Sajid Bashir

Non-linear trend estimation of cardiac repolarization using wavelet thresholding for improved T-wave alternans analysis

The phenomenon of cardiac repolarization or T-wave alternans (TWA) has attracted tremendous attention after its acceptance as a marker of malignant ventricular arrhythmias leading to sudden cardiac death. TWA manifests subtle alternation in the ST-T segment of ECG, therefore, its detection and estimation is considerably affected by deteriorated signal conditions due to noise. In this paper, we evaluate the potential of discrete wavelet transform thresholding for accurate trend estimation of ECG repolarization segment. An exhaustive experimental approach is adopted to find the optimal parameter sets for accurate trend estimation, including mother wavelets, decomposition levels and other common thresholding parameters. Validation study is carried out after shortlisting Coiflet4 and Symlet7 wavelets, subsequently applied to spectral method (SM) and modified moving average method (MMAM) for performance evaluation. For both the TWA analysis schemes, proposed method is inserted within the preprocessing stage after ST-T segmentation of ECG. When using wavelet based thresholding, SM achieves a detection gain of 3 dB in the case of Gaussian and Laplacian noises. The estimation bias and error in Gaussian noise are also improved by 40% and 62.5%, respectively, for SNR=<5 dB. Whereas, in the case of MMAM, the estimation performance improves by more than 100% for lower operating range of SNR.

Application of continuous-time wavelet entropy for detection of cardiac repolarisation alternans

Prognostic utility of microvolt T-wave alternans (TWAs) has been established since its clinical acceptance as markers for malignant ventricular arrhythmias, leading to sudden cardiac death. Accurate detection of TWA from surface electrocardiography is a challenge because of invisible nature of the phenomenon. A novel TWA detection scheme based upon analysis of continuous-time wavelet entropy (CTWE) trend of consecutive ventricular repolarisation complexes is presented. The CTWE is computed using relative wavelet energy coefficients of continuous wavelet transform. Variety of simulated alternan waveforms, wavelet functions, frequency bands and noise levels are used to test the algorithm. The algorithm achieves a sensitivity of 100% at signal-to-noise ratio (SNR) >35 dB for all the selected wavelet functions and sensitivities of 99.5, 97 and 92% for Symlet4, Mexican Hat and truncated Morlet functions, respectively, at 30 dB SNR. A performance improvement of 5 dB is achieved by only computing the wavelet coefficients at the optimal frequency band. This study concludes that CTWE can successfully characterise the heterogeneity of cardiac repolarisation and detect TWA phenomenon.

An improved statistical representation for ECG electrode movement and muscular activity noises in the context of T-wave alternan estimation

Abstract Microvolt T-wave alternans (TWA) are recognized as markers for malignant ventricular arrhythmias, leading to sudden cardiac death. Its extraordinary pathological significance and life-critical application demand elaborate modeling approaches and efficient analysis schemes. Accurate statistical model encompassing the dynamics of physiological noises and other outliers is highly significant to detection and estimation of the microvolt signal. The anomalies in parametric values characterizing the distributions of the above random phenomena are apt to incur modeling errors. Recent TWA detection theoretic approaches assume Laplacian noise due to leptokurtic distribution of electrode movement ( em ) and muscular activity ( ma ) recordings. The presented statistical analysis shows that the practiced model compromises the asymmetric nature of the probability distributions for em and ma . An analytical model called Biexponential distribution is suggested to realize the leptokurtic as well as the asymmetric nature of the noise characteristics. Comparative analysis is presented using visual inspection method, χ 2 goodness-of-fit and Monte Carlo simulations. The proposed model achieves a best match of 99.14% and 98.13% for em and ma as compared to a Laplacian fit of 95.20% and 93.84%, respectively. Conversely, the worst fit values for em and ma are found to be 96.32% and 92.45% for Biexponential and 60.47% and 15.18% for Laplacian models, respectively. The augmented degree of freedom is likely to increase the complexity of the already challenging TWA detection problem; however, the proposed model achieves a more realistic representation of the real noise data by closely matching the statistical parameters.

Extended state space recursive least squares

A new extended state space recursive least squares (ESSRLS) algorithm is proposed for state estimation of nonlinear systems. It is based on state space recursive least squares (SSRLS) approach and uses first order linearization of the system. It inherits the capability of obtaining state estimate without knowledge of process and measurement noise covariance matrices (Q and R respectively). The proposed approach is considered to provide new design option for scenarios where noise statistics and system dynamics vary. ESSRLS is initialized using delayed recursion method and a forgetting factor λ is employed to optimize the performance. The selection of λ can be problem specific as shown through experimental validations. However a value closer to and less than unity is generally recommended. Theoretical bases are validated by applying this algorithm to problems of tracking a non-conservative oscillator, a damped system with amplitude death and a signal modeled by mixture of Gaussian kernels. Simulation results show an MSE performance gain of 20 dB and 23 dB over extended Kalman filter (EKF) and unscented Kalman filter (UKF) while tracking van der Pol oscillator without knowledge about noise variances. The computational complexity of ESSRLS falls within that of EKF and UKF.

Optimizing nodes proportion for intrusion detection in uniform and Gaussian distributed heterogeneous WSN

In wireless sensor networks (WSN), intrusion detection applications have gained significant importance because of diverse implementations including tracking malicious intruder in the battlefield. Network parameters such as allowable distance, sensing range, transmission range, and node density plays important role in designing a model according to specific applications. Numerous models have been proposed to efficiently deploy WSNs for these applications. However, deviated requirements of different applications make it difficult to develop a generic model. Another important factor with significant contribution towards the performance of a WSN is the strategy adopted for distribution of the sensor nodes in the area of interest. The most common method is to deploy the sensors is either through uniform or gaussian distribution. Several performance comparisons have been reported to evaluate the detection probability and analyze its dependency on various network parameters. Another aspect fundamental to the performance of a sensor network is heterogeneity. Practically, for economic or logistic reasons, it may not be possible to ensure availability of nodes with identical features e.g. sensing range, transmission/ detection capability etc. It is, therefore, important to assess the detection performance of the network when the nodes do not possess same sensing range. In this paper we analyze the impact of various node densities in calculating detection probability in a Uniform and Gaussian distributed heterogeneous network under K-sensing model. Experimental results provide optimal values of node densities for efficient deployment in heterogeneous WSN environment.

Analysis of variation in filter length on the performance of blind equalization algorithms

Unsupervised equalizers found its use in bandwidth hungry applications (wired, wireless or optical). These equalizer show non-linear behavior in their error functions that are used to update the filter weight coefficients which makes it challenging in order to evaluate their performance in terms of ISI. In this paper, we present the performance analysis of a blind equalizer by varying its taps/length in order to optimize inter symbol interference (ISI) efficiently. Two algorithms are used for blind equalization in a modern wireless communication system i.e. Multimodulus algorithm (MMA) and Square Contour algorithm (SCA) respectively. The channel selected for simulation is frequency selective and time invariant in nature. The performance metrics are average ISI and bit error rate (BER). The overall performance is degraded in terms of average ISI by increasing filter taps using MMA and SCA as a blind equalizer. MMA is more powerful, stable and robust algorithm in terms of low SNR regimes and for increased filter length as compared to SCA respectively. The performance of MMA and SCA algorithms show similar performance behavior initially but for higher filter length, the performance of SCA is going to degrade keeping the same SNR regime.

Effect of variation in filter length on adaptive equalization in frequency selective channels

Adaptive equalizer is an integral component of modern communication systems that combats the effects of channel fading especially in multipath environment. In long term evolution (LTE) the number of multipath are standardized as seven and nine for extended pedestrian and extended vehicular-A channel respectively. The number of filter taps is an important design parameter that determines the performance and the rate of convergence of the adaptive filter. This paper analyzes the effect of the mentioned parameter on the overall performance of the system and quantifies the loss for a given range of signal to noise ratio (SNR). The performance is measured in terms of bit error rate (BER) and mean square error (MSE) both for least mean square (LMS) and recursive least square (RLS) techniques. The channel under consideration is frequency selective and time invariant in nature with an exponentially decaying power delay profile (PDP). Through simulations it is demonstrated that using RLS for channel equalization in a digital communication system it has an overall of 4 dB SNR gain as compared to LMS. Further more the convergence is almost same in case of LMS by increasing filter taps but at the cost of abrupt change in BER while in RLS the convergence time becomes slow by increasing filter length but with a small change in BER based performance.

Modified Leakage Based User Selection for MU-MIMO Systems

In multiuser MIMO systems, the spatial degrees of freedom can be effectively exploited to enhance system capacity by scheduling multiple users. This paper reviews multiuser MIMO communication from scheduling perspective, discussing performance gains, fairness, and computational complexity of leakage based scheduling algorithms. We consider well known scheduling algorithm based on user leakage power reported for MU-MIMO systems and analyze the selection criteria of first user and its impact on computational complexity of the algorithm. Through experimental validations, it is shown that if the first user is randomly selected the computational requirements are reduced with negligible performance loss. Computer simulations show that the proposed modification in the said algorithm reduces the complexity without degrading the performance.

Performance comparison of scheduling algorithms for MU-MIMO systems

Scheduling more than one user could increase system capacity of multi antenna systems. Present article reviews Multi User MIMO (MU-MIMO) communication from scheduling perspective, discussing performance gains in terms of sumrate, fairness, and computational complexity. Several scheduling techniques reported for MU-MIMO based on different algorithms are reviewed, along with partial channel state information and full channel state information. Moreover, the selection procedure of first user in leakage based and interference based scheduling algorithms is analyzed. The results show that random selection of first user does not effect the performance and have relatively less computational complexity.

Performance comparison of supervised and unsupervised equalization algorithms for frequency selective channels

Adaptive equalization are mostly used in every wireless communication systems because the wireless channel is very unpredictable and time varying in nature. This fast randomly changing channel creates many issues in wireless communication system e.g. inter symbol interference (ISI) which degrade the overall performance of wireless communication system. To mitigate these types of channel affects, adaptive algorithms are the most powerful signal processing tool which helps in the removal of ISI. In this paper we compare the performance analysis of supervised/training based equalization and unsupervised/blind equalization algorithms for a wireless communication channel having exponentially decaying power delay profile (PDP). The algorithms selected for comparison are least mean square (LMS), recursive least square (RLS), Multimodulus algorithm (MMA) and Square Contour algorithm (SCA) respectively. The performance measuring parameters are ISI and bit error rate (BER). It is observed that training based equalization algorithms i.e LMS and RLS perform efficiently in terms of removal of ISI as compared to blind equalization algorithms i.e MMA and SCA respectively. The training based algorithms tends to converge at very fast rate as compared to blind algorithms and hence suitable for fast fading channels but with the cost of additional bandwidth required for training. The blind algorithms show more computational complexity as compared to training based algorithms. Achieving the same performance level in terms of BER, an SNR loss of 7 dB is noted if we use blind algorithms instead of training based algorithms.