No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Faulty Sensor Detection Approach for Wireless Sensor Networks Based on Artificial Intelligence and Information Theory
ResearchGate has not been able to resolve any citations for this publication.
This paper presents a novel methodology for Climate Network (CN) construction based on the Kullback-Leibler divergence (KLD) among Membership Probability (MP) distributions, obtained from the Second Order Data-Coupled Clustering (SODCC) algorithm. The proposed method is able to obtain CNs with emergent behaviour adapted to the variables being analyzed, and with a low number of spurious or missing links. We evaluate the proposed method in a problem of CN construction to assess differences in wind speed prediction at different wind farms in Spain. The considered problem presents strong local and mesoscale relationships, but low synoptic scale relationships, which have a direct influence in the CN obtained. We carry out a comparison of the proposed approach with a classical correlation-based CN construction method. We show that the proposed approach based on the SODCC algorithm and the KLD constructs CNs with an emergent behaviour according to underlying wind speed prediction data physics, unlike the correlation-based method that produces spurious and missing links. Furthermore, it is shown that the climate network construction method facilitates the evaluation of symmetry properties in the resulting complex networks
Wireless sensor network (WSN) is a developed wireless network consisting of some connected sensor nodes. The WSN is employed in many fields such as military, industrial, and environmental monitoring applications. These nodes are equipped with sensors for sensing the environmental variables such as temperature, humidity, wind speed, and so on. In most applications, WSN is positioned in remote places and harsh environments, where they are most probably exposed to faults. Hence, faulty sensor identification is one of the most fundamental tasks to be considered in WSN. This study suggests a hybrid
methodology based on mutual information change (MIC) and wavelet transform (WT) for faulty sensor identification. The MIC method is suggested to study correlation among sensors, while the WT technique is proposed for self-sensor detection. WT is suitable for analysing non-stationary signals into approximation and detail coefficients. The suggested algorithm performance is investigated by applying a real case study at an arbitrary location close to Cairo, Egypt. The results of each method are compared using the true positive rate (TPR), false negative rate, and accuracy measures. Obtained results have shown that combining MIC and WT techniques can achieve a higher TPR and accuracy reach 100% in most fault types.
Wireless sensor networks (WSNs) are one of the very active research area. They have many applications like military, health care, environmental monitoring and industrial monitoring. The sensor nodes have limited energy source. Since in many cases the nodes are deployed in unreachable areas, hence recharging or replacing the battery of the sensor nodes is not an option. Therefore, one must employ techniques to conserve the energy by reducing the energy consumption by the nodes. In this paper, we discuss about the different energy saving schemes investigated by different research community in WSNs to reduce the energy consumption of the nodes and thereby improving the lifetime of the overall network. Energy saving protocols such as duty cycle, energy efficient routing, energy efficient medium access control (MAC), data aggregation, cross layer design and error control code (ECC) are discussed. Sleep/wake up method is adopted by the duty cycle approach to reduce the active time of the nodes and conserve their energy. The routing and MAC protocols use suitable energy efficient algorithms for saving energy. The data aggregation aims to save energy by reducing the number of transmissions. On the other side, cross layer approach looks for a cross layer optimization solution to improve the energy efficiency of the network. ECC reduces energy consumption by virtue of coding gain it offers which allows lower signal-to-noise ratio (SNR) to achieve the same bit error rate (BER) as an uncoded system. Some techniques such as use of directional antennas, topology control and transmission power control which have been widely investigated for other ad-hoc networks for energy conservation are also discussed in brief in this paper.
Rotating machines have been widely used in industrial engineering. The fault diagnosis of rotating machines plays a vital important role to reduce the catastrophic failures and heavy economic loss. However, the measured vibration signal of rotating machinery often represents non-linear and non-stationary characteristics, resulting in difficulty in the fault feature extraction. As a statistical measure, entropy can quantify the complexity and detect dynamic change through taking into account non-linear behavior of time series. Therefore, entropy can be served as a promising tool to extract the dynamic characteristics of rotating machines. Recently, many studies have applied entropy in fault diagnosis of rotating machinery. This paper aims to investigate the applications of entropy for the fault characteristics extraction of rotating machines. First, various entropy methods are briefly introduced. Its foundation, application and some improvements are described and discussed. The review is divided into eight parts: Shannon entropy, Rényi entropy, approximate entropy, sample entropy, fuzzy entropy, permutation entropy, and other entropy methods. In each part, we will review the applications using the original entropy method and the improved entropy methods, respectively. In the end, a summary and some research prospects are given.
The Internet of Things is becoming a very promising paradigm with the extensive market adoption of the development of associated technologies, such as; cloud computing, near-field communications, wireless mobile networks, etc. This will expose the future direction of communication around the world. Wireless Sensor Networks together with the existing communication technologies are enabling the continuous integration of controlling and processing the functionality of the Internet of Things applications. Since Wireless Sensor Networks are typically deployed for gathering sensitive information from unattended or hostile environments, they are exposed for security attacks, which are strongly affecting the user privacy and the network performance. There are various security mechanisms and solutions for Wireless Sensor Networks that have been proposed in the previous works. Therefore, it is mandatory to give attention for its applicability and feasibility features in terms of related security challenges based on the Internet of Things perspectives. The purpose of this paper is to explore and show the influence of Wireless Sensor Networks security challenges within the perspective of the Internet of Things and its applications. Consequently, an exploration of the major and minor security requirements in the Wireless Sensor Networks has been made in this paper, accompanied by a classification of the available attacks and threats against these requirements. Finally, a discussion on the Internet of Things security issues and challenges in Wireless Sensor Networks is provided.
This article presents a novel impact identification algorithm that uses a linear approximation handled by a statistical inference model based on the maximum-entropy principle, termed linear approximation with maximum entropy (LME). Unlike other regression algorithms as Artificial Neural Networks (ANN) and Support Vector Machines (SVM), the proposed algorithm requires only parameter to be selected and the impact is identified after solving a convex optimization problem that has a unique solution. In addition, with LME data is processed in a period of time that is comparable to the one of other algorithms. The performance of the proposed methodology is validated by considering an experimental aluminum plate. Time varying strain data is measured using four piezoceramic sensors bonded to the plate. To demonstrate the potential of the proposed approach over existing ones, results obtained via LME are compared with those of ANN and Least Square Support Vector Machines (LSSVM). The results demonstrate that with a low number of sensors it is possible to accurately locate and quantify impacts on a structure and that LME outperforms other impact identification algorithms.
Wireless Sensor Networks (WSNs) have attracted many researchers in the past few years due to their applicability for a wide-range of applications. WSNs rely on unreliable sensing schemes in which a sensor might lose some data due to the inherent characteristics of such networks. Estimating missing values that cope with other collected ones is crucial for some applications. In this paper, we introduced a framework dedicated to predicting missing values in WSNs. The key idea is to estimate missing values according to the natural spread (i.e. dispersion) of the guilty sensors. The framework considers cases in which distance and time play a significant role in estimating missing values. Thus, accurate values might be generated as compared to state-of-the-art central tendency measurements such as mean, median, mode, and midrange.
The Internet of Things is a paradigm that allows the interaction of ubiquitous devices through a network to achieve common goals. This paradigm like any man-made infrastructure is subject to disasters, outages and other adversarial conditions. Under these situations provisioned communications fail, rendering this paradigm with little or no use. Hence, network self-organization among these devices is needed to allow for communication resilience. This paper presents a survey of related work in the area of self-organization and discusses future research opportunities and challenges for self-organization in the Internet of Things. We begin this paper with a system perspective of the Internet of Things. We then identify and describe the key components of self-organization in the Internet of Things and discuss enabling technologies. Finally we discuss possible tailoring of prior work of other related applications to suit the needs of self-organization in the Internet of Things paradigm.
Wireless Sensor Networks are vulnerable to a plethora of different fault types and external attacks after their deployment. We focus on sensor networks used in healthcare applications for vital sign collection from remotely monitored patients. These types of personal area networks must be robust and resilient to sensor failures as their capabilities encompass highly critical systems. Our objective is to propose an anomaly detection algorithm for medical wireless sensor networks. Our proposed approach firstly classifies instances of sensed patient attributes as normal and abnormal. Once we detect an abnormal instance, we use regression prediction to discern between a faulty sensor reading and a patient entering into a critical state. Our experimental results on real patient datasets show that our proposed approach is able to quickly detect patient anomalies and sensor faults with high detection accuracy while maintaining a low false alarm ratio.
Sensor node faults are a serious threat to wireless sensor networks. They can cause node crashes or lead to the transmission of corrupted data. Especially the latter endangers the quality of subsequent data analyses. Most related fault detection approaches consider the sensor nodes as black boxes. They neglect vital information available on the node level. Consequently, most of these approaches can not distinguish between (i) irregular but correctly sensed data events and (ii) data corruption caused by soft faults. In contrast, our contribution integrates node-level diagnostics with the characteristics of the sensor data. We utilize this node-level diagnostic information to present our fault detection approach. Based on simulations and practical experiments, we show the correctness and efficiency of our approach. The results show that our approach offers a high fault detection rate and can differentiate between events and faults. Furthermore, it consumes a justifiably small overhead of resources and energy.
The aim of this research is to provide an approach for extension of the capacities of wireless sensor network (WSN) by enabling possibilities of monitoring of sea environment parameters. This domain area is related with control of recognition process of the hydro meteorological situations by providing the proper structure of the decision support system (DSS), which can work under specific real-time conditions. The construction issues of such multilayered buoy’s infrastructure based on WSN causes some problems: mechanisms of such interconnected devices have to work by properly presented goals and in extreme weather conditions; the modern embedded systems requires the intellectualization possibilities with recognition of adequate evaluation of risky situations; the functionality of WSN has to be based on the technology of the Internet of Things (IoT) with quite limited capacities. Our main findings concern the proposed architectural solution with extended capacities of integrated embedded systems and other gathering sources, by enabling the assessing of water pollution situations. The originality of the proposed approach concerns the issues of network allocation, system resource allocation and the architectural solutions, enabling wireless channels to provide big streams of data. The experimental results show the possibilities of the proposed prototype of multilayer system to work under real conditions of sea with interconnected network of sensors and controllers, based on the principles of other standardized network layers. The prototype method is used for evaluation of functionality of the embedded systems and WSN. The provided experimental results show functional effectiveness of collection and gathering of data under restricted and limited capacities of the embedded systems for operative control needs.
Data-based process monitoring is gaining increasing attention, especially deep learning modeling methods. Given that process data are inherently dynamic, the dynamic relationship of process data needs to be incorporated into the monitoring model. Measured variables have diverse characteristics; some variables have strong autocorrelation, but some have no or weak autocorrelation. Hence, a new nonlinear dynamic fault detection model based on mutual information (MI) and stacked sparse autoencoders (SSAE) was proposed. With input data taken as time series data, variables with a strong autocorrelation and their lag time of maximum autocorrelation were selected. Then, dynamic augmented data that contain the values of all measured variables at the current moment and the selected data were used as input of SSAE to extract features of cross-correlation and autocorrelation between variables. T2 and Q statistical quantities were constructed to implement fault detection. The proposed MI-dynamic SSAE can not only automatically learn the key features of the present measurements but can also learn the latent information contained in the dynamic process data. Simulation studies on the Tennessee-Eastman process were conducted to evaluate the performance of the proposed monitoring model.
Given the increasing size and dimension of streaming data made available by for example in industrial sensors or wireless sensor networks (WSNs), it is an important and worthwhile task to monitor not only the data itself but also the relationships between data sources. To solve this task, we present DIMID, an online algorithm to monitor dependencies in high dimensional streaming data. DIMID uses an entropy-based measure that generalizes to non-linear as well as complex functional types of relationships, is non-parametric and can be computed incrementally.
To deal with the streaming, possibly infinite data, DIMID contains a dimensionality reducing projection method and an estimator for entropy that uses the local density of data points. This also allows the algorithm to update the current relationships with new data as it becomes available, instead of recomputing on the complete batch after every update.
Comparisons to three state-of-the-art other algorithms for dependency-monitoring on a variety of time series data sets with linear and non-linear dependencies showed significant (p < 0.01) improvements in the AUC- and F1-measure. We also achieve a reduction in run-time from linear to logarithmic in the number of observed samples.
The collaborative and low-cost nature of wireless sensor networks (WSNs) brings significant advantages over traditional communication technologies used in today's electric power systems. Recently, WSNs have been widely recognized as a promising technology that can enhance various aspects of today's electric power systems, including generation, delivery, and utilization, making them a vital component of the next-generation electric power system, the smart grid. However, harsh and complex electric-power-system environments pose great challenges in the reliability of WSN communications in smart-grid applications. This paper starts with an overview of the application of WSNs for electric power systems along with their opportunities and challenges and opens up future work in many unexploited research areas in diverse smart-grid applications. Then, it presents a comprehensive experimental study on the statistical characterization of the wireless channel in different electric-power-system environments, including a 500-kV substation, an industrial power control room, and an underground network transformer vault. Field tests have been performed on IEEE 802.15.4-compliant wireless sensor nodes in real-world power delivery and distribution systems to measure background noise, channel characteristics, and attenuation in the 2.4-GHz frequency band. Overall, the empirical measurements and experimental results provide valuable insights about IEEE 802.15.4-compliant sensor network platforms and guide design decisions and tradeoffs for WSN-based smart-grid applications.
This paper presents a fault diagnosis procedure based on discriminant analysis and mutual information. In order to obtain good classification performances, a selection of important features is done with a new developed algorithm based on the mutual information between variables. The application of the new fault diagnosis procedure on a benchmark problem, the Tennessee Eastman Process, shows better results than other well known published methods.