Majlesi Journal of Electrical Engineering
Volume:8 Issue: 1, Mar 2014

Early detection of heart diseases/abnormalities can prolong life and enhance the quality of living through appropriate treatment. Whereas clustering of electrocardiogram (ECG) signals will help to identification of heart diseases as soon as possible. In this regard, neural network and fuzzy logic have been used in many application areas while each of them has advantages and disadvantages. Thus, the present paper utilizes the proposed fuzzy neural network (FNN) with initial weights generated by genetic algorithm (GFNN) for the sake of improvement training speed, accurate and to reduce the chance of the FNN getting stuck on a local minimum.Four types of ECG beats (normal beat, congestive heart failure beat, ventricular tachyarrhythmia beat and atrial fibrillation beat) obtained from the PhysioBank database was clustered by the proposed GFNN model. Model evaluation results indicate that the proposed model can perform more accurately and less training speed than the conventional statistical methods, a single ANN and FNN. The total clustering accuracy of the GFNN model is 98.23%.

In an information retrieval system, the query can be made by user sketch. The new method presented here, optimizes the user sketch and applies the optimized query to retrieval the information. This optimization may be used in Content-Based Image Retrieval (CBIR) and Content-Based Video Retrieval (CBVR) which is based on trajectory extraction. To optimize the retrieval process, one stage of retrieval is performed by the user sketch. The retrieval criterion is based on the proposed distance metric from the user query. Retrieved answers are considered as the primary population for evolutionary optimization. The optimized query may be achieved through reproducing and minimizing the proposed measurement by using Genetic algorithm (GA). The optimized query could then be used for the retrieval of concepts from a given Data Base (DB). The proposed algorithms are evaluated for trajectory retrieval from urban traffic surveillance video and image retrieval from a DB. Practical implementations have demonstrated the high efficiency of this system in trajectory retrieval and image indexing.

Prostate cancer is one of the leading causes of death by cancer among men in the world. Ultrasonography is said to be the safest technique in medical imaging so it is used extensively in prostate cancer detection. In the other hand determining of prostate’s boundary in TRUS (Transrectal Ultrasound) images is very necessary in lots of treatment methods prostate cancer. So first and essential step for computer aided diagnosis (CAD) is the automatic prostate segmentation that is an open problem still. But the low SNR, presence of strong speckle noise, Weakness edges and shadow artifacts in these kind of images limit the effectiveness of classical segmentation schemes. The classical segmentation methods fail completely or require post processing step to remove invalid object boundaries in the segmentation results. This paper has proposed a fully automatic algorithm for prostate segmentation in TRUS images that overcomes the explained problems completely. The presented algorithm contains three main stages. First morphological smoothing and sticks filter are used for noise removing. A neural network is employed in second step to find a point in prostate region. Finally in the last step, the prostate boundaries is extracted by GVF active contour. Some experiments for the performance validity of the presented method, compare the extracted prostate by the proposed algorithm with manually-delineated boundaries by radiologist. The results show that our method extracts prostate boundaries with mean square area error lower than 4.4%.

In the modern context of electricity market deregulation, the price of the kilowatt-hour must take both power injections and withdrawals of the multiple market participants into consideration, as well as their actual grid usage. The responsibility for causing transmission losses and voltage drops therefore needs to be fairly attributed. While grid power injections and withdrawals are unequivocally attributed, it remains to date impossible to naturally share responsibility for transmission losses. Relevant literature proposes a variety of methods. This paper proposes a new method for allocating transmission losses to market participants using the network. The overall grid losses are obtained from summing the difference between injected and withdrawn power for all nodes. A set of allocation factors derived from the electrical distance between concerned buses and their voltage levels is used to attribute active power loss to each bus, after the losses arising from the mutual influencing between buses has been calculated. This method focuses on busbar current injections and assumes there is a hypothetical power flow between nodes. For mutual influencing, one of the busbars is considered a generator and the other a load. A reference bus voltage is set and then the load side is penalized depending on how far its own voltage is lower than that reference value. Results from a sample network are compared to those of previous methods.

Due to artifacts, brain magnetic resonance image (MRI) segmentation is a complicated concern. This research work presents an image segmentation approach for brain magnetic resonance (MR) images. The proposed method is based on multi dimensional fuzzy C-mean. In this technique, different features of neighbouring pixels such as mean, standard deviation and singular value are extracted and then a multi dimensional feature vector is created in feature selection stage in which the best combination of extracted features is used. The created feature vector is used as an input to the multi dimensional FCM. The results have been evaluated with manual segmentation on two publicly available datasets

Optimal sliding mode guidance (OSMG) law is proposed for tactical missiles pursuing maneuvering targets in three-dimensional space. The most important characteristic of sliding mode control (SMC) is high robustness against parameter variations or external disturbances. Using a Lyapunov function, it is demonstrated that the derived guidance law can stabilize the engagement system. Also, coefficients of controller are chosen using genetic algorithm optimally. Compared with traditional augmented proportional navigation guidance (APNG) law, the proposed guidance law not only can increase robustness against external disturbance and eliminate the effect of target maneuvers but can improve tracking performance and reduce interception time and miss distance. 3-D missile-target engagement is simulated for different target maneuvers and for some various scenarios, then results of the OSMG law are compared with conventional APNG law. Simulation Results confirmed the above mentioned pretensions.

Since development of wind power plants installation is growing, problems related to network connecting, stability and voltage effects become more important. On the other hand, wind farms are often open to lightning because of their long height and specific appearance. In this paper modeling and simulation of 1-phase, 3-phase and lightning faults in a wind farm consisting of 40 wind turbines and faults impact on wind farm and the network is investigated in EMTP-RV environment. In this field it’s necessary to develop a precise modeling out of wind power plant in order to evaluate the effects of these power plants on dynamical behavior of the power system. These models can be used in designing new protection systems, new protection algorithms, and new strategies for power plants exploitation improvement. Each wind unit in the farm is connected to the whole units connected to the network using a doubly fed induction generator (DFIG).

in this paper, embedding a watermark image is performed in discrete Contourlet transform (CT) domain. It is known that the CT is able to capture the directional edges and contours superior to discrete wavelet transform. The proposed scheme is based on embedding watermark bits into the singular value of the selected blocks within low-pass sub-band of the original gray image CT. Two algorithms, the first uses adaptive quantization and the second uses quantization step, are implemented. Our experimental results show that the method with quantization step has better fidelity in terms of peak signal to noise ratio (PSNR) and is more robust against geometrical and non-geometrical attacks in terms of normalized cross correlation (NC) in comparison with the first method, though, the method with adaptive quantization is blind in the sense only quantization strategies but not the original image is required.

Reactive power dispatch plays a key role in secure and economic operation of power systems. Optimal reactive power dispatch (ORPD) is a non-linear optimization problem which includes both continues and discrete variables. Due to complex characteristics, heuristic and evolutionary based optimization approaches have became effective tools to solve the ORPD problem. In this paper a new optimization approach based on improved differential evolution (IDE) has been proposed to solve the ORPD problem. IDE is an improved version of differential evolution optimization algorithm in which new solutions are produced in respect to global best solution. In the proposed approach, IDE determines the optimal combination of control variables including generator voltages, transformer taps and setting of VAR compensation devices to obtain minimum real power losses. In order to demonstrate the applicability and efficiency of the proposed IDE based approach, it has been tested on the IEEE 14 and 57-bus test systems and obtained results are compared with those obtained using other existing methods. Simulation results show that the proposed approach is superior to the other existing methods.

In this article the electrical parameters of the induction motor are estimated by measuring temperature of its components. This process consists of four sections: 1-Thermal resistance model 2- Electrical parameters 3- Electrical model and 4- Neural network which presents motor parameters and operational status compatible with corresponding learning data. In thermal resistance model, we use physical and geometrical properties of motor components to formulate thermal resistance of each component of the motor. In electrical model, electrical losses are calculated by electrical variables, by relationship between electrical losses and thermal resistance, these modeles gives us corresponding temperatures as output. These temperatures were used for training network (temperatures as input data, stator current and maximum torque as target data) in the neural network. After measuring temperatures of the components and using them as inputs to neural network, the corresponding stator current and maximum torque (target data) are estimated. All the stator currents referred to this paper were validated by experimental measurements.

In this paper for the first time, a circuit model for multi-finger I-gate body-contacted silicon-on-insulator MOSFET is presented. The model parameters are adjusted using simulation of a 45 nm SOI nMOSFET. Using the model, typical body voltage for a 35 finger device is obtained and applied to the transistor. The threshold voltage and drain current are obtained for the first transistor and center ones in the multi-finger structure. The drain induced barrier lowering of the center transistor is increased by 30% and off-current 40 times than that of the first transistor. Simulation results verified the I-gate body contact model in lack of controlling the body voltage comparing with the conventional body contacts e.g. H-gate.