International Journal of Intelligent Engineering and Systems

11

Authors: Usman Bashir Tayab, Muhammad Kashif

Conventional droop control is a basic control strategy for power sharing in islanded microgrid applications. This strategy has several limitations, such as low transient response, frequency and voltage deviations. This paper presents a modified droop control method for the proper operation of parallel-connected inverters in islanded microgrid. The proposed method is able to improve transient response and achieved higher output power without voltage and frequency deviations by introducing a power derivative term into a conventional droop method. A new method of average power computation is proposed which significantly reduces the delay encountered in conventional approaches and provides the power signal without DC components. The simulation was developed in MATLAB/Simulink to verify the effectiveness of the modified droop control scheme. Based on the results, it can be concluded that the modified droop control strategy improves the performance of parallel-connected power electronic inverters in an islanded microgrid as compared to conventional droop control strategy.

28

Authors: Hilal Rahali, Samir Zeghlache, Leila Benalia

This paper proposes Interval Type-2 Fuzzy Gain-Adaptive PI (IT2FGAPI) controller based on direct field oriented control (DRFOC), to control the speed of a dual star induction machine (DSIM), to get a robust performance machine. We use IT2FGAPI for speed control of the DSIM corresponding to adapting the different gains Kp and Ki corresponding the different PI logic, PI of speed, PI of flux and the four PI of currents in a vector control mode. An interval type-2 fuzzy control system is used to adapt in real-time the controller gains, to generate the PI controller parameters, he gives a fast dynamic response with no overshoot, robust and their performances are insensible to parameter variations, at different dynamic operating such as sudden change in command speed, step change in load torque and some key parameters deviation. The simulations results obtained from MatLab/Simulink are finally presented and discussed; the obtained results show that the controller in enhancing the robustness of control systems with high accuracy.

50

Authors: Narendra Kumar, Hari Shanker Shukla, Rakesh Prakash Tripathi

In this paper, we have proposed adaptive methods for image restoration in which the input images are affected by noise which is removed by fuzzy based median filter (FMF). The noise removed images from the FMF is appears to be so there is a need to restore the images with high quality. To restore the images an APSO (Adaptive particle swarm optimization) based Richardson-Lucy (R-L) algorithm is utilized. By both FMF and APSO-RL methods the denoising and restoration of the image is performed efficiently. The performance of the image denoising and restoration technique is evaluated by comparing the result of proposed technique with the existing denoising filter and GA, PSO methods. The comparison result shows a high-quality denoising and restoration ratio for the noisy images than the existing methods, in terms of peak signal-to-noise ratio (PSNR) and second-derivative-like measure of enhancement (SDME).

70

Authors: Abdul Rafiez Abdul Raziff, Md Nasir Sulaiman, Norwati Mustapha, Thinagaran Peruma

Gait signal of a person can be easily obtained using a smartphone sensor. To get the source of the signal, the smartphone need to be placed in the pocket, pouch or attached to other parts of the body. In the real world application, it is hard to place the device on the mentioned position. The easiest way is to put it on hand. In another issue, the single magnitude is known in the use of multiple orientations. However, this method may discard useful features for machine learning classification. Another problem is that the signal captured using a smartphone is not in a fix sampling rate and in the small distance, hence interpolation needs to be applied so that the sampling can be in a fix sequence with more fix point data. However, too much application of interpolation may result in low prediction rate. Finally, a multiclass dataset may contain overlapped class boundary which produces low accuracy on a single classifier mapping. In this paper, hand based smartphone placement position is implemented and evaluated. Single magnitude application is also evaluated in representing multiple positions of a person into one signal. Besides that, the linear interpolation factor is introduced in sampling the signal. Lastly, OvO classification model is implemented in binarizing the multiclass gait dataset. From the experiment, it shows that using the mentioned method do produce satisfactory result hence opening a new gateway in a better gait identification/recognition system.

91

Authors: Shubhanshi Saini, Sanjay Kumar Dubey

Jaundice is a disease caused due to alarmingly high level of bilirubin. This prevails in places with low level of sanitation and inability to access nutrient enriched diet. In the Indian perspective, jaundice is a severe menace affecting many people including infants, ladies and men. To cure jaundice, the bilirubin levels need to be lowered which can be done by a proper diet incorporating multiple critical nutrients in right proportion along with medication. This makes it an incredibly complex process. The proposed paper recommends diet to jaundice patient by using Analytic Hierarchy Process (AHP). This technique for diet recommendation has not been used before and it provides a new perspective which is different from conventional techniques. In AHP, critically important nutrients are taken as factors and alternative diet plans are evaluated on the basis of these factors in order to select the best option. The result is the diet plan which includes meals to be consumed by a jaundice patient at breakfast, lunch and dinner. The results obtained by AHP are further validated using Fuzzy Analytic Hierarchy Process (FAHP). FAHP provided the same results as obtained from AHP.

109

Authors: Marwin Kannamplavil Mathew, Rahul George John Valooran, Naveen Kumar Vaegae

In this paper, a fiber Bragg grating (FBG) based temperature sensor is designed and implemented using a conventional laser source. It is observed that the sensitivity of FBG temperature sensor is low and not effective in sensing temperature of critical applications. Hence, FBG temperature sensitization encapsulation technique is introduced to increase the sensitivity of the sensor. The sensitivity is improved by a factor of 44 with temperature sensitization encapsulation technique, but the linearity error is ±15% full scale reading. In order to retain the high sensitivity without compromising the linearity of the system, multilayer perceptron (MLP) artificial neural network (ANN) is introduced to estimate and reduce the linearity error of FBG sensor. The Levenberg-Marquardt (LM) algorithm is used for training and learning mechanism of MLP-ANN. A linearity error of approximately ±2.8% is achieved without affecting the sensitivity of the system.

126

Authors: Vanmathi Chandrasekaran, Prabu Sevugan

In this paper a reversible data hiding (RDH) algorithm for medical images is proposed. This algorithm specifically based on histogram modification in hybrid domain. The idea of this algorithm is that a histogram is created from the differences between each pixel and its neighbours. A 2D DWT haar transform is performed to convert cover image into a transform domain to select appropriate frequencies for embedding the payload. The selected DWT coefficients bit planes are used to carry the secret message. The experimental result demonstrates that our proposed method outperforms the traditional state of art methods in terms Peak Signal to Noise Ratio (PSNR), Structural Similarity Index (SSI) and the algorithm guarantees the reversibility of the host image. The proposed method avoids overflow and underflow of the pixel values and achieves highest embedding capacity.

147

Authors: Renuka Devi Saravanan, Shyamala Loganathan, Saraswathi Shunmuganathan, Yogesh Palanichamy

Distributed Denial of Service attacks are becoming a serious issue for the developers and the users of the Internet. In recent times, the attackers are targeting the online applications and web services. Detecting such application level attacks are much challenging because the attack traffic mimics the legitimate behaviour. A more sophisticated mechanism is required to detect and mitigate such attacks. In this paper, a novel method for detecting application layer Distributed Denial of Service attack is proposed. Initially, web user behaviour on different perspectives is analyzed using the system log and key dimensions that are highly responsive to attacks are identified using Principal Component Analysis. The extracted key features are analyzed to fix up the appropriate thresholds for differentiating legitimate and illegitimate access. Each incoming session is examined and if found suspicious, the detection mechanism is invoked. The detection mechanism includes a score assignment mechanism which assigns the threat score based on the history and statistical analysis of the current characteristics. The sessions having acceptable score are then scheduled to get service from the server. Remaining sessions are considered malicious and dropped. The real data sets are taken for the simulation and the results are exhibited to show the efficiency of the proposed detection method. The results show that the proposed technique performs effective detection of constant flooding and repeated shot attacks with low false positives and low false negatives.

166

Authors: Ramanan Gopalakrishnan, Edwin Raja Dhas John

Wire electric discharge machining (WEDM) has turned out to be one of the well-known machining procedures utilized in bringing up elaborate aspects on electrically conductive substances. To improve the performance of WEDM process an attempt was made to improve the material removal rate (MRR) and minimize surface roughness (SR) on AA7075-activated carbon composites. The machined surface integrity associated with MRR and SR have been analysed as a function of processing parameters, such as discharge current, pulse on time, pulse off time and servo speed rate. Analysis of variance (ANOVA) is carried out in the idea of identifying tremendous causes of parameters. Experimental outcome authorize the feasibility of proposed prediction strategy and the developed particle swarm optimization (PSO) algorithm. Experiment result indicates that the MRR and SR are more sensitive to a change in size than a change in volume fraction of reinforcement. AA7075-activated carbon composites have extensive applications in aerospace and marine sectors.

185

Authors: Firdaous Bakkali, Mourad Derra, Abdellah Amghar, Hassan Sahsah

A non-destructive ultrasonic technique presented in this paper to monitor in real time the renneted milk coagulation process in order to predict the clotting time. This one is often used as a reference to determine the cutting time of the gel, which means that the whey trapped in the gel pores has been expelled. During the enzymatic coagulation of milk, the change of the flight time has been exploited to determine the clotting time with high precision. The effects of the temperature (Standard temperature (ST), High temperature (HT) and Low temperature (LT)) and rennet concentration (Standard rennet (SR), High rennet (HR) and Low rennet (LR)) on the enzymatic coagulation were studied.

205

Authors: Sushil Kumar, Bipin Kumar Tripathi

In the last few years there have been a growing number of studies concerning the introduction of quaternions into neural networks, which demand a faster learning technique with superior performance. In this paper, we propose a fast, but novel quaternionic resilient propagation (ℍ-RPROP) algorithm for high dimensional problems. It achieves significantly faster learning over quaternionic domain back propagation (ℍ-BP) algorithm. The slow convergence and stability of weight update around the local minima are the main drawbacks of ℍ-BP. The gradient descent based ℍ-BP algorithm takes the value of partial derivative (error gradient) and scales the weight updates through a learning rate while ℍ-RPROP does not takes the value of partial derivatives, but it considers only the sign of partial derivatives that indicates the direction for each component of quaternionic weight update. The main aim of ℍ-RPROP is to eliminate the value which is a little increased by constant increasing factor in order to accelerate convergence in shallow regions. ℍ-RPROP computes an individual delta for each connection of the network, which determines the size of weight update. Therefore, the faster convergence and higher accuracy are the main key features of proposed algorithm. The intelligent behavior of the proposed learning approach is demonstrated through a wide spectrum of prediction problems with different statistical performance evaluation metrics. In order to illustrate the learning and generalization of 3D motion as its inherent behavior, a solid set of experiments is presented where the training is performed through input-output mapping over a line and the generalization ability is verified over various non-linear geometrical objects. The slow convergence problem of back-propagation algorithm has been well combated by ℍ-RPROP. It has always demonstrated drastic reduction in the training cycles.

226

Authors: Samer Hasan, Vatsavayi Valli Kumari

Cloud computing delivers computing services over the Internet based on pay as you use financial model. Cloud service providers publish service advertisements in different formats on the Internet. Thus, cloud consumers should explore all provider websites using available search engines to find the appropriate cloud service. Unfortunately, existing search engines give a huge list of unrelated results that makes consumers spend a lot of time and effort to find the best matched cloud service. In this paper, we present a layered architecture for cloud service discovery and selection system to automate cloud service discovery and selection process, and remove the barriers between cloud service providers and consumers. Additionally, we present novel numerical algorithm for cloud services matching and compare it with existing algorithms. Proposed algorithm (XNSim) is independent of any external attribute value, while existing algorithm (SNSim) depends on the max and min values of the service attribute and (MNSim) algorithm depends on the max value of the service attribute. Comparison is done based on four parameters (number of matched services, execution time, average score and recall) to find the advantages and disadvantages of each one. XNSim algorithm showed better performance and more effectiveness over MNSim and SNSim.

245

Authors: Deepa Mani, Anand Mahendran

The Fault tree furnishes an engraving and analytical structure to dissect the dependability of systems. A fault tree gives an abstractly essential demonstrating schema to embody the level of system communications among all the component reliabilities in cloud computing environment. Fault tree Investigation is that fundamental technique which concentrates on failure modes for whatever system and the probabilities for occurrence/ likelihood risk dissection connected with it. Fault Tree Analysis signifies to significant deficiency or critical failures connected to the system and reasons for the faults graphically. Here we will realise that likelihood from claiming the failure of the primary event with that assistance about which we get that probability of failure of the top event. Dynamic fault trees include a progressive thought of the conventional issue tree tactic: system failures might rely on component failure. Exceptional motivation dynamic fault tree gates can model gradual supplanting of neglected segments. Starting with pools from claiming spares, failures that happen just if others happen for certain orders, dependencies that propagate failure, also circumstances the place failures could happen best to a predefined request. With the help of the dynamic gates, the fault trees need to be tackled by automatic change with proportional Markov models regularly in a cloud environment. The proposed system uses fault tree analysis and Markov chain process to efficiently identify the repair and failures of each virtual machine and reduces the energy consumption rate and cost accordingly. The result obtained in our proposed model definitely bounds optimum solutions and deceptively specifies the triumph of an efficient fault tree analysis of failures on minimal cost and energy consumption.

265

Authors: Gopalan Meyyappan, Marthandan Govindan, Eswaran Chikkannan

In our previous work we have proposed a fuzzy based scheduling technique for multiple service flows in 802.16 WiMAX. This technique involves two entities namely evaluator and slot allocator. Where the evaluator arranges the real time and non real time service flows in two sub queues and then sorts them by their priority levels. The slot allocator obtains the priority information of service flows from the evaluator to allocate slots. For this work for we propose an extension which allocates the resources for the WiMAX Networks. This allocation is done based on the fairness of the node. During the allocation the approach considers the lifetime, transmission power, outrage probability and the power gain of the nodes. Based on these constraints the power and the bits are allocated to the source nodes in the network.

284

Authors: Ricky parmar, Ramanathan Lakshmanan

A microvascular impediment, Diabetic Retinopathy, is observed due to long term diabetes and is one of the significant reason for visual impedance because of variations in veins of the retina. Significant loss of vision due to Diabetic Retinopathy can be profoundly prevented with proper screening and medication at a more earlier stage. The evaluating procedure comprises of perceiving fine points of interest, for example, microaneurysms, to some greater elements, for example, exudates, and some of the time their position in respect to each other on pictures of the eye. Our task here is to predict the class for each of the images. We classify our images into five categories We are using Convolutional Neural Networks for our prediction model and we train our model on GPU. GPU-accelerated library of primitives aimed at Deep Neural Networks, NVIDIA CUDA Deep Neural Network (cuDNN) is used in our model. Our model has around 85% of accuracy when tested on 53576 number of retinal images. Our solution is elegant and automated, saving a lot of time and manual efforts.