Vijander Singh

ANN-based estimator for distillation using Levenberg-Marquardt approach

In modern chemical industries the purity of the distillate is the main objective and time to estimate the distillate composition is also the constraint. In the present paper, the Levenberg-Marquardt (LM) approach is proposed for predictive inferential control of distillation process. The developed estimator using LM approach predicts the composition of distillate using column pressure, reboiler duty, and reflux flow along with the temperature profile of the distillation column as inputs. In complex chemical industries where the output depends on many parameters, Steepest Descent Back Propagation (SDBP) algorithm does not work properly for estimating the composition of distillate, which results in saturated outputs and differs from the desired results. To overcome such type of situation, LM approach is used in developed estimator. The estimated results are compared with the simulation results and it is observed that the results obtained from LM approach are significantly improved than the results obtained from SDBP algorithm. To enhance the accuracy of the estimated results, the pressure, reflux flow and heat input with temperature profile of the column are used as input to train the neural network.

Prospects and limitations of non-invasive blood glucose monitoring using near-infrared spectroscopy

Abstract Diabetes is a chronic metabolic disorder which can lead to severe complications and affect all vital organs. The occurrence of complications due to diabetes can be prevented by regular monitoring and maintaining the blood glucose level in the normal range. Most of the commercially available devices for glucose measurement are invasive or minimally invasive. Invasive devices used for blood glucose monitoring are inconvenient and painful whereas minimal invasive devices have limited time span and stability. Thus, there is a need of an economic, compact, painless and convenient non-invasive device which can promote frequent blood testing which help in control of blood glucose level. In this paper various methods of glucose monitoring are reviewed and overall emphasis is laid on the development of NIRS (near-infrared spectroscopy) based non-invasive glucose monitoring. The motivation of this review is to demonstrate the prospects, limitations and technical challenges for development of NIRS based non-invasive blood glucose measurement system.

Development of soft sensor for neural network based control of distillation column

The present work is aimed at the design of Levenberg-Marquardt (LM) and adaptive linear network (ADALINE) based soft sensors and their application in inferential control of a multicomponent distillation process. Further the ADALINE sensor is trained online using past measurements, to adapt the changes in the inputs and is termed as dynamic ADALINE (D-ADALINE) sensor. The soft sensors are then used in the control loop to obtain LM based inferential controller (LMIC), ADALINE based inferential controller (ADIC) and D-ADALINE based inferential controller (DADIC) for the process. The performance of dynamic controller is also analyzed for different inputs and sampling intervals. The comparison of results shows the efficient and robust prediction capability of D-ADALINE sensor and hence DADIC proves to be the best controller.

An efficient approach for face recognition based on common eigenvalues

In this paper, a simple technique is proposed for face recognition among many human faces. It is based on the polynomial coefficients, covariance matrix and algorithm on common eigenvalues. The main advantage of the proposed approach is that the identification of similarity between human faces is carried out without computing actual eigenvalues and eigenvectors. A symmetric matrix is calculated using the polynomial coefficients-based companion matrices of two compared images. The nullity of a calculated symmetric matrix is used as similarity measure for face recognition. The value of nullity is very small for dissimilar images and distinctly large for similar face images. The feasibility of the propose approach is demonstrated on three face databases, i.e., the ORL database, the Yale database B and the FERET database. Experimental results have shown the effectiveness of the proposed approach for feature extraction and classification of the face images having large variation in pose and illumination.

Near-infrared LED based non-invasive blood glucose sensor

Diabetes is a metabolic pathological condition of concern, which affects vital organs of body if not diagnosed and treated on time. Regular monitoring of blood glucose is important to avoid complication of diabetes. Commonly used glucose measurement methods are invasive which generally involves finger puncturing. These methods are painful and frequent pricking cause calluses on the skin and have risk of spreading infectious diseases. Therefore there is need to develop a non-invasive monitoring system which can measure blood glucose continuously without much problem. The present work is focused on development of non-invasive blood glucose measurement sensor system using Near-infrared (NIR) technique. Initially in-vitro glucose measurement prototype is developed using continuous wave (CW) from NIR LED (940 nm) to check the sensitivity of the system for different glucose concentrations. Later a Sensor patch was designed using LED and a photodiode to observe diffused reflectance spectra of blood from the human forearm. Diffused reflectance spectra of the subjects obtained with this technique was also compared with commercially available invasive finger tip gluco-meter. The results are promising and show the potential of using NIR for glucose measurement.

A New Particle Swarm Optimization with Quadratic Interpolation

In this paper we have presented a new variant of Basic Particle Swarm Optimization (BPSO) algorithm named QIPSO for solving global optimization problems. The QIPSO algorithm makes use of a multiparent, quadratic crossover/reproduction operator defined by us in the BPSO algorithm. We have compared it with Basic Particle Swarm Optimization and the numerical results show that QIPSO outperforms the BPSO algorithm in all the sixteen cases taken in this study.

Differential Evolution using Quadratic Interpolation for Initializing the Population

The performance of population based search techniques like Differential Evolution (DE) depends largely on the selection of initial population. A good initialization scheme not only helps in giving a better final solution but also helps in improving the convergence rate of the algorithm. In the present study we propose a novel initialization scheme which uses the concept of quadratic interpolation to generate the initial population. The proposed DE is validated on a test bed of 10 benchmark problems with varying dimensions and the results are compared with the classical DE using random initialization, DE using opposition based learning for generating the initial population. The numerical results show that the proposed algorithm using quadratic interpolation for generating the initial population accelerates the convergence speed quite considerably.

Identification of optimal wavelet-based algorithm for removal of power line interferences in ECG signals

Identification of optimal denoising algorithm for ECG signals using DWT is described. The parameters are determined by comparing different types of wavelet functions, decomposition levels and threshold selection methods. The algorithm is used to improve the SNR of the ECG contaminated by disturbances like power line interferences, to present a clean signal for accurate auto diagnosis.

Optimal algorithm for ECG denoising using Discrete Wavelet Transforms

The ECG signal is used for various medical diagnoses. For diagnosis purposes, the ECG signal must be free from the noise and undesired disturbances. In the present work the best algorithm for de-noising the ECG signal is identified using Discrete Wavelet Transform. Statistical analysis and comparison has been made to find the best wavelet function, their decomposition level and threshold selection method to be used for signal de-noising and thus improve the Signal to Noise Ratio.

Performance Evaluation and Implementation of FPGA Based SGSF in Smart Diagnostic Applications

The main objective of the paper is to implement Savitzky Golay Smoothing Filter (SGSF) so as to apply in pre-processing of real time smart medical diagnostic systems. As very important information of EEG and ECG waveforms lies in the peak of the signal, hence it becomes absolutely necessary to filter noise and artifacts from the signal. The implemented filter should be able to reject the noise efficiently along with the least distortion from the original signal. The shape preserving characteristics of the filter are determined by introducing different noise levels in the signal. The designed filter is tested on synthetic signals of EEG and ECG by adding different types of noise and the performance is analysed on various parameters, i.e., SNR, SSNR, SNRI, MSE, COR and signal distortion of the final output. The smoothing performance comparison of SGSF with the most commonly used Moving Average Filter (MAF) proves that SGSF is more efficient. Hence it is suggested that MAF can be replaced by SGSF. For real time issues, it is further implemented on reconfigurable architectures so as to achieve high speed, low cost, low power consumption and less area. Therefore SGSF is realized on FPGA platform to combine the advantages of both. Real time EEG and ECG signals are also considered for experimentation. The experimental results show that the proposed methodology (FPGA-SGSF) significantly reduces the processing time and preserves the actual features of the signal.