index. The model allows clustering validation in a parallel and a distributed manner using Map-Reduce framework, it is termed MR-Centropy. The aim is to be able to scale with increasing dataset sizes when ground truth clustering is available. The proposed MR-Centropy is a three-jobs process where each job consists of Map and Reduce functions. Three jobs were necessary to gather all the statistics involved in the computation of the Conditional Entropy index. Each step in the proposed framework is done in parallel. Numerical tests on real and synthetic datasets demonstrate the effectiveness of our proposed model.

Social Network Analysis (SNA) is an active research topic. It arises in a broad range of fields. One important issue in SNA is the discovery of key players who are the most influential actors in a social network. Negative Key Player Problem (KPP-NEG) aims at finding the set of actors whose removal will break the social network into fragments. By another way, Multi-Agents Systems (MAS) paradigm suggests suitable ways to design adaptive systems that exhibit desirable properties such as reaction, learning, reasoning and evolution. A fortiori, the intrinsic nature of social networks and the requirements of their analysis could be efficiently handled using a MAS framework. Within this context, this paper proposes a multi-agents based-model AMAM for KPP-NEG. We first represent the social network in terms of a weighted graph. Then, a set of agents cooperate in order to identify the most important nodes. Simulation and computational results are demonstrated to confirm the effectiveness of our approach.

In this paper, an advanced evolving clustering strategy is used to design a fuzzy model-based sensor faults detection mechanism for a pilot parallel-type heat exchanger. The change in the process operating mode is detected by an incremental unsupervised clustering procedure based on participatory learning. Real experimental data is used to construct signals for fuzzy residual generation. The resulting residuals are then processed by the evolving classifier to supervise the heat exchanger operation. The obtained results clearly show the ability of the evolving fuzzy classifier to early detect the considered temperature sensor faults.

Training artificial neural networks (ANNs) is a common hard optimization problem. The process of neural nets training is generally defined on synaptic weights and thresholds of artificial neurons with the aim to find optimal or near-optimal values. Artificial bee colony (ABC) optimization has been successfully applied to several optimization problems, including the optimization of weights and biases of ANNs. This paper addresses the problem of feed-forward ANNs training by using a novel ABC variant named cooperative learning artificial bee colony algorithm (CLABC), which we have developed in our previous work. The performance of the CLABC-trained feed-forward ANN is validated on different classification problems, namely the XOR problem, the 3-bit parity, 4-bit encoder-decoder and Iris benchmark problems. The results are compared to other advanced optimization methods.