AUT Journal of Modeling and Simulation
Year:2024 | Volume:56 | Issue:1|Papers Count:8

Virtual power plant planning (VPP) has received much attention in recent years. VPP refers to the integration of multiple power units, considered as a single power plant. In this paper, three VPPs are considered, each consisting of different power plant units and expected to supply the desired load. In addition to providing the desired load, they must maximize their profits. A decentralized optimization method was used to optimize these three VPPs. The reason for using a decentralized approach is to increase network security and eliminate the need for a central computer. However, using decentralized optimization increases the speed of problem-solving. Finally, the obtained results are compared with the centralized method. Simulations show that almost the same results are achieved using different optimization methods. These results increase the trend of using decentralized methods in VPP. Another feature of decentralized methods compared to the centralized method is the reduction in the speed of problem-solving, which in this article has greatly reduced the solution time. If the considered network becomes wider and the number of problem variables and their limitations increases, the use of decentralized methods will become more efficient, and in those problems, the difference in problem-solving time by centralized and decentralized methods will increase.

The proliferation of Bluetooth Low Energy (BLE) in diverse applications has led to a dense clustering of piconets in confined spaces. BLE operates within the congested 2.4GHz ISM band, shared by numerous short-range wireless protocols. Frequency hopping and Adaptive Frequency Hopping (AFH) mechanisms have been introduced in Bluetooth to reduce interference with other protocols in the same band. However, research shows these mechanisms are ineffective in reducing inter-piconet interferences. These studies have demonstrated the relationship between the number of piconets and the level of interference. In this paper, we investigate the relationship between the similarity of channel maps and the number of neighbouring piconets in interference intensity. Then, we present a light and practical coordination framework to address this challenge. Our solution employs a BLE gateway to detect its surrounding piconet masters and computes a better channel map to reduce the impact of inter-piconet interferences. We also introduce the Isolated Channels (ICA) algorithm for channel allocation of neighbouring piconets with controlled channel overlaps for BLE5. Simulation results show a 20 to 60% reduction of interference level in environments with high to moderate inter-piconet interferences. To the best of our knowledge, this is the first practical BLE5-compatible solution for mitigating the inter-piconet interference problem and does not require modifying the standard stack.

New technologies have emerged over the last few years, such as IoT and fog computing. IoT devices and the enormous amounts of data generated every minute have led to the vast growth of the Internet of Things (IoT). In order to meet the term "Data Never Sleeps", some IoT applications require real-time services and low bit latency. To provide quick processing, storage, and services, Cisco proposed fog computing as an extension of cloud computing. The traditional methods are not capable of addressing the complex scheduling scenarios of fog computing. In this paper, we introduce a novel Fuzzy Reinforcement Learning Scheduling algorithm (FRLS) that enhances schedule accuracy in dynamic computing environments. To optimize task scheduling, the FRLS algorithm integrates fuzzy logic with reinforcement learning. To prioritize critical tasks, fuzzy logic handles uncertainty and prioritizes tasks according to deadlines, sizes, and file sizes. Then, reinforcement learning schedules the prioritized tasks, continually adjusting to dynamic conditions to ensure the best resource allocation. In addition to improving overall system performance, this combination provides a robust framework that can address the complexity and variability of fog computing environments. FRLS is designed to minimize response time while adhering to resource and deadline constraints in fog-based applications. A comparison of FRLS with existing algorithms shows that it significantly improves load balancing, deadline satisfaction, response time, and waiting time. Combining reinforcement learning and fuzzy logic leads to an efficient scheduling solution. In addition, FRLS outperforms non-prioritized algorithms.

This paper presents a novel design and analysis of a single-layer, exponentially tapered circuit analogue absorber (CAA) that is flexible and optically transparent. By modifying the edge of conventional crossed strips to an exponential taper, a wider bandwidth is achieved, analysed through current distribution and on the top layer of the unit cell. The designed unit cell comprises of ITO-Coated-PET a ground plane and CA absorber layer and a quarter-wavelength PVC dielectric substrate, achieving over 80% transparency and a relative bandwidth exceeding 85%. Comparative analysis with Numerical and experimental results and conventional CAA unit cells is conducted. Also, the current distribution is expressed mathematically. Parametric studies investigate various design parameters and CA elements to enhance impedance matching and absorption properties across the 2 to 25 GHz frequency range. Challenges such as the impact of layer count, substrate thickness, and analogue element type on transparency and absorption are mitigated by optimizing absorber dimensions and employing tapered shapes for strips to widen the bandwidth. This innovative CAA design finds applications in electromagnetic compatibility, radar cross-section reduction, offering a balance between wide bandwidth electromagnetic shielding and absorbing and high optical transparency, suitable for aircraft windshields, fighter canopies, space station windows end etc.

Skeleton-based action recognition has attracted significant attention in the field of computer vision. In recent years, Transformer networks have improved action recognition as a result of their ability to capture long-range dependencies and relationships in sequential data. In this context, a novel approach is proposed to enhance skeleton-based activity recognition by introducing Transformer self-attention alongside Convolutional Neural Network (CNN) architectures. The proposed method capitalizes on the 3D distances between pair-wise joints, utilizing this information to generate Joint Distance Images (JDIs) for each frame. These JDIs offer a relatively view-independent representation, allowing the model to discern intricate details of human actions. To further enhance the model's understanding of spatial features and relationships, the extracted JDIs from different frames are processed. They can be directly input into the Transformer network or first fed into a CNN, enabling the extraction of crucial spatial features. The obtained features, combined with positional embeddings, serve as input to a Transformer encoder, enabling the model to reconstruct the underlying structure of the action from the training data. Experimental results showcase the effectiveness of the proposed method, demonstrating performance comparable to other state-of-the-art transformer-based approaches on benchmark datasets such as NTU RGB+D and NTU RGB+D120. The incorporation of Transformer networks and Joint Distance Images presents a promising avenue for advancing the field of skeleton-based human action recognition, offering robust performance and improved generalization across diverse action datasets.

Accurate modelling and control of a soft arm has enormous importance because of the inherent softness and shape adaptability. The goal of this work is to represent an exact position control approach based on the precise Cosserat rod modelling method. To this aim, a compact Cosserat model of the pneumatically-actuated soft arm is extracted owing to the viscoelastic behaviour of the constructing material and the pneumatic pressure effects that appear as external loads. To address high nonlinearity, a PID sliding controller is suggested and formulated. The PDE equation set is solved by a recursive numerical method satisfying both boundary and domain requirements. Experimental results represented good behaviour of the proposed model, and the effectiveness of the control approach is demonstrated by simulation in both continuous and set-point positioning.

Graph embedding is the procedure of transforming a graph into a low-dimensional, informative representation. The majority of existing graph embedding techniques have given less consideration to the embedding distribution of the latent codes and more attention to the graph’s structure. Recently, Variational Graph AutoEncoders (VGAEs) have demonstrated good performance by learning smooth representations from unlabeled training samples. On the other hand, in regular VGAEs, the prior distribution over latent variables is generally a single Gaussian distribution. However, complex data distributions cannot be well-modelled under the assumption of a single Gaussian distribution. This choice of prior distribution is important because each dimension of a multivariate Gaussian can learn a separate continuous latent feature, which can result more structured and disentangled representation. In this paper, we employ the Gaussian Mixture Model (GMM) as the prior distribution in a Variational Graph Autoencoder (GMM-VGAE) framework for node classification in graphs. In this framework, GMM effectively discovers the inherent complex data distribution, and graph convolutional networks (GCNs) exploit the structure of the nodes of a graph to learn more informative representations. The proposed model incorporates several Graph Convolutional Networks (GCNs): one to map the input feature vector to the latent representation utilized for classification, another to generate the parameters of the latent distribution for learning from unlabeled data, and finally, an additional GCN is employed for reconstructing the input and delivering the reconstruction loss. Through extensive experiments on well-known Citations, Co-authorship, and Social network graphs, GMM-VGAE’s superiority over state-of-the-art methods is demonstrated.

Compared to traditional cryptography methods, physical layer secret key generation (PLSKG) can be more efficient and well-suited for Internet-of-Things (IoT) owing to its lightweight nature and scalability. In PLSKG, schemes utilizing local random generators are employed to achieve a high key generation rate. In this study, it is proposed a high-rate PLSKG in the presence of an untrusted relay. This untrusted relay assists the PLSKG process but cannot determine the secret key. We select channel probe signals from PSK signals and adopt a multi-bit quantizer at the receiver to enhance practicality. Additionally, we utilize quantization with guard bands (GB) to decrease the key error rate. We calculate the performance and security of the proposed PLSKG scheme under these conditions. Our results indicate that the relay, receiving superimposed signals from Alice and Bob, cannot ascertain the secret key. Finally, we compare the proposed PLSKG with a direct scenario where the relay is omitted during key generation.