Search

This paper investigates the throughput of a non-orthogonal multiple access (NOMA)-based cognitive radio (CR) system with multiple unmanned aerial vehicle (UAV)-assisted relays under system performance and security constraints. We propose a communication protocol that includes an energy harvesting (EH) phase and multiple communication phases. In the EH phase, the multiple UAV relays (URs) harvest energy from a power beacon. In the first communication phase, a secondary transmitter (ST) uses the collected energy to send confidential signals to the first UR using NOMA. Simultaneously, a ground base station communicates with a primary receiver (PR) under interference from the ST. In the subsequent communication phases, the next URs apply the decode-and-forward technique to transmit the signals. In the last communication phase, the Internet of Things destinations (IDs) receive their signals in the presence of an eavesdropper (EAV). Accordingly, the outage probability of the primary network, the throughput of the secondary network, and the leakage probability at the EAV are analyzed. On this basis, we propose a hybrid search method combining particle swarm optimization (PSO) and continuous genetic algorithm (CGA) to optimize the UR configurations and the NOMA power allocation to maximize the throughput of the secondary network under performance and security constraints.

In this paper, the secrecy performance of a cognitive radio networks (CRN) for Internet of Things (IoT) in the presence of an eavesdropper (EAV) is investigated. In particular, a primary transmitter (PT) communicates with multiple primary IoT device (ID) receivers (i.e., PRs) at the licensed frequency. Meanwhile a secondary transmitter (ST) uses the same licensed frequency of the primary network to send the confidential signal to an unmanned aerial vehicle (UAV) that is equipped with the reconfiguration intelligent surfaces (RIS) (called UAV-RIS) as a relay. The UAV-RIS then forwards this confidential signal to multiple secondary ID receivers (i.e., SRs). Furthermore, an EAV is located near the multiple SRs and can eavesdrop the confidential signal from the UAV-RIS. Accordingly, we propose the ST’s transmit power policy to not cause harmful interference to the primary network. In addition, the physical layer security (PLS) in terms of the secrecy outage probability (SOP) of the considered CRN is analyzed. The numerical results show that the secrecy performance of the secondary network is improved as the numbers of the RIS cells increase.

The rapid advancements in manufacturing technologies are transforming the current industrial landscape through Industry 4.0, which refers not only to the integration of information technology with industrial production, but also to the use of innovative technologies and novel data management approaches. The target is to enable the manufacturers and the entire supply chain to save time, boost productivity, reduce waste and costs, and respond flexibly and efficiently to consumers’ requirements.\ud \ud Industry 4.0 moves the digitization of manufacturing components and processes a step further by creating smart factories. Within this context, one of the key enabling technologies for Industry 4.0 is the adoption and integration of the Digital Twin (DT). However, most of the DT solutions provided by the current leading vendors are in fact digital models or digital shadows, and not digital twins. This is due to the fact that there is no common understanding of the definition of the DT amongst the leading vendors, and its usage is slightly different but showcased under the same umbrella of DT. In this paper, a DT framework is proposed that replicates the processes of a real production line for product assembly using the Festo Cyber Physical Factory for Industry 4.0 located at Middlesex University. Moreover, the paper introduces a viable framework for interlinking the physical system with its digital instance in order to offer extended predictive maintenance services and form a fully integrated digital twin solution.

Deep learning is an effective approach to many real-world problems. The effectiveness of deep learning models depends largely on the amount of data being used to train the model. However, these data are often private or sensitive, which make it challenging to collect and apply deep learning models in practice. In this paper, we introduce an anonymous deep neural network training protocol called ATP (Anonymous Training Protocol), in which each party owns a private dataset and collectively trains a global model without any data leakage to other parties. To achieve this, we use the technique of sharing random gradients with large aggregate mini-batch sizes combined with the addition of temporary random noise. These random noises will then be sent back through an anonymous network to be filtered out during the update phase of the aggregate server. The proposed ATP model allows protection of the shared gradients even when the aggregating server colludes with other n-2 participants. We evaluate the model on the MNIST dataset with the CNN network architecture, resulting in an accuracy of 98.09%. The results show that the proposed ATP model has high practical applicability in protecting privacy in deep learning.

Several flexible substrate materials potentially usable for wearable antennas are presented in this paper, including rubber, leather and foam. The measured electrical characteristics are compared to show their pros and cons in view of wearable antenna design. Two planar inverted-F antennas based on two promising materials are designed, fabricated and measured. Material characteristics and experimental results suggest that natural rubber, which has a relative permittivity of around 6.4 is an appropriate flexible material for wearable antennas.

A reconfiguration module based on a small printed circuit board (PCB) and commercial snap-on buttons for reconfigurable flexible textile antennas is proposed. The concept consists of a small PCB loaded with a reconfigurable circuitry including active and lumped components, which is combined with a pair of commercial male snap-on buttons for electrical connection to conductive textiles. The reconfiguration module can provide a secure and robust coplanar electrical connection between rigid electronics and flexible textile conductors, which is usually challenging for textile antennas. To validate this module concept, a frequency-reconfigurable planar inverted-F antenna using a varactor-based module realization is designed and simulated. The antenna is predicted to exhibit a frequency tuning range of 24.9% from 2.14 to 2.75 GHz with efficiency from 60.1% to 89.4%.

In recent years, research of optimal algorithms for short-term load forecasting has become popular. Optimal algorithms can improve forecast results in two main directions: precision and speed of convergence. To achieve both of these goals, researchers often use a combination of different algorithms. This paper describes a new combination of two optimal algorithms: Genetic Algorithm (GA) and Particle Swarm Optimization (PSO). The forecasting of the 24-hour daily-load on the normal working day of the SPC-Southern Power Corporation (a big local Vietnam Company) showed a significant improvement in the error as well as the speed of convergence with several selected days. In the most accurate case, the average error of prediction reaches 1.15% while the biggest error is 3.74% and the smallest is 0.02%

Statistical language models (LMs) have been applied in several software engineering applications. However, they have issues in dealing with ambiguities in the names of program and API elements (classes and method calls). In this paper, inspired by the success of Deep Neural Network (DNN) in natural language processing, we present Dnn4C, a DNN language model that complements the local context of lexical code elements with both syntactic and type contexts. We designed a context-incorporating method to use with syntactic and type annotations for source code in order to learn to distinguish the lexical tokens in different syntactic and type contexts. Our empirical evaluation on code completion for real-world projects shows that Dnn4C relatively improves 11.6%, 16.3%, 27.1%, and 44.7% top-1 accuracy over the state-of-the-art language models for source code used with the same features: RNN LM, DNN LM, SLAMC, and n-gram LM, respectively. For another application, we showed that Dnn4C helps improve accuracy over n-gram LM in migrating source code from Java to C# with a machine translation model.

Cloud storage providers can reduce storage costs by detecting identical files and storing only one instance of them. While appealing to the storage providers, this deduplication set-up raises various privacy concerns among clients. Various techniques to retrofit content confidentiality in deduplication have been studied in the literature. Nevertheless, data encryption alone is insufficient to protect users' privacy, for the ownership and equality information of the outsourced data left unprotected may have serious privacy implications. In this paper, we investigate a three-tier architecture that saves bandwidth otherwise incurred by server-side deduplication solutions, yet does not admit the client-side deduplication's leakage on file existence. Leveraging trusted SGX-enabled processors, we construct the first privacy-preserving data deduplication protocol that protects not only the confidentiality, but also the ownership and equality information of the outsourced data, offering better privacy guarantees in comparison with existing works on secure data deduplication. Our experiments show that the proposed protocol incurs low performance overhead over conventional solutions that provide weaker level of privacy protection.

API example code search is an important application in software engineering. Traditional approaches to API code search are based on information retrieval. Recent advance in Word2Vec has been applied to support the retrieval of API examples. In this work, we perform a preliminary study that combining traditional IR with Word2Vec achieves better retrieval accuracy. More experiments need to be done to study different types of combination among two lines of approaches.

Word2Vec is a class of neural network models that as being trained from a large corpus of texts, they can produce for each unique word a corresponding vector in a continuous space in which linguistic contexts of words can be observed. In this work, we study the characteristics of Word2Vec vectors, called api2vec or API embeddings, for the API elements within the API sequences in source code. Our empirical study shows that the close proximity of the api2vec vectors for API elements reflects the similar usage contexts containing the surrounding APIs of those API elements. Moreover, api2vec can capture several similar semantic relations between API elements in API usages via vector offsets. We demonstrate the usefulness of api2vec vectors for API elements in three applications. First, we build a tool that mines the pairs of API elements that share the same usage relations among them. The other applications are in the code migration domain. We develop api2api, a tool to automatically learn the API mappings between Java and C# using a characteristic of the api2vec vectors for API elements in the two languages: semantic relations among API elements in their usages are observed in the two vector spaces for the two languages as similar geometric arrangements among their api2vec vectors. Our empirical evaluation shows that api2api relatively improves 22.6% and 40.1% top-1 and top-5 accuracy over a state-of-the-art mining approach for API mappings. Finally, as another application in code migration, we are able to migrate equivalent API usages from Java to C# with up to 90.6% recall and 87.2% precision.

To support code migration, we introduce JV2CS, a tool to generate a sequence of C# API elements and related control units that are needed to migrate a given Java code fragment. First, we mine the mappings between single APIs in Java and C#. To overcome the lexical mismatch between the names of Java and C# APIs, we represent an API by its usages instead of its name. To characterize an API with its context consisting of surrounding APIs in its usages, we take advantage of Word2Vec model to project the APIs of Java JDK and C# .NET into corresponding continuous vector spaces. The transformation matrix between the two vector spaces is learned from a small set of human-written pairs of mappings. We use the transformation matrix to derive other mappings, and then use the mappings to generate the corresponding API sequence in C# via a phrase-based translation model. The video demo for JV2CS can be found at https://www.youtube.com/watch?v=MjTfmr9AmR8&feature=youtu.be.

The ubiquity of mobile sensors makes spatial crowd-sourcing a very promising platform for acquiring spatial tasks (i.e., tasks that are related to a location). Some frameworks have been successfully developed for crowdsourcing spatial tasks to a set of workers. Most of the current frameworks assume that all tasks belong to the same category and that workers are self-motivated to voluntarily perform tasks. However, the assumptions may not be practical in reality since different tasks may belong to different expertise and workers may not be self-incentivised to voluntarily perform tasks. In this paper, we introduce a reward-based approach for crowdsourcing spatial expert tasks (i.e., spatial tasks that are related to specific expertise). We formally define the Maximum Task Minimum Cost Assignment (MTMCA) problem and propose a solution for it. Subsequently, we perform various experiments to prove the usability and scalability of our approach as well as investigate factors that may effect the overall assignment. The experimental evaluation was conducted using both real-world and synthetic data sets.

Weather factors such as temperature, moisture, and air pressure are considered as geographic phenomena distributed continuously in space and without boundaries. Weather factors have field characteristics, meanwhile their data are collected discretely at nodes which are considered as spatial objects. In this article, the model of multivariate cube is employed to visualize the data of weather factors in two modes, object-based visualization and field-based visualization. On a multivariate cube, the 2-D Cartesian coordinate systems representing various factors at a node are embedded in a space-time cube at the position of the node on map plane, where the data of each factor are represented as histogram bars with respect to time. The representation of factors on a multivariate cube supports the object-based visualization and the field-based visualization. The mode of object-based visualization displays the variation of one or more factors over time at one or more nodes, the difference between the values of a factor at various spatial positions, as well as the correlation between various factors at one or more spatial positions at the same time. The mode of field-based visualization displays each factor on layers associated with time. Each factor layer is constituted by converting point data of the factor recorded at nodes to surface data. The mode of field-based visualization approaches the models of stopped process and dynamics to infer surface data from point data. The mode of field-based visualization indicates the value of factors at a certain spatial position, where the mode of object-based visualization may be applied to display data similarly to at nodes. The mutual transformation of data between two modes of object-based visualization and field-based visualization on a multivariate cube expands analytical problems from some locations of nodes to every point in space.

International audience; This paper deals with the experimental characterization of discharges propagating over pressboard of different thicknesses immersed in mineral and vegetable oils, under AC and DC voltages using a point-plane electrode arrangement. It's shown that the thickness of pressboard and the type of oil significantly influence the characteristics of creeping discharge and especially the stopping length Lf and the density of branches □. For a given thickness, Lf increases with the voltage; it decreases when the thickness increases. For a given voltage and thickness of pressboard, Lf is longer under AC than under DC whatever the oil type. And under DC, Lf is longer with positive point than with a negative point. However, the characteristics Lf(U) in presence of vegetable oil and mineral oil are very close except for thin pressboard.

This paper is devoted to a comparative study of partial discharge (PD) characteristics initiated in pressboard/mineral oil and pressboard/vegetable oil insulating systems. We mainly analyze the average number N and maximum apparent charge q ap.max of discharges versus the voltage. The results are discussed with respect to the thickness of pressboard and type of oil. It's shown that for a given voltage and pressboard of 4 mm thickness, any significant difference is observed in the value of N and q ap.max in both mineral and vegetable oils while for a thickness of 2 mm N and q ap.max are much higher in vegetable oil than in mineral oil. As concerns the phase, the maximum of discharge activity moves to the left (i.e. towards the low phases) when increasing the voltage. This would indicate that the influence of the electric field and capacitive effects decrease when the thickness increases. As concerns the phase, the maximum of discharge activity moves to the left (i.e. towards the low phases) when increasing the voltage.

International audience; This paper is devoted to the streamers propagation in natural esters (vegetable oils) in a point - plane electrode arrangement stressed by standard lightning impulse voltage (1.2/50 µs). We mainly investigate the shape and stopping length of streamers as well as the associated current. Six commercial oils (untreated natural ester) extracted from grape seeds (GS), sunflower (SF) and rape seed (RS), corn (CO), rice (RI) and sesame (SS) that could constitute potential liquids for high voltage applications are tested. A mineral oil (MO) is also tested for comparison. It's shown that the streamers are filamentary in the different oils for both polarities. For a given voltage, the stopping lengths (Lf) of streamers are longer when the point is positive than when it is negative; also, except mineral oil when the point is negative, the values of Lf are very close enough in all tested oils.

Acoustic source positioning plays an important role in military tracking unwelcome objects. A system for this application must be capable of dealing with the input recorded convolved mixture signals while minimizing the high communication and computation cost. This paper describes a distributed system for positioning multiple independent moving sources relying on acoustic signals. The sensors pre-process the sensed data to obtain the frequency features before compressing and sending it to the base. At the base, the source positioning are carried out via two clustering stages and an optimization method. Analysis and simulation results show that our system provides high accuracy and needs neither much communication nor complex computation in a distributed manner. It is robust even when there exists high noise with Rayleigh multi-path fading under Doppler effect and when the number of independent sources is greater than the microphone number.

International audience; This paper presents the experimental results of a comparative study of mineral oils, synthetic ester oil and natural ester oils (vegetable oils) issued from different origins through the analyse of the initiation threshold voltage of streamers, their pattern and their stopping length (final length) Lf as well as the associated current and electrical charge, in a point - plane electrode arrangement submitted to a lightning impulse voltage (1.2/50 µs). It's shown that the streamers are filamentary in the different tested oils whatever the polarity of voltage. Lf is longer when the point is positive than when it's negative whatever the oil. It's is generally longer in ester oils (natural and synthetic) than in some mineral oils. When the point is negative, Lf can be ten times higher in ester oils than in mineral ones. Basing on the stopping length of streamers, the positive polarity appears as the most unfavorable one for industrial applications and the dielectric strength of mineral oils we tested are generally better (in the considered experimental conditions) than that of ester oils.