International Journal of Intelligent Systems Design and Computing (14 papers in press)
Ontology Based Approach for Document Semantic Similarity Using Concept Map
by Poonam Chahal, Manjeet Singh
Abstract: Ontology plays an important role in the process of semantic similarity computation. To extract the relevant and important information from a given document it is necessary to understand the semantic associated with the document. This understanding of semantic information comes through the concepts representing the words that are present in the given documents. These concepts and relationships between these concepts are used to construct the concept map which is the first step in the construction of ontology of a document. However, to extract the concepts and their relationships, the document set of words needs to be considered. These set of words further represents the concepts which are then used in the document ontology construction process. Finally, the matching process of constructed document ontology is applied to find the semantic similarity between any two given documents. However, it can be possible that a concept present in different ontologies may give different meaning depending upon the domain of the document. In our approach of document similarity, we provide a novel method of construction of document ontology based on concept map for computation of similarity between any two given documents.
Keywords: Concept Map; Ontology; Concepts; Relationships; Semantic; Similarity.
Hardware and Software Load Power Control in Smart Home applications Based on Taguchi Optimization Technique
by Yasmina Azzougui, Abdelmadjid Recioui
Abstract: Power control is one of the concerns in smart grid implementation. The balance between the supplied power and the demand must be maintained so that blackouts are avoided. Smart meters play an important role in establishing this balance. On the other hand, power control implementation is a challenge as one would have to find the right mix between the hardware and the software parts. The purpose of this work is to optimize and implement a small power control system of home appliances. The objective of the optimization is to reschedule some tasks if the power demand exceeds a certain peak level. The optimization is based on Taguchi method which is known of its robustness and relatively fast convergence. The system casts a real life situation and can be considered as a small-scale prototype that can be extended to larger systems.
Keywords: Load side management; Software; Hardware; Optimization; Taguchi method.
A Novel Multi-Objective Hybrid QoS Optimization Technique for MANETs
by Rangaraj Jayavenkatesan, Anitha Mariappan
Abstract: Mobile Ad-Hoc Network (MANET) requires the support of optimization techniques for improving the Quality of Service (QoS) it delivers. Optimization becomes mandatory due to the autonomous and infrastructure less characteristics of the network. In order to improve the performance of the network, this manuscript formulates a hybrid optimization by assimilating Ant Colony Optimization (ACO) and Fitness Distance Ratio (FDR) based Particle Swarm Optimization (PSO). In order to ensure better QoS, ACO is aimed at optimal solution based on throughput maximization. Contrarily, the FDR PSO considers overhead and residual energy for precising the optimal solution further. The proposed hybrid technique is implemented in a scenario consisting of 100 mobile nodes and the performance of the proposed technique is investigated using the metrics throughput, packet delivery ratio, delay, overhead and energy.
Keywords: ACO; FDR PSO; QoS Routing; MANET; Hybrid optimization.
An Analysis of Adaptive Neural Networks for Speech Enhancement
by Rashmirekha Ram, Mihir Narayan Mohanty
Abstract: Adaptive algorithms are used for many different applications since a long. It has the versatile function in the area of signal processing as well as nonlinear system like Neural Networks (NN). In this paper, authors have attempted to enhance the speech signal using adaptive methods. In the first stage, Adaptive Linear Neuron (ADALINE) model is used for five different noisy speech enhancements. As it is the preliminary stage of neural network, it is used to verify the existing model. Further the same signals are used with the Deep Neural Network (DNN) model for generalized used. In both the models, four hidden layers are used to analyze the noisy speech signal. However for ADALINE case, the learning method used is weights and bias, whereas the Restricted Boltzmann Machines (RBMs) learning algorithm is used in DNN. Perceptual Evaluation of Speech Quality (PESQ) and Signal-to-Noise-Ratio (SNR) are used to verify the performance. Also these parameters are considered for the comparison purpose. The spectral components of input and enhanced speech are exhibited and presented for comparing the performance with different models. The DNN is found to be better performance in terms of measuring parameters as compared to ADALINE model. Nevertheless the ADALINE model can be omitted in such comparison to prove the adaptability for the developments of automated system.
Keywords: Neural Networks; Adaptive Linear Neuron; Deep Neural Network; Speech Enhancement; Perceptual Evaluation of Speech Quality; Signal-to-Noise Ratio.
Automatic shadow elimination in a high-density scene
by Hocine Chebi, Dalila Acheli, Mohamed Kesraoui
Abstract: This paper presents an automatic system for monitoring high-density images to estimate parameters from video sequences using a single camera. Unlike traditional methods, which can detect and monitor this behavior under the right conditions, the proposed method has a good ability to segment more specific objects (vehicles and people) by introducing a new shadow elimination algorithm. In addition, the proposed system may well address the problem of vehicle occlusions caused by shadows, which often results in failure of vehicle counting and grading. This problem is solved by a new linear shadow algorithm that uses simple tools to eliminate all undesirable shadows. Therefore, an automatic algorithm for detecting shadows is also proposed. The experimental results show that the proposed method is more robust, more precise and more powerful than other traditional methods.
Keywords: Shadow; high-density; automatic detection; video surveillance.
Investigations on Applicability of Meta-Heuristics for Survivability of WDM Optical Networks
by Himanshi Saini, Amit Kumar Garg
Abstract: Wavelength Division Multiplexing (WDM) technology have increased the optical fiber bandwidth to THz and beyond by multiplexing number of wavelengths in a fiber. Survivability is of prime concern in high-speed networks as any single failure in WDM optical networks leads to tremendous information and revenue loss. These networks demand optimum planning. In the paper, the networks are configured with pre-planned traffic model. Ant Colony Optimization (ACO), Genetic Algorithm (GA) and Simulated Annealing (SAN) meta-heuristics are applied in networks with link and node failures with an objective to minimize network congestion. The novelty of the work lies in investigations of network performance under failures with different optimization techniques and proposed traffic model. This work can assist in optimizing overall network planning of survivable networks. It is observed that GA optimisation has reduced network congestion for network under failure. Least variation in congestion for different failures is also obtained for GA meta-heuristic.
Keywords: ACO; GA; SAN; optimization; survivability; congestion.
Robust Integral Back-stepping Control for a Quadrotor: Experiments and Simulation
by Scott Cardimen, Mohamed Zohdy, Moath Sababha, Andrew Rusek
Abstract: This paper presents a robust integral back-stepping (RIB) control method applied to a constructed quadrotor prototype. The controller provides stabilization of attitude and motion under various assumptions. The controller is robust to reject external disturbances and compensate for potential parameter uncertainties. An analysis of the benefits of the proposed control method is presented by comparison to classic linear control methods. This analysis is supported by simulation and implementation on a physical prototype. The simulation parameters match that of the prototype to provide a consistent means of validation.
Keywords: quadrotor; control; robust integral back-stepping (RIB); controller.
Study of Skew Mitigation Techniques in Map Reduce Applications
by Narinder Seera
Abstract: Data skew is one of the reasons due to which MapReduce has been criticized for years. Skew occurs as a result of uneven assignment of workload to computational nodes. Many real world applications such as PageRank, CloudBurst etc severely suffer from the problem of skew which occurs either at map side or at reduce side in MapReduce model. Unfair task distribution in such applications shows the negative impact of skew on overall job execution and its performance. This study attempts to explore various types of skew, their causes and existing solutions for skew mitigation. The study observed that unfair task distribution in distributed environment leaves the potential parallelism unexploited. The paper also presents few applications which show the presence of skew and possible improvements.
Keywords: MapReduce; Skew; Straggler; Skew Mitigation; Task Distribution.
Intelligent control of nonlinear inverted pendulum system using Mamdani and TSK fuzzy inference systems: A performance analysis without & with disturbance input
by Lal Bahadur Prasad, Hari Om Gupta, Barjeev Tyagi
Abstract: Fuzzy control, an intelligent control technique uses the human expert knowledge to make control decisions. The inverted pendulum, a highly nonlinear unstable system is used as a benchmark for implementing the control methods. In this paper the modeling and control design of nonlinear inverted pendulum-cart dynamic system with & without disturbance input have been presented. Here PID control and fuzzy control methods using Mamdani type and Takagi-Sugeno-Kang (TSK) type fuzzy inference systems (FIS) have been implemented to control the cart position and to stabilize the inverted pendulum in vertically upright position. The MATLAB-SIMULINK models have been developed for simulation of the control schemes. The simulation results and performance analysis justify the comparative advantages of fuzzy control methods.
Keywords: Inverted pendulum; disturbance input; nonlinear system; PID control; fuzzy control; intelligent control; Mamdani FIS; TSK FIS.
Special Issue on: IJISDC ICCASP2018 Next-Generation Technologies in Computing, Communication and Signal Processing
Selection of Energy Efficient Path by Applying Particle Swarm Optimization Method in Wireless Sensor Networks
by Nada Al-Humidi, Girish Chowdhary
Abstract: Power and resource limitations of the sensor nodes, the possibility of packet loss and delay are the requirements should be considered while designing routing protocol for the wireless sensor network. To meet and achieve these requirements, several routing techniques have been proposed. Clustering-based routing protocol puts a network structure to satisfy energy efficiency, stability, and scalability of the network. In such protocols, the network is organized into clusters in which one node will be selected as a cluster head for the cluster. Selecting cluster head and forming the clusters are the key issues in these protocols, as a result, many routing protocols based clustering have been proposed. With the objective of solving of these issues, reducing the energy consumption and extending the lifetime of the network, in this paper, Energy Efficiency based Clustering and Particle Swarm Optimization (EECPSO) method is proposed. EECPSO performance is evaluated and justified through extensive analysis, comparison, and implementation. The results show that the proposed method is highly efficient and effective in term of balancing the consumption of energy and prolonging network lifetime.
Keywords: WSN; Routing Protocol; Hierarchical Cluster-based; Network Lifetime; Clustering; Optimization; PSO.
Web Phishing Detection: Feature Selection using Rough Sets and Ant Colony Optimization
by Ravi Kiran Varma Penmatsa, Padmaprabha K
Abstract: Abstract: Phishing has become a global issue which is doing fraud by stealing online data. Because of Phishing, many users may lose trust in online services which cause a negative effect on organizations. Predictive, Preventive and Counteractive measures taken for phishing is a crucial step towards protecting online business transactions. The accuracy of classifying any website as phished necessarily depends on the goodness of features selected. Using Feature selection algorithms combined with Optimization techniques, appropriate features can be identified. Removal of a feature should not affect the accuracy of classification. This paper proposes Rough-set and Ant colony optimization technique for attribute minimization on standardized phishing data set. Experiment results show improvement in performance with reduced attributes for web Phishing detection.
Keywords: Phishing; web phishing detection; feature selection; rough sets; ant colony optimization; classification accuracy; feature reduction.
THE DETRENDED FLUCTUATION AND CROSS-CORRELATION ANALYSIS OF EEG SIGNALS
by SUNIL HIREKHAN, Ramchandra Manthalkar
Abstract: The Detrended Fluctuation Analysis(DFA) and Detrended Cross-Correlation Analysis(DCCA) are the widely used methods for analysis of non-stationary time series. The Detrended Fluctuation and Cross-Correlation Analysis of the EEG signals Pre- and Post-meditation (mindfulness) intervention are compared. It is observed that the EEG data obtained from 7 subjects out of total 11 subjects shows reduction in the DFA and DCCA exponent values. The reduction in DFA exponent values represents the lower intrinsic fluctuations in the EEG time series, which is a measure of better (higher) complexity of these vital rhythms. The reduced DFA exponent values after 8-weeks of Focused Attention (mindfulness) meditation practice in more number of subjects, indicate that the meditation practice enhances the ability to handle complexity. The Detrended Cross Correlation Analysis(DCCA) of EEG signals shows that the EEG signals of electrode pair of two brain lobe hemispheres obeys the power-law relationship, and the DCCA exponent is reduced in value after the meditation intervention. The reduced DFA and DCCA exponents indicate improved neuronal functioning of these subjects.
Keywords: Detrended Fluctuation Analysis(DFA); Detrended Cross-Correlation Analysis; power-law correlation; mindfulness meditation; complexity indicator.
Performance Analysis of VoIP under the effect of interference and during Conference Call in WLAN Network using OPNET Modeler
by Poonam Chakraborty, Aparna Telgote
Abstract: Circuit Switched Network or Packet Switched Networks are used for both Visual and Vocal communication. Public Switched Telephone Network (PSTN) is not an affordable option when comparing with existing packet switched network. Voice over Internet Protocol (VoIP) has become a preferable alternative due to its reduced cost and flexibility compared to VoLTE(Voice over LTE) where carriers are still building out their 4G networks. However, despite its reduced cost it has to face so many challenges which affect its successful deployment. The reason is that the quality of VoIP is mainly affected by jitter, delay, packet loss and various other parameters. This research was carried out to evaluate the quality of voice in VoIP experimentally, under the effect of interference and during conference calls. The simulations were carried out using Riverbed modeler academic edition 17.5 The results of the analysis and the performance evaluation are presented in this paper. This work can guide researchers and designers to design a network for VoIP services and its deployment based on their requirements.
Keywords: Voice over Internet Protocol (VoIP); Mean Opinion Score (MOS); Jitter; interference; Conference Call.
IDENTIFICATION AND CLASSIFICATION OF HISTORICAL KANNADA HANDWRITTEN DOCUMENT IMAGES USING LBP FEATURES
by Parashuram Bannigidad, Chandrashekar Gudada
Abstract: Most of the manuscript preservation centres are working on the digitization of historical handwritten manuscript documents. Digitization process is very much essential for handwritten documents due to degraded qualities in the manuscript such as ink bleed, arbitrary geometric distortions (folding and wrapping of manuscripts), storage and maintenance condition, age, etc. It is also very much important to transfer the contents written on handwritten documents about ancient heritage culture and its evolution based on the age-type of the manuscript to our future generation. In this work, the age-type identification and classification of historical Kannada handwritten document images is done by applying text-block wise segmentation method, extracting the LBP features and using LDA, K-NN and SVM classifiers. The purpose of the present work is to identify the document script of the dynasties, whether it belongs to Hoysala dynasty or Vijayanagara dynasty or Mysore Wodeyars dynasty? The average classification accuracy of the proposed method with LDA classifier is 94.6%, whereas K-NN classifier yielded 98.3% and SVM classifier yielded 99.3%. As per the results obtained from the experimentation, it is proved that the SVM classifier has got good classification ability comparatively with respect to LDA and K-NN classifiers for historical types of Kannada handwritten document images of all age-type scripts.
Keywords: Manuscripts; Kannada; Handwritten documents; LBP features; Historical documents; Hastaprati.