فصلنامه روشهای هوشمند در صنعت برق
پیاپی 10 (تابستان 1391)

Substrate Integrated Waveguide (SIW) is a new type of guided wave structure which could be implemented in both microwave and millimeter wave integrated circuits. In this paper a novel waveguide type directional coupler based on substrate integrated waveguide consisting a double layer dielectric substrate is proposed. The design process is carried out using coupling theory and Chebyshev method. The proposed structure is fully based on SIW structure except at the input-output ports which are microstrip lines. The SIW bends are optimized for the lowest insertion loss using HFSS software. The center frequency of designed 20dB coupler is at 10GHz with 6GHz bandwidth. Based on the simulations the coupling remains almost constant over the coupler bandwidth with less than 2 dB variations. Larger bandwidth can be achieved by increasing the number of apertures. Also if we used a second set of apertures opposite side wall، the coupling would be increased by 6 dB.

Bearing current sininduction motorsare considered a sone of the most damaging factors. Induced shaft voltage through the parasitic capacitors cause this type of current. Inthispaper،given the increasing importance of direct torque control of induction motorin industry، various switching tables are assessed in order to ensure the lowest common voltage while maintaining the performance characteristics of the drive. Finally best switching table based on the minimum CMV، less torque rippleand better quality out put reference tracking is proposed.

Electrocardiogram (ECG) signals are the most popular non-invasive approach for diagnosis of heart irregularities and indications of possible heart diseases. Previous studies have shown that ECG signals do not have a linear distribution and contain a variety of non-linear dimensions. In the present research we have treated the ECG signals as time-series data and applied chaos indices analysis. Utilizing data from MIT_BIH Database، the present study has improved the past research by analysing chaotic indices such as Lyapunov Exponent (λmax)، and Correlation Dimension to ECG signal data from healthy individuals and heart patients. We present appropriate algorithms for reconstruction of Phase Space and estimations of the model parameters using Lyapunov Exponent and CorrelationDimension. We then present the results from reconstruction of Phase Space based on chaotic indices، and fuzzy classifier، to discriminate healthy individuals (NSR) from the heart patients. The heart patients include those with Arterial Fibrillation (AF) and those with Left Bundle Branch Block (LBBB). These results ascertain the effectiveness of application of chaotic distribution to ECG data in improving the heart disease diagnosis.

In switching power converters، electromagnetic interference (EMI) is produced due to the fast changes of the voltage and current. Among switching topologies used in low and medium powers، forward converter is vastly employed which needs resetting transformer core. In this paper، a forward converter with reset winding is modeled to predict conducted EMI. To evaluate the EMI model، EMI measurement results are presented. To select an optimum reset scheme from EMI viewpoint، three reset techniques are compared using experimental results. After selecting the proper reset scheme، two EMI mitigation methods are proposed and the practical EMI reduction is presented.

Today، the connection between human and computer is possible using mouse، keyboard and etc. These devices have some limitations like application speed. Making easy the interaction between human and computer is final objective of technology. In this paper، we propose one method for gesture recognition using curvature of B-Spline curves. First، the image of hand is extracted from different frames and some numbers of points are selected on the contour of hand in equal distances. Then، B-spline curves for groups of 4 points are calculated. Next، the slope of curvature for B-spline curves is calculated and used as feature vector of HMM classifier. The proposed method is inariant to rotation، movement and size of image because of using the slope of curvature. The results on 15 video sequences (with 150 frames in average) give recognition rate of 92. 21%.

Watermarking is used to protect copyright proof. Robustness index is the most important parameter that evaluates watermarking algorithm against different attacks such as noise and compression. In this paper، a novel semi- blind image watermarking algorithm based on joint wavelet transform (WT) and singular value decomposition (SVD) transform is proposed. In this algorithm، a new strategy is used to joint WT and SVD effectively. The most important advantage of this algorithm is robustness against a number of common attacks. Experimental results show that the proposed algorithm improves evaluation parameters more than other methods presented previously in the literature.

In this paper، grouping and recognition of online Farsi discrete characters are presented according to their structural features. The letters are divided into 9 groups based on the form and structure of their main bodies. After feature extraction، grouping is performed using a decision tree. Final recognition of letters is carried out in each group by delayed strokes. The proposed method is a rapid method in character recognition because time-consuming methods have not been used. Our proposed method was tested on TMU-OFS dataset، and a recognition rate of 94% and 92% was achieved for character grouping and recognition، respectively. The mean processing time for recognizing a letter was 3ms.

Distillation is a complex and highly nonlinear industrial process. In general it is not always possible to obtain accurate first principles models for high-purity distillation columns. On the other hand the development of first principles models is usually time consuming and expensive. To overcome these problems، empirical models such as neural networks can be used. One major drawback of empirical models is that the prediction is valid only inside the data domain that is sufficiently covered by measurement data. Modeling distillation columns by means of neural networks is reported in literature by using recursive networks. The recursive networks are proper for modeling purpose، but such models have the problems of high complexity and high computational cost. The objective of this paper is to propose a simple and reliable model for distillation column. The proposed model uses feed forward neural networks which results in a simple model with less parameters and faster training time. Simulation results demonstrate that predictions of the proposed model in all regions are close to outputs of the dynamic model and the error in negligible. This implies that the model is reliable in all regions.