Browse

You are looking at 1 - 10 of 3,134 items for :

  • Electrical Engineering x
Clear All
Open access

Yun Wei, Ying Yu, Lifeng Xu, Wei Huang, Jianhua Guo, Ying Wan and Jinde Cao

Abstract

Vehicle emission calculation is critical for evaluating motor vehicle related environmental protection policies. Currently, many studies calculate vehicle emissions from integrating the microscopic traffic simulation model and the vehicle emission model. However, conventionally vehicle emission models are presented as a stand-alone software, requiring a laborious processing of the simulated second-by-second vehicle activity data. This is inefficient, in particular, when multiple runs of vehicle emission calculations are needed. Therefore, an integrated vehicle emission computation system is proposed around a microscopic traffic simulation model. In doing so, the relational database technique is used to store the simulated traffic activity data, and these data are used in emission computation through a built-in emission computation module developed based on the IVE model. In order to ensure the validity of the simulated vehicle activity data, the simulation model is calibrated using the genetic algorithm. The proposed system was implemented for a central urban region of Nanjing city. Hourly vehicle emissions of three types of vehicles were computed using the proposed system for the afternoon peak period, and the results were compared with those computed directly from the IVE software with a trivial difference in the results from the proposed system and the IVE software, indicating the validity of the proposed system. In addition, it was found for the study region that passenger cars are critical for controlling CO, buses are critical for controlling CO and VOC, and trucks are critical for controlling NOx and CO2. Future work is to test the proposed system in more traffic management and control strategies, and more vehicle emission models are to be incorporated in the system.

Open access

Yuchen Hou and Lawrence B. Holder

Abstract

Deep learning has been successful in various domains including image recognition, speech recognition and natural language processing. However, the research on its application in graph mining is still in an early stage. Here we present Model R, a neural network model created to provide a deep learning approach to the link weight prediction problem. This model uses a node embedding technique that extracts node embeddings (knowledge of nodes) from the known links’ weights (relations between nodes) and uses this knowledge to predict the unknown links’ weights. We demonstrate the power of Model R through experiments and compare it with the stochastic block model and its derivatives. Model R shows that deep learning can be successfully applied to link weight prediction and it outperforms stochastic block model and its derivatives by up to 73% in terms of prediction accuracy. We analyze the node embeddings to confirm that closeness in embedding space correlates with stronger relationships as measured by the link weight. We anticipate this new approach will provide effective solutions to more graph mining tasks.

Open access

Igor Aizenberg, Antonio Luchetta, Stefano Manetti and Maria Cristina Piccirilli

Abstract

A procedure for the identification of lumped models of distributed parameter electromagnetic systems is presented in this paper. A Frequency Response Analysis (FRA) of the device to be modeled is performed, executing repeated measurements or intensive simulations. The method can be used to extract the values of the components. The fundamental brick of this architecture is a multi-valued neuron (MVN), used in a multilayer neural network (MLMVN); the neuron is modified in order to use arbitrary complex-valued inputs, which represent the frequency response of the device. It is shown that this modification requires just a slight change in the MLMVN learning algorithm. The method is tested over three completely different examples to clearly explain its generality.

Open access

Keiko Ono, Yoshiko Hanada, Masahito Kumano and Masahiro Kimura

Abstract

In evolutionary computation approaches such as genetic programming (GP), preventing premature convergence to local minima is known to improve performance. As with other evolutionary computation methods, it can be difficult to construct an effective search bias in GP that avoids local minima. In particular, it is difficult to determine which features are the most suitable for the search bias, because GP solutions are expressed in terms of trees and have multiple features. A common approach intended to local minima is known as the Island Model. This model generates multiple populations to encourage a global search and enhance genetic diversity. To improve the Island Model in the framework of GP, we propose a novel technique using a migration strategy based on textit f requent trees and a local search, where the frequent trees refer to subtrees that appear multiple times among the individuals in the island. The proposed method evaluates each island by measuring its activation level in terms of the fitness value and how many types of frequent trees have been created. Several individuals are then migrated from an island with a high activation level to an island with a low activation level, and vice versa. The proposed method also combines strong partial solutions given by a local search. Using six kinds of benchmark problems widely adopted in the literature, we demonstrate that the incorporation of frequent tree information into a migration strategy and local search effectively improves performance. The proposed method is shown to significantly outperform both a typical Island Model GP and the aged layered population structure method.

Open access

Gustavo Botelho de Souza, Daniel Felipe da Silva Santos, Rafael Gonçalves Pires, Aparecido Nilceu Marana and João Paulo Papa

Abstract

Biometric systems have been widely considered as a synonym of security. However, in recent years, malicious people are violating them by presenting forged traits, such as gelatin fingers, to fool their capture sensors (spoofing attacks). To detect such frauds, methods based on traditional image descriptors have been developed, aiming liveness detection from the input data. However, due to their handcrafted approaches, most of them present low accuracy rates in challenging scenarios. In this work, we propose a novel method for fingerprint spoofing detection using the Deep Boltzmann Machines (DBM) for extraction of high-level features from the images. Such deep features are very discriminative, thus making complicated the task of forgery by attackers. Experiments show that the proposed method outperforms other state-of-the-art techniques, presenting high accuracy regarding attack detection.

Open access

Ondřej Macíček, Radovan Jiřík, Jan Mikulka, Michal Bartoš, Andrea Šprláková-Puková, Miloš Keřkovský, Zenon Starčuk, Karel Bartušek and Torfinn Taxt

Abstract

Dynamic contrast enhanced MRI (DCE-MRI) and dynamic susceptibility contrast MRI (DSC-MRI) are perfusion imaging techniques used mainly for clinical and preclinical measurement of vessel permeability and capillary blood flow, respectively. It is advantageous to apply both methods to exploit their complementary information about the perfusion status of the tissue. We propose a novel acquisition method that combines advantages of the current simultaneous and sequential acquisition. The proposed method consists of a DCE-MRI acquisition interrupted by DSC-MRI acquisition. A new method for processing of the DCE-MRI data is proposed which takes the interleaved acquisition into account. Analysis of both the DCE- and DSC-MRI data is reformulated so that they are approximated by the same pharmacokinetic model (constrained distributed capillary adiabatic tissue homogeneity model). This provides a straightforward evaluation of the methodology as some of the estimated DCE- and DSC-MRI perfusion parameters should be identical. Evaluation on synthetic data showed an acceptable precision and no apparent bias introduced by the interleaved character of the DCE-MRI acquisition. Intravascular perfusion parameters obtained from clinical glioma data showed a fairly high correlation of blood flow estimates from DCE- and DSC-MRI, however, an unknown scaling factor was still present mainly because of the tissue-specific r2* relaxivity. The results show validity of the proposed acquisition method. They also indicate that simultaneous processing of both DCE- and DSC-MRI data with joint estimation of some perfusion parameters (included in both DCE- and DSC-MRI) might be possible to increase the reliability of the DCE- and DSC-MRI methods alone.

Open access

Ruoyu Wei and Jinde Cao

Abstract

This paper investigates the global exponential synchronization and quasi-synchronization of inertial memristive neural networks with time-varying delays. By using a variable transmission, the original second-order system can be transformed into first-order differential system. Then, two types of drive-response systems of inertial memristive neural networks are studied, one is the system with parameter mismatch, the other is the system with matched parameters. By constructing Lyapunov functional and designing feedback controllers, several sufficient conditions are derived respectively for the synchronization of these two types of drive-response systems. Finally, corresponding simulation results are given to show the effectiveness of the proposed method derived in this paper.

Open access

Imrich Andráš, Pavol Dolinský, Linus Michaeli and Ján Šaliga

Abstract

This paper presents a way of acquiring a sparse signal by taking only a limited number of samples; sampling and compression are performed in one step by the analog to information conversion. The signal is recovered with minimal information loss from the reduced data record via compressed sensing reconstruction. Several methods of analog to information conversion are described with focus on numerical complexity and implementation in existing embedded devices. Two novel analog to information conversion methods are proposed, distinctive by their computational simplicity - direct subsampling and subsampling with integration. Proposed sensing methods are intended for and evaluated with real water parameter signals measured by a wireless sensor network. Compressed sensing proves to reduce the data transfer rate by >80 % with very little signal processing performed at the sensing side and no appreciable distortion of the reconstructed signal.

Open access

A.C. Gheorghe and M.S. Chiran

Abstract

The paper proposes the development of a weather station made up from a Raspberry Pi 3 and the Sense Hat shield. The shield used in this application has sensors for temperature, humidity and pressure. The weather station can connect to the internet wirelessly or with a standard UTP (Unshielded twisted pair) connection, this connection gives us the ability to take part in a scientific community for the sole purpose to monitor the weather all around the world. The program for the weather station is made in Python, the program language is easy to used and very versatile.

Open access

A. C. Gheorghe

Abstract

The paper proposes the development of a wireless video camera that is used to capture images from any microscope, and is universally compatible with the standard C-Mount. The purpose behind developing this camera is to offer an economical alternative to the current high-priced microscopy camera without compromising performance. In addition this camera has the technology to work wirelessly.