Sign LMS Research Articles

Comparison of the performance of artificial neural network with variable step-size adaptive algorithms for the beamforming of smart antenna for cellular networks

A smart antenna is an antenna array that uses spatial diversity to identify the desired mobile station (MS) and reject the unwanted interference signal in a cellular network. Generally, adaptive signal processing algorithms are used for smart antenna beamforming, and one of the most common algorithms is the least mean square (LMS) algorithm. Here, the artificial neural network (ANN) is used for beamforming of smart antennas, and the performance of the ANN is compared with the performance of variable step-size LMS (VS-LMS) and variable step-size sign LMS (VS-SLMS) algorithms. The ANN has better performance than the VS-LMS and VS-SLMS algorithms for the determination of user and null directions. Lower side lobe levels (SLLs) are achieved using ANN compared to the VS-LMS and VS-SLMS algorithms. The reduction of SLL from about 3.5 dB to 8.5 dB is achieved using ANN compared to signal processing algorithms.

A novel heuristic for the analysis of block sign LMS algorithm

AbstractThis paper presents a new heuristic for the analysis of adaptive filtering algorithms. It combines Price's Theorem (which is strictly valid when the random variables are jointly Gaussian) with a probabilistic model of the input data that assumes statistical independence between the radial and angular distributions. Moreover, the last distribution is modeled as discrete, which allows obtaining concise and useful closed‐form estimates of the algorithm's performance. The proposed method directly provides a derivation of a traditional result for the steady state performance of the Sign Least Mean Squares algorithm. Furthermore, it can be used to gain new insights into the asymptotic performance of the Block Sign Least Mean Squares algorithm. The analysis results are confirmed through simulations.

Real Time Implementation of SIGN LMS Adaptive Filters using Xilinx System Generator

Sign Least Mean Square (SLMS) adaptive filter can adapt dynamically based on corresponding filter output. One of the major applications of adaptive filter is Noise cancellation. In real time applications like medical computing, speed of the process developing hardware is essential hence the hardware realization of SLMS adaptive filter using Xilinx System generator is proposed in this work. The propose architecture aims to reduce convergence rate, path delay and increasing speed. In this work (i) Modified architecture is designed for a 8-tap SLMS adaptive filter and (ii) multiplier less structure for Modified DLMS Filter. The designed architecture tested for ECG signal. The functionality of the algorithm is verified in MATLAB with various ECG data from the MIT-BIH database as input. Both LMS and SLMS are designed, simulated, synthesized and implemented in Virtex-5 FPGA using Xilnix ISE 14.3 . The result shows 5% decrease in total real time router completion and also decrease in the number of adders and subtractors, the maximum combinational path delay has been reduced by 48.84% in Systolic Sign LMS Filter when compared to LMS Filter.

Analysis of ECG Signal Processing and Filtering Algorithms

Electrocardiography (ECG) is a common technique for recording the electrical activity of human heart. Accurate computer analysis of ECG signal is challenging as it is exceedingly prone to high frequency noise and various other artifacts due to its low amplitude. In remote heath care systems, computer based high level understanding of ECG signals is performed using advanced machine learning algorithms. The accuracy of these algorithms relies on the Signal-to-Noise-Ratio (SNR) of the input ECG signal. In this paper, we analyse various methods for removing the high frequency noise components from the ECG signal and evaluate the performance of several adaptive filtering algorithms. The result suggest that the Normalized Least Mean Square (NLMS) algorithm achieves high SNR and Sign LMS is computationally efficient.

A high-speed realization of the delayed dual sign LMS algorithm

A high-speed realization of the delayed dual sign LMS algorithm

Beamforming in Smart Antenna using Some Variants of Least Mean Square Algorithm

Beamforming and side lobe level reduction of smart antenna are important tasks in mobile network. Adaptive signal processing algorithms are used for beam forming in smart antenna. In this paper, variable step-size sign least mean square (VS-SLMS) algorithm is used for beam forming of smart antenna with linear antenna array. The results are compared with the results obtained using sign least mean square (SLMS) algorithm. Variable step-size algorithm shows good results for beam forming compared to ordinary constant step-size algorithm.

Stochastic analysis of soft limiters in the LMS algorithm for stationary white Gaussian inputs—A unified theory

The effects of saturation-type nonlinearities on the input and the error in the weight update equation for LMS adaptation are investigated for a stationary white Gaussian data model for system identification. Nonlinear recursions are derived for the transient and steady-state weight first and second moments that include the effect of soft limiters on both the input and the error driving the algorithm. By varying a single parameter of the soft limiter, a general theory is presented that is applicable to LMS, soft limiting of the input, error or both and sign–sign LMS.

Acoustic Echo Cancellation using Computationally Efficient Adaptive Algorithm Techniques


 
 
 Several sign based LMS adaptive filters, which are computationally free having multiplier free weight update loops, are proposed for acoustic echo cancellation. The adaptive filters essentially minimizes the mean- squared error between a primary input, which is the echo, and a reference input, which is either echo that is correlated in some way with the echo in the primary input. The results show that the performance of the signed regressor. LMS algorithm is superior than conventional LMS algorithm, the performance of signed LMS and sign- sign LMS based realizations are comparable to that of the LMS based filtering techniques in terms of Average Attenuation and computational complexity.
 
 

FPGA Implementation of Combination of Error Normalised Sign LMS and Zero Attractor-ENSLMS Algorithms for Denoising ECG Signals

The electrical potential of the heart known as Electrocardiogram(ECG) are contaminated by noises like Power Line Interference, Baseline wander and Electrode Motion etc.,. To improve the reliability of the signal, it is necessary to eliminate these noises. Fixed Coefficient filters are very difficult for denoising the ECG signals. The dynamic changes in the ECG signals can be tracked by means of adaptive filters. Different adaptive filters available currently are Least Mean Square (LMS) and Recursive Least Square (RLS) algorithms. The LMS algorithm is considered for its less complexity and efficiency. The computational complexity of the LMS algorithm can be still be improved by using various other algorithms like Sign LMS Algorithm(SLMS), Error Nonlinear Sign LMS algorithm(ENSLMS). When there is a sparse impulse response, the system should dynamically inherit the sparseness level and this is achieved by exploiting the information about the system sparseness. In this paper, a new approach of combining the ENSLMS and ZA-ENLMS is proposed for noise cancellation in ECG. The designed structures are simulated using MATLAB 2014b, Xilinx System Generator and implemented in Virtex 5 FPGA. The performance analysis is done in terms of SNR and computational complexity. The implementation results show that the number of sliced LUT's and the number of slice registers has been reduced to 5.1% and 3.22% respectively compared with the proposed combination.

Analysis of the SNR Estimator for Speech Enhancement Using a Cascaded Linear Model

Elimination of tainted noise and improving the overall quality of a speech signal is speech enhancement. To gain the advantage of individual algorithms we propose a new linear model and that is in the form of cascade adaptive filters for suppression of non-stationary noise. We have successfully deployed NLMS (Normalized Least Mean Square) algorithm, Sign LMS (Least Mean Square) and RLS (Recursive Least Square) as the main de-noising algorithms. Moreover, we are successful in demonstrating that the prior information about the noise is not required otherwise it would have been difficult to estimate for fast-varying noise in non-stationary environment. This approach estimates clean speech by recognizing the long segments of the clean speech as one whole unit. During experiment/implementation we used in-house database (includes various types of non stationary noise) for speech enhancement and proposed model results have shown improvement over conventional algorithms not only in objective but in subjective evaluations as well. Simulations present good results with a new linear model that are compared with individual algorithm results.

ECG Noise Canceller: Studying and Performance Improvement under Different Algorithms

In Electrocardiogram (ECG) application, we face a main problem of power line noise added to the ECG signal. Different Researchers have been interested in this problem due to its importance. In this paper we introduce a study of different algorithms and their effects on the performance of the ECG noise canceller. We have used many kinds of algorithms such as: LMS (Least Mean Square), NLMS (Normalized Least Mean Square), Signed Regressor LMS (SRLMS), Sign LMS (SLMS), Sign-Sign LMS (SSLMS) and a new proposed modified LMS called Variable Step size Least Mean Square (VSLMS) using MATLAB software package as well as Unbiased Linear output Neural Network (ULNN) and Unbiased Non Linear output Neural Network (UNLNN). It is promising to clarify the difference among these algorithms with the aim of obtaining better performance.

Partial-Update Normalized Sign LMS Algorithm Employing Sparse Updates

Partial-Update Normalized Sign LMS Algorithm Employing Sparse Updates

Sign least mean squares-based deconvolution technique for ultrasonic testing

Sign LMS algorithms, members of the simplified adaptive least-mean-squares class, have been developed to reduce computational complexity and simplify hardware implementation. These advantages make them suitable to utilize in ultrasonic testing units, a class of applications requiring simple and efficient signal processing algorithms. This paper proposes a specific sign LMS adaptive filters-based deconvolution technique for ultrasonic straight beam pulse-echo inspections. It extracts only two of the interface echoes multiple reflections for enhanced resolution and quality-enriched presentation; this technique is named “selective deconvolution”. Resolution enhancement and presentation’s quality enrichment performance among the different sign LMS algorithms were investigated by experiments, and based on performance, the methods themselves were compared. Computational requirements are also presented. The proposed technique with the various adaptive sign LMS filters gave satisfactory results.

Mixed analog-digital crossbar-based hardware implementation of sign–sign LMS adaptive filter

Recently announcement of a physical realization of a fundamental circuit element called memristor by researchers at Hewlett Packard (HP) has attracted so much interest worldwide. Combination of this newly found element with crossbar interconnect technology, opened a new field in designing configurable or programmable electronic systems which can have applications in signal processing and artificial intelligence. In this paper, based on the simple memristor crossbar structure, we will propose a new mixed analog-digital circuit as a hardware implementation of the sign---sign least mean square (LMS) adaptive filter algorithm. In this proposed hardware, any multiplication and addition is performed with infinite precision and there is no necessity for the quantization of the input signal. Since the coefficients of the filter are stored in the switches of the crossbar, they can remain unchanged theoretically for an infinite period of time.

An Efficient Implementation of the Sign LMS Algorithm Using Block Floating Point Format

An efficient scheme is presented for implementing the sign LMS algorithm in block floating point format, which permits processing of data over a wide dynamic range at a processor complexity and cost as low as that of a fixed point processor. The proposed scheme adopts appropriate formats for representing the filter coefficients and the data. It also employs a scaled representation for the step-size that has a time-varying mantissa and also a time-varying exponent. Using these and an upper bound on the step-size mantissa, update relations for the filter weight mantissas and exponent are developed, taking care so that neither overflow occurs, nor are quantities which are already very small multiplied directly. Separate update relations are also worked out for the step size mantissa. The proposed scheme employs mostly fixed-point-based operations, and thus achieves considerable speedup over its floating-point-based counterpart.

Asymptotic analysis of stochastic gradient-based adaptive filtering algorithms with general cost functions

This paper presents an analysis of stochastic gradient-based adaptive algorithms with general cost functions. The analysis holds under mild assumptions on the inputs and the cost function. The method of analysis is based on an asymptotic analysis of fixed stepsize adaptive algorithms and gives almost sure results regarding the behavior of the parameter estimates, whereas previous stochastic analyses typically considered mean and mean square behavior. The parameter estimates are shown to enter a small neighborhood about the optimum value and remain there for a finite length of time. Furthermore, almost sure exponential bounds are given for the rate of convergence of the parameter estimates. The asymptotic distribution of the parameter estimates is shown to be Gaussian with mean equal to the optimum value and covariance matrix that depends on the input statistics. Specific adaptive algorithms that fall under the framework of this paper are signed error least mean square (LMS), dual sign LMS, quantized state LMS, least mean fourth, dead zone algorithms, momentum algorithms, and leaky LMS.

A new efficient LMS adaptive filtering algorithm

A new efficient LMS adaptive filtering algorithm is proposed. The algorithm has comparable performance to that of the direct-form LMS algorithm (DLMS), while it costs N/2-1 fewer multiplications at the expense of N/2+5 more additions than the DLMS algorithm, where N is the number of filter taps. The new algorithm has one more parameter adaptation than the DLMS algorithm. Further, the algorithm was combined with sign LMS algorithm (SA), signed regressor algorithm (SRA) and zero forcing (ZFA) algorithm for more complexity reduction. Simulation results showed that the new combined algorithms converge as fast as the direct-form SA, SRA and ZFA algorithms, while still maintaining comparable performances.