فهرست مطالب

Science and Technology Transactions of Electrical Engineering - Volume:37 Issue: 2, 2013

Iranian Journal of Science and Technology Transactions of Electrical Engineering
Volume:37 Issue: 2, 2013

  • تاریخ انتشار: 1392/08/02
  • تعداد عناوین: 7
|
  • Pages 101-120
    The problem of sparse signal reconstruction from the well-known Compressed Sensing measurement is considered in this paper. The measured signal is assumed to be corrupted with additive white Gaussian noise with zero mean and known variance. Based on detection theory, two iterative algorithms are developed for detection and estimation of nonzero elements of sparse signal. The principle of the proposed methods is based on applying composite multiple hypothesis test to the underlying problem at each iteration. Simulation results show the satisfactory performance of the proposed algorithms in sparse signal recovery. The proposed approach has the potential of being applied to other models for noise and signal.
    Keywords: Sparse signal reconstruction, compressed sensing, detection theory, composite multiple hypothesis test
  • Pages 121-132
    We present a modified examplar-based inpainting method in the framework of patch sparsity. In the examplar-based algorithms, the unknown blocks of target region are inpainted by the most similar blocks extracted from the source region, using the available information. Defining a priority term to decide the filling order of missing pixels ensures the connectivity of the object boundaries. In the exemplar-based patch sparsity approach, a sparse representation of missing pixels is considered to define a new priority term and the unknown pixels of the fill-front patch is inpainted by a sparse combination of the most similar patches. Here, we modify this representation of the priority term and take a measure to compute the similarities between fill-front and candidate patches. Also, a new definition is proposed for updating the confidence term to illustrate the amount of the reliable information surrounding pixels. Comparative reconstructed test images show the effectiveness of our proposed approach in providing high quality inpainted images.
    Keywords: Image inpainting, texture synthesis, patch sparsity
  • Pages 133-145
    In general, identification and verification are done by passwords, pin number, etc., which are easily cracked by others. To overcome this issue, biometrics has been introduced as a unique tool to authenticate an individual person. Biometric is a quantity which consists of individual physical characteristics that provide more authentication and security than the password, pin number, etc. Nevertheless, unimodal biometric suffers from noise, intra class variations, spoof attacks, non-universality and some other attacks. In order to avoid these attacks, the multimodal biometrics, i.e. a combination of more modalities is adapted. Hence this paper has focused on the integration of fingerprint and Finger Knuckle Print (FKP) with feature level fusion. The features of Fingerprint and (FKP) are extracted. The feature values of fingerprint using Discrete Wavelet Transform and the key points of FKP are clustered using K-Means clustering algorithm and their values are fused. The fused values of K-Means clustering algorithm is stored in a database which is compared with the query fingerprint and FKP K-Means centroid fused values to prove the recognition and authentication. The comparison is based on the XOR operation.Hence this paper provides a multimodal biometric recognition method to provide authentication with feature level fusion. Results are performed on the PolyU FKP database and FVC 200 fingerprint database to check the Genuine Acceptance Rate (GAR) of the proposed multimodal biometric recognition method. The proposed multimodal biometric system provides authentication and security using K-Means clustering algorithm with GAR=99.4%, FRR=0.6% and FAR 0% with security of 128 bits for each modality.
    Keywords: Biometrics, feature level fusion, fingerprint, FKP feature extraction, K, Means clustering algorithm, multimodal biometric systems
  • Pages 147-159
    In this paper, we consider cognitive radio network in which two cognitive radio sources communicate with two cognitive destinations via a relay node. The decode and forward (DF) relay node employs physical layer network coding (PLNC) to improve the data rate. Based on the availability of the spectrum bands at the source, relay and destination, the network employs three different diversity schemes namely source to relay diversity, relay to destination diversity and combination of earlier two diversity schemes with overall source to destination diversity schemes. The optimal allocation of channel and power with per band and sum power constraints of a node in the network is formulated as convex optimization problem to improve the end to end throughput of the cognitive radio network. Simulation results show that the resultant joint channel and power allocation are superior to the equal power allocation in terms of both end to end throughput and outage probability.
    Keywords: Cognitive radio, physical layer network coding, channel, power allocation, throughput
  • Pages 161-182
    Diagnostics of synchronous generator turn-to-turn faults can be accomplished by analyzing the anomalies of machine local variable such as magnetic flux linkage. On the other hand, because current transformers and voltage transformers are usually installed in the machine for different purposes as default, the monitoring schemes that depend on the study of these inputs have become a topic of interest. As a consequence, in the last two decades, a variety of methods have been proposed based only on the global external variable (such as stator current and voltage). It must be noted that the global external variable analysis techniques are not capable of detecting all of the potential faults in electrical machines and some researches, including this work, aim to transfer a part of the global external variables analysis knowledge to techniques that monitor the signatures produced by the fault to the electromagnetic flux signals (local variable). Hence, this contribution deals with the analysis of the stator turn-to-turn short circuit fault of salient pole synchronous generator. In this paper, a fault indicator based on harmonic analysis of the magnetic flux linkage is investigated. Experimental results derived from a 4-pole, 380V, 1500 rpm, 50 Hz, 50 KVA, 3-phase salient-pole synchronous generator, show that the proposed method is useful and effective.
    Keywords: Synchronous generator, inter, turn fault, magnetic flux linkage, harmonic analysis
  • Pages 183-191
    The major problems in switched reluctance motors (SRMs) are radial force and torque ripple which cause increased undesirable acoustic noise. This paper describes an approach to determine optimum magnetic circuit parameters to minimize both radial force and torque ripple for such motors. There is no publication for simultaneous reduction of both radial force and torque ripple. In previous works, torque ripple was decreased without any research on the radial force or counter. In this paper, a procedure for radial force and torque ripple reduction in SR motors is proposed. To decrease the acoustic noise, the air gap width is increased while the radial force is maximized. On the other hand, by increasing the air gap width, torque decreases. By varying the angular interval and consequently the air gap width, the optimum angular interval is achieved. In the optimum angular interval, the radial force decreases while the torque remains constant. A twodimensional (2-D) finite element (FE) analysis carried out on the 6/4 SRM. By using the method of the compensated current, the ripple torque can be reduced to zero, radial force decreases 3.7%, and the acoustic noise power decreases 7.3% in the non-uniform air gap in comparison with the static case. Radial force decreases 5.6% and the acoustic noise power decreases 10.9% in the uniform air gap in comparison with the static case.
    Keywords: Switched reluctance motor, radial force, torque ripple, acoustic noise
  • Pages 193-198
    In recent years Fuzzy Wavelet Neural Networks (FWNNs) have been used in many areas. Function approximation is an important application of FWNNs. One of the main problems in effective usage of FWNN is tuning of its parameters. In this paper several different evolutionary algorithms including Genetic Algorithm (GA), Gravitational Search Algorithm (GSA), Evolutionary Strategy (ES), Fast Evolutionary Strategy (FES) and variants of Differential Evolutionary algorithms (DE) are used for adjusting these parameters on five test functions. The obtained results are compared based on some measures by using multiple non-parametric statistical tests. The comparison reveals the superiority of some variants of DE in terms of convergence behavior and the ability of function approximation.
    Keywords: Fuzzy wavelet neural networks, function approximation, evolutionary algorithms, nonparametric statistical test