Current Issue : April - June Volume : 2020 Issue Number : 2 Articles : 5 Articles
Recently, device-to-device (D2D) communications have been attracting substantial attention\nbecause they can greatly improve coverage, spectral efficiency, and energy efficiency, compared to\nconventional cellular communications. They are also indispensable for the mobile caching network,\nwhich is an emerging technology for next-generation mobile networks. We investigate a cellular\noverlay D2D network where a dedicated radio resource is allocated for D2D communications to\nremove cross-interference with cellular communications and all D2D devices share the dedicated\nradio resource to improve the spectral efficiency. More specifically, we study a problem of radio\nresource management for D2D networks, which is one of the most challenging problems in D2D\nnetworks, and we also propose a new transmission algorithm for D2D networks based on deep\nlearning with a convolutional neural network (CNN). A CNN is formulated to yield a binary vector\nindicating whether to allow each D2D pair to transmit data. In order to train the CNN and verify the\ntrained CNN, we obtain data samples from a suboptimal algorithm. Our numerical results show that\nthe accuracies of the proposed deep learning based transmission algorithm reach about 85% Approximately 95% in\nspite of its simple structure due to the limitation in computing power....
Image captioning is a comprehensive task in computer vision (CV) and natural language\nprocessing (NLP). It can complete conversion from image to text, that is, the algorithm automatically\ngenerates corresponding descriptive text according to the input image. In this paper, we present an\nend-to-end model that takes deep convolutional neural network (CNN) as the encoder and recurrent\nneural network (RNN) as the decoder. In order to get better image captioning extraction, we propose\na highly modularized multi-branch CNN, which could increase accuracy while maintaining the\nnumber of hyper-parameters unchanged. This strategy provides a simply designed network consists\nof parallel sub-modules of the same structure. While traditional CNN goes deeper and wider to\nincrease accuracy, our proposed method is more effective with a simple design, which is easier to\noptimize for practical application. Experiments are conducted on Flickr8k, Flickr30k and MSCOCO\nentities. Results demonstrate that our method achieves state of the art performances in terms of\ncaption quality....
An adaptive saturated neural network (NN) controller is developed for 6 degree-of-freedom (6DOF) spacecraft tracking, and its\nhardware-in-the-loop experimental validation is tested on the ground-based test facility. To overcome the dynamics\nuncertainties and prevent the large control saturation caused by the large tracking error at the beginning operation, a saturated\nradial basis function neural network (RBFNN) is introduced in the controller design, where the approximate error is\ncounteracted by an adaptive continuous robust term. In addition, an auxiliary dynamical system is employed to compensate for\nthe control saturation. It is proved that the ultimate boundedness of the closed-loop system is achieved. Besides, the proposed\ncontroller is implemented into a testbed facility to show the final operational reliability via hardware-in-the-loop experiments,\nwhere the experimental scenario describes that the simulator is tracking a planar trajectory while synchronizing its attitude with\nthe desired angle. Experimental results illustrate that the proposed controller ensures that the simulator can track a preassigned\ntrajectory with robustness to unknown inertial parameters and disturbances....
Aimed at reducing the switching loss and common-mode voltage amplitude of\nhigh-power medium-voltage three-level inverter under low modulation index conditions, an\nimproved synchronous space vector PWM strategy is proposed in this paper. The switching times\nin each fundamental period are reduced by the re-division of small regions and the full use of the\nredundant switching state. The sum of switching algebra is introduced as an evaluation index and\nthe switching state with the minimum value of the sum of switching algebra are adopted. Then, the\ncommon mode voltage amplitude is reduced. The theoretical analysis and experimental results\nshow that the improved modulation strategy proposed in this paper can effectively reduce the\nswitching loss and common-mode voltage amplitude of the inverter under the condition of the low\nmodulation index. Moreover, the neutral-point voltage ripple is also reduced simultaneously....
Although various algorithms have widely been studied for bankruptcy and credit risk\nprediction, conclusions regarding the best performing method are divergent when using different\nperformance assessment metrics. As a solution to this problem, the present paper suggests the employment\nof two well-known multiple-criteria decision-making (MCDM) techniques by integrating their preference\nscores, which can constitute a valuable tool for decision-makers and analysts to choose the prediction\nmodel(s) more properly. Thus, selection of the most suitable algorithm will be designed as an MCDM\nproblem that consists of a finite number of performance metrics (criteria) and a finite number of classifiers\n(alternatives). An experimental study will be performed to provide a more comprehensive assessment\nregarding the behavior of ten classifiers over credit data evaluated with seven different measures, whereas\nthe Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS) and Preference Ranking\nOrganization METHod for Enrichment of Evaluations (PROMETHEE) techniques will be applied to rank\nthe classifiers. The results demonstrate that evaluating the performance with a unique measure may lead\nto wrong conclusions, while theMCDMmethods may give rise to a more consistent analysis. Furthermore,\nthe use of MCDM methods allows the analysts to weight the significance of each performance metric\nbased on the intrinsic characteristics of a given credit granting decision problem....
Loading....