Palash Kumar Kundu

Electronic Tongue System for Water Sample Authentication: A Slantlet-Transform-Based Approach

This paper proposes the development of a new approach for water sample authentication, in real life, using a pulse-voltametry-method-based electronic tongue instrumentation system. The system is developed as a parallel combination of several neural network classifiers, each dedicated to authenticate a specific category of water sample, and can be extended for more categories of water sample authentication. The system employs a slantlet-transform (ST)-based feature extraction module and two popular variants of neural networks for classification. The proposed system hybridizes ST with two variants of backpropagation-neural-network-based binary classifiers to develop an automated authentication tool. ST is regarded as an improved version of orthogonal discrete wavelet transform that can provide improved time localization with simultaneous achievement of shorter supports for the filters. This proposed system, implemented in a laboratory environment for various water samples available in India, showed encouraging average authentication percentage accuracy, on the order of over 80% for most water categories and even producing accuracy results exceeding 90%, for several categories.

Classification and authentication of unknown water samples using machine learning algorithms

This paper proposes the development of water sample classification and authentication, in real life which is based on machine learning algorithms. The proposed techniques used experimental measurements from a pulse voltametry method which is based on an electronic tongue (E-tongue) instrumentation system with silver and platinum electrodes. E-tongue include arrays of solid state ion sensors, transducers even of different types, data collectors and data analysis tools, all oriented to the classification of liquid samples and authentication of unknown liquid samples. The time series signal and the corresponding raw data represent the measurement from a multi-sensor system. The E-tongue system, implemented in a laboratory environment for 6 numbers of different ISI (Bureau of Indian standard) certified water samples (Aquafina, Bisleri, Kingfisher, Oasis, Dolphin, and McDowell) was the data source for developing two types of machine learning algorithms like classification and regression. A water data set consisting of 6 numbers of sample classes containing 4402 numbers of features were considered. A PCA (principal component analysis) based classification and authentication tool was developed in this study as the machine learning component of the E-tongue system. A proposed partial least squares (PLS) based classifier, which was dedicated as well; to authenticate a specific category of water sample evolved out as an integral part of the E-tongue instrumentation system. The developed PCA and PLS based E-tongue system emancipated an overall encouraging authentication percentage accuracy with their excellent performances for the aforesaid categories of water samples.

The e‐tongue‐based classification and authentication of mineral water samples using cross‐correlation‐based PCA and Sammon's nonlinear mapping

The present paper elaborates on the design of classifiers based on cross‐correlation‐based principal component analysis (PCA) and Sammons nonlinear mapping (NLM) using current signals obtained from electronic tongue (e‐tongue) with commercial mineral water samples available in the Indian market. The pulse‐voltammetric method is used to capture the electroanalytical/electrochemical characteristics of the sampled mineral waters by considering a real model for the liquid–electrode interface in a given e‐tongue apparatus. Then the cross‐correlation coefficients between the output and input signals are determined. Both PCA and Sammons NLM create a subspace from high‐dimensional mineral water data by considering the principal eigenvectors and minimising the stress function, respectively. The proposed cross‐correlation‐based PCA and Sammons classifiers establish the highest separation distance among the investigated water brands and carries out the authentication of more than one unknown sample of the same brand with a certain degree of variability with respect to their sources. Copyright

Modified DCSK communication scheme for PLCC based DAS with improved performance

Differential Code Shift Keying (DCSK) scheme developed in MATLAB environment require longer execution time at decoding end particularly with longer strings. The problem can be avoided by using suitable data compression technique. In this paper, Huffman Coding based data compression technique is implemented for improving the performance of DCSK Communication scheme developed for integer strings which can be extended for any character string with slight modification.

A novel DCSK communication scheme for PLCC based DAS

With the increasing popularity of power line communication technologies, modifications over Spread Spectrum Technologies were made so as to develop a suitable alternative for power line communication. Differential Code Shift Keying (DCSK) is one such technology developed by Yitran Technologies & patented on May 16th, 2000. In this paper, it is proposed to develop this DCSK Communication Scheme which can be used for PLCC system for DAS application.

Dissimilarity factor based classification of inferior myocardial infarction ECG

Electrocardiography (ECG) is popular non-invasive technique for preliminary level investigation on cardiovascular assessment. Computerized analysis of ECG can significantly contribute towards assisted diagnosis and in early detection of many cardiac diseases. Conventional automated ECG classifiers employing soft computing tools may suffer from the inaccuracies that may result in different clinical feature extraction stages. In this paper, we propose the use of a statistical index, namely, dissimilarity factor (D) for classification of normal and Inferior Myocardial Infarction (IMI) data, without the need of any direct clinical feature extraction. Time aligned ECG beats were obtained through filtering, wavelet decomposition processes, followed by PCA based beat enhancement to generate multivariate time series data. The T wave and QRS segments of IMI datasets from Lead II, III and aVF were extracted and compared with corresponding segments of healthy patients using Physionet ptbdb data. With 35 IMI datasets, the average composite dissimilarity factor Dc between normal data was found to be 0.39, and the same between normal and abnormal data were found to be 0.65. This paper shows the promise of descriptive statistical tools as an alternative for medical signal analysis.

A novel data compression approach to reduce string size for PLCC based DAS

MATLAB based DCSK communication scheme developed for PLCC based DAS requires longer time of execution at receiving end particularly with longer strings. In order to avoid the problem, basic arithmetic coding based lossless data compression method is proposed which can reduce the string size of string and consequently improve the performance of the system significantly. As conventional DAS deals mostly with integer values only, the algorithm is developed for integer string only. But this proposed method can be modified suitably and can be implemented for character strings successfully.

Performance analysis of SINEMO and NEMO BSP for vehicle-infrastructure-vehicle communications

Network mobility (NEMO) basic support protocol (BSP) and seamless IP diversity based NEMO (SINEMO) have been designed to provide seamless and uninterrupted services to the mobile hosts in NEMO. SINEMO outperforms NEMO BSP by utilizing the advanced loss recovery and multi-homing feature of stream control transmission protocol (SCTP). To improve the performance of NEMO BSP, we could utilize the services of on-board TCP (obTCP) which implements local retransmission mechanism in the wireless links to recover wireless losses quickly and effectively. In this paper, we analytically compare NEMO BSP along with obTCP and SINEMO based on end-to-end packet loss probability, end-to-end packet delivery delay, handoff latency, and throughput degradation time during handoff. The numerical results show that NEMO BSP when used with obTCP performs better than SINEMO.

PID controller design for an interacting tank level process with time delay using MATLAB FOMCON toolbox

This paper presents a new way to design PID controller for both integer order and fractional order with a time delay for a typical interacting cylindrical tank system using MATLAB FOMCON toolbox. Here, our work aims to study the performance characteristics of integer order and fractional order PID controller on the current integer order plant obtaining minimum objective function by Nelder - Mead optimization technique with different performance metrics ISE, ITSE and IAE. Next our work shows to make comparison between integer order PID controller based on AMIGO model performance and fractional order PID controller on time domain characteristics. The proposed method aims finally to analyze overall desired performance on fractional order PID controller by adding two extra degrees of freedom over the integer order PID controller with different performance criteria.

Electrocardiogram synthesis using Gaussian and fourier models

Electrocardiogram (ECG) is an important tool for investigation of cardiac functions. ECG synthesis or modeling can be useful for biomedical applications involving data compressions, signal analysis and testing of medical systems. In this paper, we present a morphological modeling method of single lead ECG by two different approaches, viz., Fourier and Gaussian models. Single lead ECG data was preprocessed to remove unwanted noise and segmented in three zones, P-R, Q-R-S and S-T. The individual segments were then modeled to extract model coefficients. The residual of each segment, computed as difference between original and reconstructed samples were also modeled using Fourier model. The algorithms were validated with lead V2 from normal and Anterior Myocardial Infarction (AMI) data from Physionet. With 25 AMI datasets, the average PRDN and SNR were found to be 9.33 and 20.71 respectively with Gaussian model and 5.43 and 22.16 respectively with Fourier model. The Fourier model showed better reconstruction performance, but less memory efficient compared to the Gaussian model.