Neelam Rup Prakash

A neuro-genetic algorithm approach for solving the inverse kinematics of robotic manipulators

The inverse kinematics solution of a robotic manipulator requires the solution of non-linear equations having transcendental functions and involving time-consuming calculations. Artificial neural networks with their massively parallel architecture are natural candidates for providing a solution to this problem. In this work, a neuro-genetic algorithm approach is used to obtain the inverse kinematics solution of a robotic manipulator. A multi-layered feed-forward neural network architecture is used. The weights of the neural network are obtained during the training phase using a real-coded genetic algorithm. This training algorithm does not suffer from the usual drawbacks of the backpropagation learning algorithm. The approach is used to obtain the inverse kinematics solution of a planar robotic manipulator.

Prolonging Network Lifetime Using Ant Colony Optimization Algorithm on LEACH Protocol for Wireless Sensor Networks

This paper presents a fair comparison of Low energy adaptive clustering hierarchy (LEACH) and ant colony applied on LEACH on the basis of the death of first node in the wireless sensor network (WSN) and data transfer. The simulation results show that when the ant colony algorithm is applied on the existing LEACH protocol the results show significant improvement in the network lifetime by delaying the death of first node in the WSN and thus increasing the efficiency of the system.

An augmented image gradients based supervised regression technique for iris center localization

This paper describes a robust and accurate technique for iris center localization by combining supervised regression based approach and image gradients. The proposed work consist of two stages. The first stage comprises regression approach which is based upon learning of local binary features to detect the periocular regions. In the second stage, image gradients were applied to the extracted eye patch regions to detect the accurate iris centers. The proposed augmented image gradients based supervised regression approach tested on the two publicly available challenging datasets show good accuracy. The results proved that supervised regression technique when augmented with image gradients approach improved the accuracy of iris center detection on the face image acquired under unconstraint conditions. The outcome of the proposed work suggests that by augmenting effective unsupervised techniques such as image gradients improves the accuracy and robustness of the supervised approaches used for face alignment applications. This work may be extended towards the development of accurate and fast eye gaze tracking systems.

An Efficient IFIR Filter Based Prototype Filter Design for Cosine Modulated Transmultiplexers

Cosine modulated filter banks and transmultiplexers are widely used because of the ease of implementation, since they have the advantage of higher obtainable attenuation in the stopband. But this leads to certain drawbacks because it would mean longer length filters and longer delays. Longer length filters would mean that that computational complexity increases. The IFIR approach to filter design is known to yield great savings in computation. In this paper an earlier approach to prototype filter design based on the Kaiser window is revisited and the same is implemented using an IFIR approach. Saving in computation will be highlighted and any deviation in performance from the previous approach will be determined. Factors governing choice of the stretch factor will be determined

A new algorithm for detection of Atrial Fibrillation

The main objective of this paper is to discuss about alarming increase in deaths and disability due to Atrial Fibrillation (AF), less efficiency in detection of AF, algorithms to detect AF and in the end a new algorithm is proposed to detect AF. This algorithm is also tested for its performance analysis using pre-recorded ECG signals. Atrial Fibrillation (AF) is a common Coronary Artery disease (CAD). In AF, the heart beats irregularly and rate of rhythm of heart becomes asynchronous as compared to rate of rhythm of heart of a healthy person. AF has higher rate of mortality and quality of life also worsens due to frequent hospitalization. With such high rate of mortality and less efficiency in detection of AF through Electrocardiogram (ECG) & even Holter, this cardiovascular disease is becoming threatening to lives of masses. Health Professionals and Scientists are working in synergy to find different ways to detect the AF cases in advance by developing the algorithms which can detect AF. An ECG signal is formed due to electrical activity of heart. Various Signal processing algorithms developed by research scientists can be applied upon open ECG databases, to detect the various parameters of ECG signals. These algorithms are tested on various open ECG and AF databases to find out their specificity and sensitivity. Algorithms for analysis of signals can be broadly classified in three different domains which are: (i) Time Domain Analysis, (ii) Frequency Domain Analysis and (iii) Non-Linear Analysis. In this paper, All of these techniques are briefly explained and a new technique is proposed for detection of AF. This technique involves statistical analysis of RR intervals and continuous slope analysis using four sample window of ECG signal to detect the presence of normal or abnormal P wave.

EDA wavelet features as Social Anxiety Disorder (SAD) estimator in adolescent females

Social Anxiety Disorder(SAD) effects individuals social behaviour and results in excessive self-consciousness, negative judgmental thoughts and uncontrollable fear. It is visible not only in behavior but also pattern of physiological signals (such as electrodermal activity) of individuals as it is associated with autonomic nervous system (ANS). Previous studies have used various features of Electrodermal Activity (EDA) such as Mean SCR, Min SCR, Range, Slope and Max SCL etc to distinguish between groups of anxious and control group subject during rest and anxious task/situations. This research explores the use of EDA wavelet features to estimate the social anxiety disorder of female subjects via Multi Layer Perceptron (MLP). In this study joint time-frequency domain features of EDA signal via wavelet analysis were extracted. The Backward regression model with p<;0.05 was used in this study for feature selection. The machine learning algorithm developed in this research was able to classify the SAD with accuracy of 82.3% during training, 85.7% during testing and 80% in holdout cases.

Design of low power multiple output power supply for embedded systems

This paper presents the design of multiple output power supply from AC power supply that is used as an input to this power supply system. As an Embedded system works on different voltage levels because the sub modules used in any embedded system required different voltage levels as per their specifications, so power supply proposed in this paper produces +15V, -15V, +5V, -5V, +3.3V, +5V Isolated voltage levels. The input for this power supply is 220 Volt AC supply. This AC supply is converted into DC supply of 48 volts but this range can be 36-72 volts. This power supply is capable of providing approximate 41W output power with 85% efficiency. The power supply design topology selected for this paper is flyback converter because flyback converter is an efficient method for generating multiple output voltage levels. This power supply is designed by utilizing low cost, highly efficient discrete components. This supply designed here can be used for almost all applications due to its versatile range of constant output voltage levels.

Intelligent planning of trajectories for pick-and-place operations

Intelligent industrial robots, unlike their current re-programmable counterparts, should be able to work in a time varying environment and must have the capability to respond to unanticipated situations. The work presented is concentrated on the trajectory planning phase of an online prediction planning execution strategy for such situations. A neural network based solution for the generation of time based control set points is proposed. This solution uses a feedback neural network, which can be trained using the extensively used backpropagation algorithm. The proposed neural based trajectory generator ensures a real time solution consistent with electric actuator safety requirements. The generator is evaluated for pick-and-place operations of a RRR manipulator.

Audio-video emotional response mapping based upon Electrodermal Activity

Abstract In this paper, a machine learning algorithm is proposed for emotional pattern recognition during audio-visual stimuli (music videos) using Electrodermal Activity (EDA). For emotion prediction apart from conventional time domain features of EDA signal, various features in different signal representation i.e. frequency and wavelet were analysed. The comparative result indicated that the wavelet features subset outperformed the conventional time domain features in term of classification accuracy. For identification of optimal network configuration, various combination of optimization algorithms (i.e. backpropagation algorithms) and error function were explored. The best performance of 79% for arousal, 69.8% for valence and 71.2% for dominance were obtained for emotion recognition respectively.

An Accurate and Simple Approach to Detect Eye Centers in Low Resolution Face Images

ABSTRACT This paper introduces a new approach for accurate eye center localization based upon optimized image gradients algorithm. The proposed approach requires pre-fetched feature descriptors to detect the accurate iris centers amongst the possible eye center candidates computed by image gradients over the face image data-set. The results of the proposed algorithm for detecting eye, iris, and iris centers over the test set images of publicly available low resolution and challenging data-set show an accuracy percentage of 98.9, 95.7, and 89.2, respectively. The comparison results obtained by the proposed approach were at par with the state-of-the-art techniques involving complex calculations and training requirements. So, the superior performance and outcome of the proposed approach show the usefulness of optimizing the results of simple image gradients by pre-fetched Scale Invariant Feature Transform feature descriptors in detecting iris centers under unconstrained environments. The proposed approach may be useful for the development of real-time eye gaze tracking application with improved robustness and accuracy.