International Journal of Communication Systems最新文献

A wireless sensor network (WSN) is deployed to monitor agricultural environments, helping to overcome issues like suboptimal crop selection that can result in lower yields and diminished quality. To enhance crop recommendation, this work proposes the Kronecker LeNet Forward Harmonic Net (KLeFHNet). The workflow begins with WSN simulation and energy prediction using a DRNN, after which clusters are formed using the Dung Namib Optimizer (DNO) and routing is handled by the Fractional DNO (FDNO). The combination of data normalization, Wave–Hedges metric-based feature selection, and oversampling for data augmentation boosts the performance of the model. The integration of the Deep Kronecker Network, LeNet, and Forward Harmonic Network within KLeFHNet boosts energy efficiency and prolongs network longevity. Experimental results show that FDNO achieves an energy consumption of 0.353 J, a network lifetime of 74.271, and a trust of 86.266, whereas KLeFHNet attains a TPR of 0.900, TNR of 0.895, and FPR of 0.105, supporting sustainable, data-driven agriculture.

The recent technologies in IoT-based smart healthcare services are gaining attention because of their quality of service and reliability. These remote health monitoring systems offer enhanced quality of comfort to individuals and also meet emergency situation management requirements. Automatic data recording, processing, and communication with a third party like a doctor, caretaker, or hospital wirelessly enables the system to supersede conventional healthcare techniques. This paper focuses on the major challenging issues faced in the implementation of wireless body area networks (WBANs) in a dynamic environment, and the performance evaluation of the currently used systems. As the data collected by the sensor nodes are highly critical, a machine learning algorithm, integrating a rule-based screening and a bidirectional LSTM, is proposed to detect data manipulation by intruders. The performance of the network is then evaluated with the help of various parameters. The study results provide better insight into system performance and optimization of system parameters to achieve better reliability and throughput in the presence of false data.

Channel state information (CSI) is necessary for wireless systems supported by reconfigurable intelligent surfaces (RIS) to regulate wireless channels and maximize bandwidth as well as energy effectiveness. In this paper, an Adaptive Multilayer Contrastive Graph Neural Networks for Channel Estimation in Reconfigurable Intelligent Surface-Aided MIMO Schemes (RIS-MIMO-CE-MCGNN) is proposed. Using reconfigurable intelligent surfaces, multiuser millimeter wave big MIMO systems with lower overhead may estimate frequency flat and frequency-selective cascaded channels. Then uses a double-structured sparsity property of the angular cascaded channel matrices and the common sparsity property among the different subcarriers since distinct angle cascaded channels that are visible to various users have completely shared nonzero rows together with user-specific column supports. Support Detection using Adaptive Multilayer Contrastive Graph Neural Networks (AMCGNN). Then, the Channel estimation is performed using the Wader Hunt Optimization Algorithm (WHOA) technique. Finally, the system seeks to minimize the quantity of pilot overhead needed for precise channel estimate. The proposed technique attains 6.18%, 5.58% and 7.29% higher SNR and 7.88%, 5.88%, and 7.19% higher PSNR comparing with the existing methods: Channel estimation for reconfigurable intelligent surface-dependent full-duplex MIMO with hardware impairments (CE-RIS-MIMO-HI), RIS-supported Multiuser MIMO Systems Exploiting Extreme Learning Machine (RIS-MU-MIMO-EELM), and RIS-assisted mmWave-MIMO channel estimation utilizing DL and compressive sensing (RIS-MIMO-CE-CS), respectively.

Many mobile and sensor nodes comprised wireless sensor networks (WSN). Yet, it is quite challenging to locate these sensor and mobile nodes. Because of the time-varying movements, analysis of the current positions of sensor nodes in WSN is quite challenging. Because of locating all known sources in unknown nodes, the typical localization approaches are used to find the position of these nodes, producing a lot of inaccuracy when forecasting the distance between the source and unknown nodes. Also, it is very expensive to use Global Positioning System (GPS) technology for node detection. Although numerous localization procedures for WSNs in a three-dimensional topology have been proposed, it is still important to create and refine new localization algorithms to further increase the accuracy of the node positioning method. In this research work, an advanced heuristic algorithm and a deep learning technique are developed for localizing the unknown nodes in a three-dimensional wireless sensor network (3D-WSN). Initially, the distance between the unknown node as well as the anchor node is evaluated using efficient hybrid deep learning techniques named bidirectional long short-term memory (Bi-LSTM) and gated recurrent unit (GRU). Hybrid position of mine blast and chameleon swarm (HP-MBCS) is developed for tuning the parameters in deep learning techniques. An objective function of minimizing the average localization error (ALE) on node localization is obtained by optimally selecting the position of unknown nodes with the support of computed distance from the developed Bi-LSTM-GRU technique. The experimental simulation is carried out between the proposed and traditional models to show that the proposed model is efficient in minimizing localization error. The resultant outcome shows that the MEP value of the proposed HP-MBCS-Bi-LSTM-GRU model is 24.191, which is better than the other existing algorithms like EHO, EOO, MBO, and CSO, respectively. Thus, it was confirmed that the proposed Bi-LSTM-GRU not only improves the precision and effectiveness of node localization but also enhances the overall energy efficiency.

WSNs are essential in industrial and environmental areas, where they can be used to monitor things and track objects in real time. In WSNs that achieve accurate target localization, it is a primary problem in object tracking, primarily in dynamic environments. Additionally, tracking of objects is complicated due to the motion of objects being nonlinear in real-life conditions, which may negatively affect the performance of the tracking process. To address it, the current research will introduce a new hybrid tracking algorithm, which is a combination of Extended Kalman Filters (EKF) and Particle Swarm Optimization (PSO) to measure target state and adjust covariance parameters in WSNs. In this case, the adaptive parameter tuning will be carried out by PSO, whereas EKF modulates the real-time alterations dynamically, which will increase the monitoring efficiency under unpredictable circumstances. This combination enhances precision in tracking whereby filter parameters are constantly diminished with respect to the changing motion of the target. EKF has been applied to predict the state of a target in real-life situations because of its success in estimating the state of a free nonlinear system with Gaussian noise. A static filter configuration can, however, be inadequate in dynamic environments. PSO guarantees reliable tracking of objects regardless of the unpredictable situations by addressing this limitation by modulating EKF parameters in real time. In order to evaluate the performance of the proposed EKF-PSO approach, the study is applied in MATLAB using two network topologies: Random Topology (RT) and Hybrid Topology (HT). The various parameters that are employed to quantify the workability of the suggested approach include power usage, delay, and tracking error. The results of simulations for 100 repeated runs indicate that root mean square error decreases by 28% and energy savings increase by 15% compared with conventional EKF. The algorithm is dynamic in conditions of nonlinear motion and motion noise. From the results, it is shown that HT-EKF-PSO significantly outperforms RT-EKF-PSO with all metrics.

This paper describes the design and implementation of a dual-layer frequency selective surface (FSS) based dual-band high-gain circularly polarized (CP) meander-shaped monopole antenna for the applications of Wi-Fi and C-band. To achieve the final configuration, the design process involves several steps: designing the meander-shaped antenna, modeling the FSS, circuit analysis, and practical realization. The antenna prototype (electrical dimension: 0.333 λ₀ × 0.266 λ₀, λ₀ at 2 GHz and physical dimension of 50 × 40 mm²) comprises a meander-shaped strip and a 3 × 3 matrix dual layer periodic FSS embedded under the antenna structure without disturbing the impedance bandwidth. The low-cost FR4 dielectric substrate is used to design both the antenna and FSS layer. Initially, only the antenna part was designed, which yielded −10 dB impedance bandwidths (IBWs) and 3-dB axial ratio bandwidths (ARBWs) of 1160 and 600 MHz in the lower band, and 560 and 700 MHz in the upper band, respectively. Without an FSS structure, the peak gains of the antenna are 3.6 dBi (lower band) and 3.8 dBi (upper band). The proposed FSS-based antenna is fabricated and measured in the microwave test bench. The measured results show 3-dB ARBWs of 350 MHz (2.35–2.7 GHz) and 600 MHz (3.9–4.5 GHz) with LHCP waves. The measured peak gains are 8 dBi in the lower band and 8.5 dBi in the upper band. With small tolerance, the measured results agree with simulations.

Sensor localization is a crucial factor in ensuring reliable operation in wireless sensor network (WSN). The existing research on localization focuses more under isotropic conditions, assuming homogeneous environments with uniformly distributed nodes. In contrast, real-world deployments frequently exhibit Anisotropic WSN (AWSN) characteristics, including irregular topologies, obstacles, and non-uniform connectivity, which significantly challenge the localization process. Among many localization methods, the DV-Hop algorithm is widely used approach for its cost-effectiveness and easy implementations. However, the DV-Hop algorithm performs poorly when there is an obstacle encountered; the average hop distance might deteriorate, and it is hard to locate the coordinates of the location of unknown nodes. In order to achieve better localization accuracy with cost-effective, this study proposes Domination-based Anchor Placement for AWSNs (DAPA) to obtain minimum anchor requirement and optimal anchor placement. Further, a Q-learning-based Crayfish Optimization Algorithm (QCOA) is proposed to enhance the DV-Hop algorithm localization accuracy. The proposed QCOA was compared using 10 benchmark functions with some similar existing optimization algorithms. The proposed DAPA approach was compared with random anchor deployment method using three variants of DV-Hop algorithms. DAPA outperform the random anchor deployment in each three comparisons. Further, the proposed DAPA-based QCOA algorithm simulated in C-, S-, and O-shaped topologies based on localization error from varying the AWSN metrics. The proposed algorithm achieves high localization accuracy among the existing algorithms.

This research presents a wearable radiator with dual-band and quad-terminal that is compact and made on polyimide substrate. Dual metallic rings integrated with the feed line create the band notch in between 2.66 and 3.28 GHz and convert the wideband into dual band feature. By taking use of spatial and polarization diversity, the separation level among various terminals is more than 25 dB. A single-negative (SNG) metamaterial-based metasurface (MS) reflector-cum-absorber surface located just below the radiator reduces the specific absorption rate (SAR) for the 1- and 10-g tissue models by more than 90%, while also increasing the radiator gain to 5.65 and 4.55 dBi. ANSYS HFSS 2023 R2 was used for full-wave simulations, and a Keysight E5071C VNA in both flat and bending configurations was used for experimental validation. Using a three-layer human tissue phantom (skin, fat, and muscle) with 1- and 10-g averages, SAR assessment was conducted in accordance with IEEE/IEC 62209-1 and ICNIRP recommendations. Its performance in two frequency bands, 2.3–2.65 and 3.3–3.65 GHz, is confirmed by measurement findings. All these features make the radiator applicable for WBAN application.

In wireless sensor networks (WSNs) used for continuous surveillance, the problem of monitoring critical data transmitted infrequently is an extreme challenge of energy usage and latency requirements. Current medium access control (MAC) protocols often have high energy consumption, primarily owing to idle listening, collision, and excessive data transmission, and as a result, are not suitable for such uses. This study proposes a novel protocol to optimize energy consumption and transmission delays in WSNs used to monitor infrequent critical data. This protocol is named OWuR-MAC, that is, “Optimized Wake-up Radio based Medium Access Control.” OWuR-MAC implements an event-driven wake-up strategy utilizing wake-up receivers so that devices can stay in the low-power sleep mode until data transmission is necessary. Sensor nodes use wake-up receivers, which allow them to remain at low-energy sleep times until there is relevant data transmission that can wake them up. However, OWuR-MAC dynamically modifies the wake-up receiver sensitivity and transmission timing based on the characteristics of networked activity and environmental conditions. The protocol was implemented and compared with Fully Asynchronous Wake-up Radio MAC (FAWR-MAC) and Opportunistic Wake-up Radio MAC (OPWUM) protocols of a similar category. The results indicate that OWuR-MAC achieves lower rates of energy consumption, lower latency, and higher packet delivery ratios than the other two protocols.

Vehicular Ad Hoc Networks, in short VANETs, a category of mobile ad hoc networks, facilitate Vehicle-to-Vehicle (V2V) and Vehicle-to-Infrastructure (V2I) communication for improved traffic management, safety, and autonomous driving. This work proposes a traffic flow prediction model that integrates metaheuristic optimization techniques such as Sand Cat Swarm Optimization (SCSO) and Raven Roosting Optimization (RRO) with the deep learning model Bi-directional Long Short Term Memory with Stacked Autoencoder (Bi-LSTM-SA) to improve prediction accuracy. To optimize the Bi-LSTM-SA network, a hybrid SCSO-RRO approach encodes neuron parameters like weights, biases, and activation functions. The SCSO explores the search space while RRO refines solutions by having ravens follow high-fitness sand cats. A fitness function evaluates performance, and the process iterates until convergence, after which the optimized network is validated on a separate dataset. The proposed model Combined SCSO-RRO-Bi-LSTM-SA (C-SC-RRO-Bi-LSTM-SA) is compared with existing algorithms such as Random Forest (RF), Artificial Neural Network (ANN), Bi-LSTM-SA, and evaluation parameters utilized to quantify the network's performance include accuracy, prediction, recall, F1-score, and cross-entropy loss.