Arun Khosla

QRS detection using K-Nearest Neighbor algorithm (KNN) and evaluation on standard ECG databases

The performance of computer aided ECG analysis depends on the precise and accurate delineation of QRS-complexes. This paper presents an application of K-Nearest Neighbor (KNN) algorithm as a classifier for detection of QRS-complex in ECG. The proposed algorithm is evaluated on two manually annotated standard databases such as CSE and MIT-BIH Arrhythmia database. In this work, a digital band-pass filter is used to reduce false detection caused by interference present in ECG signal and further gradient of the signal is used as a feature for QRS-detection. In addition the accuracy of KNN based classifier is largely dependent on the value of K and type of distance metric. The value of K = 3 and Euclidean distance metric has been proposed for the KNN classifier, using fivefold cross-validation. The detection rates of 99.89% and 99.81% are achieved for CSE and MIT-BIH databases respectively. The QRS detector obtained a sensitivity Se = 99.86% and specificity Sp = 99.86% for CSE database, and Se = 99.81% and Sp = 99.86% for MIT-BIH Arrhythmia database. A comparison is also made between proposed algorithm and other published work using CSE and MIT-BIH Arrhythmia databases. These results clearly establishes KNN algorithm for reliable and accurate QRS-detection.

Fuzzy controller for rapid nickel-cadmium batteries charger through adaptive neuro-fuzzy inference system (ANFIS) architecture

ANFIS architecture is a class of adaptive networks, which is functionally equivalent to fuzzy inference systems. The architecture has been employed for fuzzy modeling that learns information about a data-set in order to compute the membership functions and rule-base that best follow the given input-output data. ANFIS employs hybrid learning that combines the gradient method and the least squares estimates to identify premise and consequent parameters respectively. In this paper the fuzzy controller for rapidly charging nickel-cadmium (Ni-Cd) batteries charger has been designed through ANFIS. The behavior of Ni-Cd batteries was not known for higher charging rates and the input-output data of batteries has been obtained through rigorous experimentation with an objective to charge the batteries as quickly as possible, but without doing any damage to them. Takagi-Sugeno-Kang (TSK) model has been considered for the controller.

A Framework for Identification of Fuzzy Models through Particle Swarm Optimization Algorithm

This paper presents a framework for the identification of fuzzy models from the available input-output data through Particle Swarm Optimization (PSO) algorithm. Like other evolutionary algorithms, PSO is a population-based stochastic algorithm and is a member of the broad category of swarm intelligence techniques based on metaphor of social interaction. The suggested framework has the capability to identify optimized Mamdani and Singleton fuzzy models. For the presentation and validation of the proposed framework, the data from the rapid Nickel-Cadmium (Ni-Cd) battery charger developed by the authors has been used.

A Comparison of Computational Efforts between Particle Swarm Optimization and Genetic Algorithm for Identification of Fuzzy Models

Fuzzy systems are rule-based systems that provide a framework for representing and processing information in a way that resembles human communication and reasoning process. Fuzzy modeling or fuzzy model identification is an arduous task, demanding the identification of many parameters that can be viewed as an optimization process. Evolutionary algorithms are well suited to the problem of fuzzy modeling because they are able to search complex and high dimensional search space while being able to avoid local minima (or maxima). The particle swarm optimization (PSO) algorithm, like other evolutionary algorithms, is a stochastic technique based on the metaphor of social interaction. PSO is similar to the genetic algorithm (GA) as these two evolutionary heuristics are population-based search methods. The main objective of this paper is to present the tremendous savings in computational efforts that can be achieved through the use of PSO algorithm in comparison to GA, when used for the identification of fuzzy models from the available input-output data. For realistic comparison, the training data, models complexity and some other common parameters that influence the computational efforts considerably are not changed. The real data from the rapid nickel-cadmium (Ni-Cd) battery charger developed has been used for the purpose of illustration and simulation purposes.

Power Optimization of Linear Feedback Shift Register (LFSR) for Low Power BIST

This paper proposes a low power Linear Feedback Shift Register (LFSR) for Test Pattern Generation (TPG) technique with reducing power dissipation during testing. The correlations between the consecutive patterns are higher during normal mode than during testing. The proposed approach uses the concept of reducing the transitions in the test pattern generated by conventional LFSR. The transition is reduced by increasing the correlation between the successive bits. The simulation result show that the interrupt controller benchmark circuits testing power is reduced by 46% with respect to the power consumed during the testing carried by conventional LFSR.

Electrocardiogram beat classification using empirical mode decomposition and multiclass directed acyclic graph support vector machine

In this paper, a classifier motivated from statistical learning theory, i.e., support vector machine, with a new approach based on multiclass directed acyclic graph has been proposed for classification of four types of electrocardiogram signals. The motivation for selecting Directed Acyclic Graph Support Vector Machine (DAGSVM) is to have more accurate classifier with less computational cost. Empirical mode decomposition and subsequently singular value decomposition have been used for computing the feature vector matrix. Further, fivefold cross-validation and particle swarm optimization have been used for optimal selection of SVM model parameters to improve the performance of DAGSVM. A comparison has been made between proposed algorithm and other two classifiers, i.e., K-Nearest Neighbor (KNN) and Artificial Neural Network (ANN). The DAGSVM has yielded an average accuracy of 98.96% against 95.83% and 96.66% for the KNN and the ANN, respectively. The results obtained clearly confirm the superiority of the DAGSVM approach over other classifiers.

Markovian Queuing Model for Dynamic Spectrum Allocation in Centralized Architecture for Cognitive Radios

Dynamic Spectrum Allocation (DSA) is one of the latest technologies serving to cater to rising demand of spectra. Since a decade, cognitive radios (CR) are looked upon as a solution to increase spectral efficiency. However, to make CR a reality, development of air interface is a big challenge. In this paper, we have proposed a CR architecture followed by its equivalent mathematical model. We consider a centralized network where the master/controller of this ad-hoc network coordinates for spectrum allocation with the surrounding CR in the network. This CR ad-hoc network is assumed to coexist with the network of licensed users where the controller of licensed users is updated with CR coordinating engine. The mathematical model is formed of two Markovian distributed queues. One of these queues is modeled as M/M/1 and the second as M/G/S/N. Mathematical analysis of our proposal is focused to evaluate bandwidth access latency of an unlicensed user.

A Mobility Improvement Handover Scheme for Mobile-WiMAX

This paper presents a mobility improvement handover algorithm with less scan time implementation for Mobile WiMAX. Mobile WiMAX is a wireless technology based on IEEE802.16e for broadband wireless access. Mobile WiMAX introduces the most significant new feature, mobility to support for handovers, which can be considered as a basic requirement for mobile communication system. The mandatory handoff method is Hard Handoff for Mobile WiMAX and other two optional soft handoff methods are Macro Diversity Handoff and Fast Base Station Switching Handoff method. Handover delay generates during data transmission but handover delay should be less than 50milli second for real time applications such as VoIP. The existing draft standard considers only the received signal strength (RSS) when deciding handover. The mobility improvement handover scheme that depends upon the velocity factor has been taken into consideration in this paper. The proposed scheme skips some unnecessary handover stages, reduces handover delay by adjusting the handover parameters viz. handover RSS trigger and threshold handover RSS trigger value to improve the velocity of the mobile station. The proposed scenario has been implemented using QualNet network simulator which has lead to less handover time.

Influence of stimuli colour in SSVEP-based BCI wheelchair control using support vector machines.

Abstract This study aims to develop a Steady State Visual Evoked Potential (SSVEP)-based Brain Computer Interface (BCI) system to control a wheelchair, with improving accuracy as the major goal. The developed wheelchair can move in forward, backward, left, right and stop positions. Four different flickering frequencies in the low frequency region were used to elicit the SSVEPs and were displayed on a Liquid Crystal Display (LCD) monitor using LabVIEW. Four colours (green, red, blue and violet) were included in the study to investigate the colour influence in SSVEPs. The Electroencephalogram (EEG) signals recorded from the occipital region were first segmented into 1 s windows and features were extracted by using Fast Fourier Transform (FFT). Three different classifiers, two based on Artificial Neural Network (ANN) and one based on Support Vector Machine (SVM), were compared to yield better accuracy. Twenty subjects participated in the experiment and the accuracy was calculated by considering the number of correct detections produced while performing a pre-defined movement sequence. SSVEP with violet colour showed higher performance than green and red. The One-Against-All (OAA) based multi-class SVM classifier showed better accuracy than the ANN classifiers. The classification accuracy over 20 subjects varies between 75–100%, while information transfer rates (ITR) varies from 12.13–27 bpm for BCI wheelchair control with SSVEPs elicited by violet colour stimuli and classified using OAA-SVM.

An efficient spectrum mobility management strategy in cognitive radio networks

In this paper, we propose to design a spectrum mobility strategy using Fuzzy Logic System for cognitive radio networks (CRN). The proposed strategy enables cognitive radios (CR) to vacate the spectrum if primary user (PU) needs it back or to adjust its transmit power in order to avoid interference with PU or nearby CR. Using our scheme, CR switches between the bands only if it is not able to modify its transmit power within the tolerable interference limits. Thus, our work is divided in two modules, one with the priority to control CR transmit power within tolerable range and second to switch to another frequency band in order to avoid interference. Simulated results show that using our proposed fuzzy power control scheme, we can decrease transmit power consumption and achieve lower number of spectrum handovers.