Search

In addressing multi-agent reinforcement learning (MARL) challenges, Multi-Agent Transformer (MAT) has demonstrated a number of notable successes. In various benchmarks, MAT consistently showed a strong performance. A key observation in the latest MAT frameworks is MARL modeling with sequence modeling (SM) to represent the inter-agent relationships by self-attention mechanisms. This study applies graph-based modeling to represent the inter-agent relationships present in agent interactions to improve performance. To this end, we introduce the so-called MAT-GAT model, which leverages Graph Attention Networks (GAT) to allow for individualized consideration of interactions between agents. This enables MAT to pay more attention to information relative to inter-agent interactions within a cooperative MARL environment. To evaluate the performance of MAT-GAT, we conducted a series of benchmark tests across three different levels of StarCraft Multi-Agent Challenge (SMAC) tasks and the MuJoCo Half-cheetah task. The test results indicate that MAT-GAT outperforms both the original MAT and state-of-the-art baselines such as QMIX, particularly in complex environments. This demonstrates MAT-GAT’s improved performance with respect to its representation capabilities and learning.

Abstract Scimitar syndrome is a rare condition characterized by partial or complete anomalous pulmonary venous drainage of the lung to the inferior vena cava, right lung hypoplasia and dextroposition of the heart. Haemoptysis is uncommon in adults, although the clinical spectrum is wide. We report a case of a 38‐year‐old male with scimitar syndrome who had low grade haemoptysis persisting over several years secondary to bronchiectatic changes in his hypoplastic right lung. Conservative measures to manage haemoptysis were unsuccessful and the patient proceeded to bronchial artery embolization. The post‐procedure course was complicated by lung infarction and the patient ultimately required pneumonectomy. Deformities of the blood vessels and lungs are complex in scimitar syndrome. Bronchial artery embolization should be approached cautiously to protect pulmonary perfusion.

With the growing interest in both public and private 5G services based on millimeter wave (mmWave) communication for indoor usage scenarios such as smart factories, site design specialists are seeking sophisticated methods and tools for simulating indoor radio coverage based on highly accurate path loss prediction models. Although machine learning approaches can be used in path loss modeling thanks to the highly accurate prediction capability, their performance can be limited by the size of available measurement data set used for training. In this paper, we propose new approaches to train path loss models in the few-shot learning scenarios of smart factories. The proposed approaches are based on meta-learning with slight modifications to perform fine-tuning over an entire train data set rather than a meta-test data set. It is shown that the indoor path loss models based on convolutional neural networks (CNNs) trained by meta-learning based on three different meta-train task assignment schemes outperform both a conventional CNN model and an empirical model.

An accurate and efficient path loss modeling method for millimeter wave communications plays a significant role in the large-scale deployment of a fifth-generation (5G) mobile communication system. Conventional path loss modeling methods such as deterministic methods, empirical methods, and machine learning-based methods, in practice, cannot achieve the desired level of performance in terms of accuracy. This paper proposes a novel path loss model for 5G communications in suburban scenarios using deep learning with dilated convolution and attention. Dilated convolution is used to alleviate the locality of feature extraction and capture the global information on input images. Attention rendered by global context blocks helps the attention-enhanced convolutional neural network (AE-CNN) model utilize the global information on inputs to extract essential features of propagation environments. A distance-embedded local area multi-scanning algorithm which generates input images that can improve learning the latent features with dependency on distance is proposed. The experimental results indicate that the AE-CNN model can outperform state-of-the-art deterministic and empirical methods in terms of root mean square error in test scenarios.

The performance of traditional routing protocols for mesh networks can decrease significantly in disaster recovery situations due to dynamic changes to the network environment. A number of multi-path routing protocols have been proposed to address such issues. MMQR is a multi-path routing protocol for OFDM–TDMA mesh networks which use multiple paths that can be replaced immediately by one another in case of link failures. In this paper we extend the MMQR such that the alternative paths are reordered adaptively as the communication environment changes based on the heuristic model of attractor selection mechanism by which biological entities are known to adapt to dynamically changing environment. Simulation results show that the multi-path routing protocol with attractor selection mechanism outperforms existing multi-path routing protocols in terms of packet delivery ratio, throughput and QoS.

The design of conventional electric vehicles (EVs) is affected by numerous limitations, such as a short travel distance and long charging time. As one of the first wireless charging systems, the Online Electric Vehicle (OLEV) was developed to overcome the limitations of the current generation of EVs. Using wireless charging, an electric vehicle can be charged by power cables embedded in the road. In this paper, a model and algorithm for the optimal design of a wireless charging electric bus system is proposed. The model is built using a Markov decision process and is used to verify the optimal number of power cables, as well as optimal pickup capacity and battery capacity. Using reinforcement learning, the optimization problem of a wireless charging electric bus system in a diverse traffic environment is then solved. The numerical results show that the proposed algorithm maximizes average reward and minimizes total cost. We show the effectiveness of the proposed algorithm compared with obtaining the exact solution via mixed integer programming (MIP).

To allow rapid growth of the number of base stations, reducing the energy consumption of the stations, as the main energy consumers in cellular networks, has become an important research topic. In this paper, we attempt to find an adaptive cell zooming method to reduce the energy consumption of base stations. The cell zooming mechanism was formulated as an optimization problem with consideration of varying traffic patterns and interference, as well as the service availability of the whole area. Simulations were then conducted to verify the performance of the proposed cell zooming method. The simulations considered varying traffic conditions, both timely and spatially, in traditional 19-cell configuration. The proposed scheme demonstrated reduction of energy consumption of up to 4.72 times for urban environments and 3.78 times for rural environments against traditional static cell operation.

This paper presents a location-based interactive model of Internet of Things (IoT) and cloud integration (IoT-cloud) for mobile cloud computing applications, in comparison with the periodic sensing model. In the latter, sensing collections are performed without awareness of sensing demands. Sensors are required to report their sensing data periodically regardless of whether or not there are demands for their sensing services. This leads to unnecessary energy loss due to redundant transmission. In the proposed model, IoT-cloud provides sensing services on demand based on interest and location of mobile users. By taking advantages of the cloud as a coordinator, sensing scheduling of sensors is controlled by the cloud, which knows when and where mobile users request for sensing services. Therefore, when there is no demand, sensors are put into an inactive mode to save energy. Through extensive analysis and experimental results, we show that the location-based model achieves a significant improvement in terms of network lifetime compared to the periodic model.

Three-dimensional (3D) video is expected to be a “killer app” for OFDMA-based broadband wireless systems. The main limitation of 3D video streaming over a wireless system is the shortage of radio resources due to the large size of the 3D traffic. This paper presents a novel resource allocation strategy to address this problem. In the paper, the video-plus-depth 3D traffic type is considered. The proposed resource allocation strategy focuses on the relationship between 2D video and the depth map, handling them with different priorities. It is formulated as an optimization problem and is solved using a suboptimal heuristic algorithm. Numerical results show that the proposed scheme provides a better quality of service compared to conventional schemes.

The IEEE Wireless Access for Vehicular Environment (WAVE) standard consists of the IEEE 802.11p and the IEEE 1609 specifications supporting fast link establishment as well as broadband data communication in vehicular environment. Although the standard has been widely accepted by industry as well as highway authorities in many countries to be used for the radio communication infrastructure of the next-generation Intelligent Transport System (ITS), it remains open for further enhancements such as seamless handover support. This paper presents two proactive caching and forwarding schemes for seamless handover in IEEE WAVE networks: one for straight highway sections and the other for crossroads and junctions. Practical and efficient handover procedures, based on the proposed schemes, are also presented. Specifically, an on-board unit (OBU) informs the old road-side unit (RSU) about its departure from the coverage area such that the old RSU can forward the buffered packets to candidate RSUs for proactive caching. In handover scenarios around crossroads and junctions, the IEEE 802.11f Move-notify message is used by the new RSU to request the rest of candidate RSUs to purge the cached packets in order to prevent the waste of radio resource. Simulation results show that the proposed schemes outperform conventional schemes.

Scimitar syndrome is a rare condition characterized by partial or complete anomalous pulmonary venous drainage of the lung to the inferior vena cava, right lung hypoplasia and dextroposition of the heart. Haemoptysis is uncommon in adults, although the clinical spectrum is wide. We report a case of a 38-year-old male with scimitar syndrome who had low grade haemoptysis persisting over several years secondary to bronchiectatic changes in his hypoplastic right lung. Conservative measures to manage haemoptysis were unsuccessful and the patient proceeded to bronchial artery embolization. The post-procedure course was complicated by lung infarction and the patient ultimately required pneumonectomy. Deformities of the blood vessels and lungs are complex in scimitar syndrome. Bronchial artery embolization should be approached cautiously to protect pulmonary perfusion. [ABSTRACT FROM AUTHOR]

Purpose: To prospectively evaluate four magnetic resonance sequences [ProSet, fluid attenuation inversion recovery (FLAIR), balanced turbo field echo (B-TFE), T2 Drivel for the lumbosacral plexus and its branches. Materials and Methods: Ten healthy volunteers who underwent four MRI sequences on lumbosacral area were evaluated for image quality (ito 5; 1 = poor, 5 = excellent), the number of visualized bilateral spinal nerves from L2 to Si, the overlapping vessels obscuring the plexus (1 = many, 2 = some, 3 = few), and image quality defining spinal nerves (0 = nonvisualized, 1 = poor, 2 = moderate, 3 = good). Results: The ProSet (mean = 4.2, range 3-5) and B-TFE (mean = 3.7, range 3-5) showed better image quality than others. The number of visualized spinal nerves was the largest on ProSet image (mean = 9.2, range 8-10). FLAIR (mean = 2.1, range 1-3) and 12 Drive sequences (mean = 2.1, range 1-3) discriminated the nerves well from the vessels. The main branches of the lumbosacral plexus were well visualized on both ProSet (mean = 2.9, range 2-3) and FLAIR images (mean = 2.6, range 1-3). All of these were statistically significant. Conclusion: ProSet is the best sequence in the evaluation of the lumbosacral plexus and its major branches while FLAIR can be a complementary sequence for the evaluation of nerves overlapping vascular structures. [ABSTRACT FROM AUTHOR]