The Long-Range (LoRa) system for the Internet of Things (IoT) has proved effective in infrastructure, energy consumption, resource reduction, and pollution control. However, in terms of security, LoRa faces several challenges. LoRa has a centralised architecture where all sensor data is transmitted to a fog (edge) server via LoRa gateways and processed at the fog server. This exposes sensitive sensor data to some major security threats, such as data tampering, injection of fake data and data loss. LoRa gateways in such architecture simply forward packets to the server. We present a blockchain-based open-source solution for LoRa Gateways that addresses the security vulnerabilities of this architecture and utilises the computing and storage resources of LoRa Gateways. The main challenges in implementing blockchain at LoRa gateways are limited processing, storage capabilities and battery power. The number of gateway nodes improves security and reduces the workload on the fog server. With the introduction of packet processing at the gateway, unnecessary packets are dropped at the gateway itself. Simulation results depict a reduced CPU usage at the fog server for up to 2 % when 16 end devices are being used. Further, our work reduces the overall packet handling time by up to 17% compared to existing works.

Edge computing (EC) has emerged as a promising technology to meet the demand for computational resources in Internet of Things (IoT) networks. With EC, the processing of massive data-intensive tasks can occur in proximity to IoT users. Thus, required constraints related to tasks, such as latency and Quality of Service (QoS) can be guaranteed. However, determining the task offloading strategy under various constraints, including resources, distance, and cost, remains an open issue. In this article, we study the task offloading problem from a matching perspective and propose an edge-user assignment algorithm (EUAA) that aims to maximize the resource utilization of edge servers and the number of assigned IoT users. A key concern in any matching algorithm is how to generate the preference order for either side. To generate preference orders for edge servers, we apply the analytical hierarchy process (AHP), considering criteria, such as distance from users to the server, latency, resource requirements, and pricing. This approach establishes the priority of users for matching to edge servers. From the IoT users’ perspective, we use cost and QoS parameters to enhance their satisfaction. We evaluate the performance of the proposed model based on the number of assigned users, server profit, number of satisfied users, edge server resource utilization, and execution time, comparing it with state-of-the-art schemes.

Real data plays a fundamental role in determining various features related to the collection site, such as monitoring, controlling, predictions, etc. Several systems in the Internet of Things (IoT) environment produce millions of data from the sensing node, and transmitting each of the data is costly in terms of bandwidth requirements, energy consumption, and protecting data from being corrupt or from potential threats such as man-in-the-middle attacks. In many environmental applications such as agriculture, the absolute change in the consecutive data points is usually very small (we call it slow changing environment). So, there is a need for a system that can predict the next data point within a predefined tolerable limit, then transmitting each data point can be avoided. To address this issue, we proposed an analytical prediction algorithm using estimations (APAEs). The algorithm is deployed and runs simultaneously in the three layers of architecture: sensing, fog, and cloud layers. The algorithm predicts the next data sensed by the sensor. If the difference between the actual sensed data point and the predicted data point is beyond the predefined tolerance, then the sensed value is sent to the fog node and further to the cloud; otherwise, the estimated value is accepted. We have implemented the proposed algorithm on a real testbed and also tested it on two data sets. We compare the amount of data points transmitted with the state of the state-of-the-art scheme. We also highlighted the reduction in energy consumption and high accuracy of our algorithm on the two data sets and a real testbed.

In an emergency healthcare situation, delay between injury and treatment is one of the most critical parameters with regard to survivability. Reduction in diagnosis/pre-treatment time by processing real-time ambulance data while en route to hospital can cut back the delay in treatment of the patient. However, several research challenges arise in accessing real-time patient data from ambulance to hospital while moving along different Road Side Units (RSUs). Due to the severity of medical data, there is a need to minimize computational losses along with costs due to migration and ambulance perceived latency. Considering the above scenarios, this paper formulates an average cost minimization problem keeping latency, energy, and loss function into deliberation as NP-hard. To solve the formulated problem, Minimum Cost Algorithm (MCA) using Federated Averaging (FedAvg) algorithm utilizing RSUs for effectively transferring real-time patient data to hospitals has been proposed considering above stated constraints altogether. Moreover, to handle imbalances in health data across different hospitals during processing, FedAvg algorithm combines augmentation techniques. Through experimental and prototype demonstration, the efficacy of proposed framework is shown by achieving$12.5 \%, 27 \%,$and$38 \%$reduction in an average total cost compared to other state-of-the-art techniques on real-world data sets, respectively.

The rapid and accurate identification of plant diseases is a critical challenge in agriculture, impacting cost, crop health and agricultural productivity. Traditional methods of disease detection heavily rely on manual inspection, which can be time-consuming, subjective, and prone to human error. In this work, we propose an end-to-end system for automated detection of disease with severity level in tomato plants using a fusion of deep learning methods with image processing techniques. Farmers can reduce the needless and excessive use of pesticides by using targeted and optimized treatment plans by precisely measuring the severity level of the disease. We have developed a testbed to show the effectiveness of our proposed end-to-end spray control in terms of accuracy of disease classification, severity detection, and amount of pesticide reduction. In comparison to the traditional spray approach, our suggested setup resulted in 20% pesticide reductions and a 29.68% energy savings.

Leader election is a fundamental task in distributed computing. It is a symmetry breaking problem, calling for one node of the network to become the leader , and for all other nodes to become non-leaders . We consider leader election in anonymous radio networks modeled as simple undirected connected graphs. Nodes communicate in synchronous rounds. In each round, a node can either transmit a message to all its neighbours, or stay silent and listen. A node v hears a message from a neighbour w in a given round if v listens in this round and if w is its only neighbour transmitting in this round. If v listens in a round in which more than one neighbour transmits, then v hears noise that is different from any message and different from silence. We assume that nodes are identical (anonymous) and execute the same deterministic algorithm. Under this scenario, symmetry can be broken only in one way: by different wake-up times of the nodes. In which situations is it possible to break symmetry and elect a leader using time as symmetry breaker? In order to answer this question, we consider configurations . A configuration is the underlying graph with nodes tagged by non-negative integers with the following meaning. A node can either wake up spontaneously in the round shown on its tag, according to some global clock, or can be woken up hearing a message sent by one of its already awoken neighbours. The local clock of a node starts at its wakeup and nodes do not have access to the global clock determining their tags. A configuration is feasible if there exists a distributed algorithm that elects a leader for this configuration. Our main result is a complete algorithmic characterization of feasible configurations. More precisely, we design a centralized decision algorithm, working in polynomial time, whose input is a configuration and which decides if the configuration is feasible. Using this algorithm we also provide a dedicated deterministic distributed leader election algorithm for each feasible configuration that elects a leader for this configuration in time O ( n 2 σ, where n is the number of nodes and σ is the difference between the largest and smallest tag of the configuration. We then ask the question whether there exists a universal deterministic distributed algorithm electing a leader for all feasible configurations. The answer turns out to be no, and we show that such a universal algorithm cannot exist even for the class of 4-node feasible configurations. We also prove that a distributed version of our decision algorithm cannot exist.

The design and development of energy and spectrum-efficient solutions are important in the success of the Internet of Things (IoT). Due to the presence of an enormous number of smart devices, such as sensors, actuators, and different household devices achieving such scalable and efficient solutions are challenging. A wireless sensor network (WSN) with dynamic spectrum access (DSA) capability, known as the cognitive radio sensor network (CRSN) is recently introduced to deal with spectrum scarcity problem. Although the spectrum scarcity is reduced with DSA paradigm, the energy-efficient solutions are still required to be addressed due to the involvement of energy constrained devices in CSRN. Clustering is one of the efficient ways to optimize the energy consumption in the networks. Due to combination of both WSN and cognitive radio network (CRN), existing solutions of WSN and of CRNs are not applicable to CRSN. In this article, we propose a neighbor discovery algorithm and two greedy k-hop clustering schemes (k-SACB-WEC and k-SACB-EC) for CRSN with the aim focusing on IoT applications, which require constant intracluster and intercluster communications. We focus on achieving bichannel connectivity while maximizing network life. In our clustering different parameters, such as nodes' residual energy, spectrum awareness, appearance probability of primary users (PUs) of channels, channel quality, robustness on PUs' arrival, and the Euclidean distance between nodes are taken into consideration to select the hop count and common channels for clusters. Through simulation, we have highlighted the performance improvements of our proposed schemes in terms of the lifetime of the network, number of clusters, stability of networks, and frequency of reclustering over recently reported clustering algorithm in CRSN. The simulation results show that k-SACB-WEC generates at least 40% less number of clusters as compared to k-SACB-EC, network stability-aware clustering (NSAC), Prolong-SEP (PSEP), SAC-WCM, and Cognitive LEACH (CogLEACH). Also, in terms of network stability, the k-SACB-WEC achieves at least approximately 100% higher number of rounds before the first node dead than the compared competitive approaches.

Leader election is a fundamental task in distributed computing. It is a symmetry breaking problem, calling for one node of the network to become the leader, and for all other nodes to become non-leaders. We consider leader election in anonymous radio networks modeled as simple undirected connected graphs. Nodes communicate in synchronous rounds. In each round, a node can either transmit a message to all its neighbours, or stay silent and listen. A node v hears a message from a neighbour w in a given round if v listens in this round and if w is its only neighbour transmitting in this round. If v listens in a round in which more than one neighbour transmits then v hears noise that is different from any message and different from silence. We assume that nodes are identical (anonymous) and execute the same deterministic algorithm. Under this scenario, symmetry can be broken only in one way: by different wake-up times of the nodes. In which situations is it possible to break symmetry and elect a leader using time as symmetry breaker? In order to answer this question, we consider configurations. A configuration is the underlying graph with nodes tagged by non-negative integers with the following meaning. A node can either wake up spontaneously in the round shown on its tag, according to some global clock, or can be woken up hearing a message sent by one of its already awoken neighbours. The local clock of a node starts at its wakeup and nodes do not have access to the global clock determining their tags. A configuration is feasible if there exists a distributed algorithm that elects a leader for this configuration. Our main result is a complete algorithmic characterization of feasible configurations. More precisely, we design a centralized decision algorithm, working in polynomial time, whose input is a configuration and which decides if the configuration is feasible. Using this algorithm, we also provide a dedicated deterministic distributed leader election algorithm for each feasible configuration that elects a leader for this configuration in time $O(n^2σ)$, where n is the number of nodes and σ is the difference between the largest and smallest tag of the configuration. We then ask the question if there exists a universal deterministic distributed algorithm electing a leader for all feasible configurations. The answer turns out to be no, and we show that such a universal algorithm cannot exist even for the class of 4-node feasible configurations. We also prove that a distributed version of our decision algorithm cannot exist.

Convolutional neural networks (CNNs) based on the discriminative learning model have been widely used for image denoising. In this study, a feed‐forward denoising CNN (DnCNN) with a parametric rectified linear unit (PReLU) is used to improve the denoising performance. PReLU enhances the model fitting of the DnCNN network without affecting computational cost. This network learns the leaky parameter of negative inputs in an activation function and therefore finds a proper slope in a negative direction. The proposed denoising network is based on residual learning, which comprises repeated convolutional and PReLU units along with batch normalisation. Residual learning with batch normalisation accelerates the network training, which can be used for blind Gaussian denoising. In this network, feature maps are processed by principal component analysis and transferred to subsequent convolution layers. An adaptive bilateral filter further processes the output image of the proposed CNN for image smoothening and sharpening. The mean and variance of the Gaussian kernel of adaptive filter vary from pixel to pixel. The performance of this network is analysed on BSD‐68 and Set‐12 datasets, and it exhibits an improvement in peak signal‐to‐noise ratio and structural similarity index metric and visual representation over other state‐of‐the‐art methods.

An electrocardiogram (ECG) records the electrical signal from the heart to check for different heart conditions, but it is susceptible to noises. ECG signal denoising is a major pre‐processing step which attenuates the noises and accentuates the typical waves in ECG signals. Researchers over time have proposed numerous methods to correctly detect morphological anomalies. This study discusses the workflow, and design principles followed by these methods, and classify the state‐of‐the‐art methods into different categories for mutual comparison, and development of modern methods to denoise ECG. The performance of these methods is analysed on some benchmark metrics, viz., root‐mean‐square error, percentage‐root‐mean‐square difference, and signal‐to‐noise ratio improvement, thus comparing various ECG denoising techniques on MIT‐BIH databases, PTB, QT, and other databases. It is observed that Wavelet‐VBE, EMD‐MAF, GAN2, GSSSA, new MP‐EKF, DLSR, and AKF are most suitable for additive white Gaussian noise removal. For muscle artefacts removal, GAN1, new MP‐EKF, DLSR, and AKF perform comparatively well. For base‐line wander, and electrode motion artefacts removal, GAN1 is the best denoising option. For power‐line interference removal, DLSR and EWT perform well. Finally, FCN‐based DAE, DWT (Sym6) soft, MABWT (soft), CPSD sparsity, and UWT are promising ECG denoising methods for composite noise removal.

Two anonymous mobile agents navigate synchronously in an anonymous graph and have to meet at a node, using a deterministic algorithm. This is a symmetry breaking task called rendezvous, equivalent to the fundamental task of leader election between the agents. When is this feasible in a completely anonymous environment? It is known that agents can always meet if their initial positions are nonsymmetric, and that if they are symmetric and agents start simultaneously then rendezvous is impossible. What happens for symmetric initial positions with non-simultaneous start? Can symmetry between the agents be broken by the delay between their starting times? In order to answer these questions, we consider space-time initial configurations (abbreviated by STIC). A STIC is formalized as [(u,v),δ], where u and v are initial nodes of the agents in some graph and δ is a non-negative integer that represents the difference between their starting times. A STIC is feasible if there exists a deterministic algorithm, even dedicated to this particular STIC, which accomplishes rendezvous for it. Our main result is a characterization of all feasible STICs and the design of a universal deterministic algorithm that accomplishes rendezvous for all of them without any a priori knowledge of the agents. Thus, as far as feasibility is concerned, we completely solve the problem of symmetry breaking between two anonymous agents in anonymous graphs. Moreover, we show that such a universal algorithm cannot work for all feasible STICs in time polynomial in the initial distance between the agents.

In this study, the authors propose a simple and tighter approximation to an integral representing the average of the product of two Gaussian Q functions over parametric and fading channels. This facilitates the symbol error probability computation of most of the coherent digital modulation techniques over and fading distributions. In order to improve the error performance of these digital modulation schemes, maximal‐ratio combining (MRC) diversity reception is also employed. Moreover, since and are generic fading scenarios, the widely used popular fading models like Rayleigh, one‐sided Gaussian, Nakagami‐ m , Nakagami‐ q and Rician can be unified under one umbrella. This study also gives an insight of the asymptotic approximations of the proposed integrals. In addition, the analytical framework is validated with the help of Monte Carlo simulations, which enables an easy practical implementation of the proposed work.

Convolutional neural networks (CNNs) are deep neural networks that can be trained on large databases and show outstanding performance on object classification, segmentation, image denoising etc. In the past few years, several image denoising techniques have been developed to improve the quality of an image. The CNN based image denoising models have shown improvement in denoising performance as compared to non‐CNN methods like block‐matching and three‐dimensional (3D) filtering, contemporary wavelet and Markov random field approaches etc. which had remained state‐of‐the‐art for years. This study provides a comprehensive study of state‐of‐the‐art image denoising methods using CNN. The literature associated with different CNNs used for image restoration like residual learning based models (DnCNN‐S, DnCNN‐B, IDCNN), non‐locality reinforced (NN3D), fast and flexible network (FFDNet), deep shrinkage CNN (SCNN), a model for mixed noise reduction, denoising prior driven network (PDNN) are reviewed. DnCNN‐S and PDNN remove Gaussian noise of fixed level, whereas DnCNN‐B, IDCNN, NN3D and SCNN are used for blind Gaussian denoising. FFDNet is used for spatially variant Gaussian noise. The performance of these CNN models is analysed on BSD‐68 and Set‐12 datasets. PDNN shows the best result in terms of PSNR for both BSD‐68 and Set‐12 datasets.

In this study, the authors propose a simple yet tighter approximations for the special two‐dimensional Gaussian Q functions using the Trapezoidal rule of numerical integration. This enables a simplified and accurate symbol error probability (SEP) approximation of the hexagonal‐quadrature amplitude modulation (HQAM) in additive white Gaussian noise channel. The proposed approximation further simplifies the SEP calculation of HQAM in parametric , , and fading distributions. Also, the SEP of HQAM over log‐normal shadowing is calculated in this study. The accuracy of the analytical framework is verified using computer simulations.

Extreme learning machine (ELM) is a generalized single hidden layer feed forward network in which weights and biases between the input layer and hidden layer are randomly assigned whereas, the weights between the hidden layer and the output layer are analytically determined. The optimal number of hidden neurons in ELM is evaluated by varying the number of hidden neurons in some range. Most of the recently published articles, quote the number of hidden neurons at which the ELM gives the maximum testing accuracy. Ideally, the model should not be selected using the testing accuracy because testing dataset is unseen. Selecting the number of hidden neurons by observing the training accuracy might be misleading as solution with higher training accuracy might suffers from overfitting. In this work, we have developed a variant of ELM which does not required manual tuning of the number of hidden neurons. The proposed ELM variant also has a minimum network structure, with slightly less testing performance compared to original ELM. In original ELM highest testing performance is quoted without any description to select the optimal number of hidden neurons. The proposed ELM variant initially set the number of hidden neurons to higher value and then removes the highly correlated hidden neurons to minimize the network structure. The proposed work also removes the overfitting problem in original ELM. Experimental results have been shown on some popular datasets taken from keel repository.

Equalization refers to any signal processing technique used at the receiver to combat intersymbol interference in dispersive channels. This paper reviews the applications of artificial neural networks (ANNs) in modeling nonlinear phenomenon of channel equalization. The literature associated with different feedforward neural network (NN) based equalizers like multilayer perceptron, functional-link ANN, radial basis function, and its variants are reviewed. Feedback-based NN architectures like recurrent NN equalizers are described. Training algorithms are compared in terms of convergence time and computational complexity for nonlinear channel models. Finally, some limitation of current research activities and further research direction is provided.

In this paper, a learning algorithm for a single integrate-and-fire neuron (IFN) is proposed and tested for various applications in which a multilayer perceptron based neural network is conventionally used. It is found that a single IFN is sufficient for the applications that require a number of neurons in different hidden layers of a conventional neural network. Several benchmark and real-life problems of classification and function-approximation have been illustrated. It is observed that the inclusion of some more biological phenomenon in an artificial neural network can make it more powerful.