Adaptive Filter Structure Research Articles

GoLDFormer: A global–local deformable window transformer for efficient image restoration

Thanks to the powerful modeling capabilities of multi-head self attention (MSA), transformers have shown significant performance gains in vision tasks. However, as transformers require heavy computation, more efficient designs are required. In this paper, we present an efficient transformer architecture named GoLDFormer for image restoration. GoLDFormer extends the capability of window-based self-attention through two core designs. First, We propose a globally-enhanced window-based transformer block (G-WTB), which applies transposed attention to a compressed window representation rather than the spatial features, thus establishing connections between all windows with less computational complexity. Second, since the interactions between image content and window attention weights can be interpreted as spatially varying convolution, we introduce an adaptive filter structure into transformer models and propose a deformable filtering block (DFB) to enable cross-window connections. By adjusting the shape of the generated filters in the DFB, we can balance the computational costs and the degree of adjacent window interaction. Extensive experiments on several image restoration tasks demonstrate that GoLDFormer achieves competitive results against recent methods but with optimal computational costs.

Study on the Mechanism and Suppression of Harmonic Vibration of AMB-Rotor System

The AMB-rotor system is complex and has strong coupling characteristics, which allows multi-harmonic disturbances to enter the system through different ways to produce vibrations with rich spectrum components, which has a great influence on the improvement of micro-vibration accuracy of the rotor system. To further achieve active control of the micro-vibration in the AMB-rotor system, firstly, the mechanism of multi-source disturbance is analyzed according to the working principle of the AMB-rotor system, and the mathematical and physical relationship between the mechanism of disturbance generation and the inducement is deeply studied. Then, the structure of a novel adaptive notch filter, the method of adaptive frequency estimation and analysis of harmonic current suppression in the AMB system are presented. Finally, simulation and experimental research using an MSCMG system demonstrate the feasibility of the proposed method regarding the elimination of harmonic control current.

A robust sliding window adaptive filtering technique for phonocardiogram signal denoising

AbstractObjectiveThe digital sound recording of several heart sounds (HSs) is named phonocardiogram (PCG) signal. Analysing these PCG signals is essential for diagnosing diverse sorts of heart disorders. Nevertheless, owing to troubling surrounding noise signals, PCG signal recording is challenging. So, before utilising PCG for advanced processing, de‐noising the PCG signal is executed. A new sliding window adaptive noise cancellers (SWANC)‐centred filter technique is proposed in this paper for de‐noising along with recovering the PCG signal effectively.MethodUtilising the least mean square (LMS), an SW optimum adaptive filter (AF) structure is introduced in this work for estimating a De‐noised signal (DS) with better accuracy. Here, via the SWAF stage, a noisy signal is processed. An SW of fixed duration slides over the signal; also, the signal is filtered utilising the AF in each window. Utilising this SWAF architecture, this method approximates the PCG signal's clean version with better accuracy.ResultsThe proposed robust SWAF is analogized to experimental PCG signals that were corrupted by Gaussian noise (GN) together with pink noise (PN) with distinct noise levels (NLs). From the physionet database, the experiential data are acquired. The outcomes exhibited that a remarkable performance was attained by the robust SWAF model.DiscussionA 2–10 times decrease in MSE values was attained by the proposed filter structure when analogized with conventional LMS filter configuration. Further, the SNR is improved by 3 times and comparatively, the PSNR enhancement was 4%–25%. The association betwixt the clean signal along with its estimate is greater than 0.92.ConclusionA cost‐efficient hardware installation of ANC with higher accuracy was offered by the SW LMS AF model. For obtaining desired convergence speed and accuracy, such models are tested for real‐time performances in the future.

The design and implementation of folded adaptive lattice filter structures in FPGA for ECG signals

An adaptive filter is the utmost essential filter castoff in statistical signal dealing. The fine-tuning of the filter factor in relation to the response signal is the adaptive filter's key feature due to fewer calculations, Least Mean Square (LMS) adaptive filters are widely used to remove noise from Electrocardiograms (ECG). The adaptive filters are realized as signal processing algorithms in Digital Signal Processors (DSPs) or in VLSI Signal Processors (VSPs). The technique provides a way to create a folded adaptive lattice LMS filter, which requires less hardware than an adaptive lattice filter. Folding is an algorithm that uses a time scheduling technique that combines arithmetic operations into one operation which reduces Register and silicon chip areas. The design and implementation of a folded lattice adaptive filter remove Power Line Interference (PLI) noise from ECG signals. The MATLAB Xilinx System Generator tool is used to design the Adaptive Lattice LMS Filter and Folded Adaptive Lattice LMS Filter with Folding Order K = 2 and K = 4 and realized in the Virtex 5 FPGA KIT. The results of the folded architecture show that the area is reduced for K = 2 and K = 4 by 82.60% and 91.05%, respectively compared with a normal adaptive lattice filter.

Online estimation of state-of-charge inconsistency for lithium-ion battery based on SVSF-VBL

It is of great significance to estimate cell inconsistency for improving the service life and safety performance of the power battery pack. To accurately estimate the state-of-charge (SOC) inconsistency of cells with different performance parameters and working states, a method combining adaptive robust unscented Kalman filter and smooth variable structure filter with time-varying smoothing boundary layer (SVSF-VBL) is proposed. The cell mean-difference model is used to simulate the behavior characteristics of the battery module, including the cell mean model expressed by the dual polarization (DP) model and the cell difference model characterized by the hypothetical Rint model. Firstly, the improved forgetting factor recursive least square is applied to identify parameters of the DP model, and the unscented Kalman filter incorporating robust estimation and adaptive filter tuning is employed to estimate the SOC of the battery module. Then, SVSF-VBL is used to estimate the SOC difference between each cell and module based on the Rint model for improving the estimation accuracy and robustness. In addition, the comprehensive inconsistency of the cells can be captured by the secondary performance indicator inherent in SVSF-VBL, which contributes to the in-depth study of cell inconsistency. Finally, a series of tests are carried out to verify the performance of the proposed method, and the results show that the method can improve the estimation accuracy and convergence performance while effectively suppressing the system disturbance.

A robust low-cost adaptive filtering technique for phonocardiogram signal denoising

ObjectivePhonocardiogram (PCG) represents the recordings of various heart sounds. To diagnose the different ailments of the heart, it is required to analyze these PCG signals. However, recording PCG signals is challenging since it is prone to surrounding noise signals. Therefore, there is a need to denoise the PCG signal before being used for advanced processing. This paper proposes an Adaptive Noise Cancellers-based filter model for effectively denoising and recovering the PCG signal. MethodThis work introduces an optimum adaptive filter structure for estimating a noise-free signal with high accuracy using Least Mean Square (LMS) algorithm. A noisy signal is processed through multiple adaptive filter stages connected in series in the proposed work. Multiple stages are automatically added, and each stage filter’s step size process is dynamically changed. The estimate of clean PCG signal approximated using this multistage cascaded adaptive filter architecture is subsequently used in the next module to recover the clean PCG signal with high accuracy. ResultsThe proposed robust multistage adaptive filter is evaluated for denoising synthetic and experimental PCG signals corrupted by Gaussian and pink noise of various input Signal to noise (SNR) levels. The experimental data are taken from the physionet database (Classification of Heart Sound Recordings: The PhysioNet/Computing in Cardiology Challenge 2016). The results demonstrate that the robust multistage filter model performs remarkably well. DiscussionCompared with various filter configurations, the proposed filter structure achieves an 8-50% reduction in MAE values and the 45–87% reduction in MSE values. Further, there is an improved SNR of 15–60%, ANR of 15–65%, and PSNR improvement by 7–25% comparatively. The correlation between the clean signal and its estimate obtained using the proposed filter model is more than 0.99. ConclusionUsing an LMS adaptive filter in the proposed filter model offers a cost-effective hardware implementation of Adaptive Noise Canceller with high accuracy. In the future, the suggested robust multistage adaptive filter model can be tested for real-time performance when improved convergence speed and accuracy are desired.

Noise reduction in speech signal of Parkinson’s Disease (PD) patients using optimal variable stage cascaded adaptive filter configuration

In hands-free communications, including mobile phones, teleconferencing, and video conferencing, the ultimate aim is to obtain a noise-free signal. The adaptive filters that utilize the LMS adaptation algorithm are the best suited for speech signal de-noising among the existing speech signal enhancement techniques. LMS adaptive filter is preferred due to its low computational complexity and robustness, but it fails to simultaneously achieve a faster convergence speed and minimum steady-state MSE. This work introduces a novel cascaded adaptive filter structure for speech signal de-noising wherein the objective is to estimate the noise-free signal with high accuracy. The proposed optimal variable stage cascaded adaptive filter model utilizes a variable stage cascaded adaptive filter structure to estimate the noise-free signal, which is then employed to recover the clean signal with high accuracy. The proposed filter model is tested for reducing the noise from a normal speech signal taken from the NOIZEUS database and an impaired speech signal of Parkinson’s Disease affected patients taken from the MDVR-KCL dataset. The signals are corrupted by various stationary and non-stationary noise signals of different input SNR levels. The noisy signal is used to test the performance of the proposed optimal filter model. The signal de-noising performance of the proposed filter is evaluated in MSE, SNR, and ANR. The results have shown that the proposed filter model provides remarkable performance and an output SNR of 10–15 dB higher than the existing cascaded filter structures. Further, the proposed filter model employs LMS adaptive algorithm, offering a cost-effective and straightforward hardware implementation of ANC with improved accuracy.

Dual phase dependent RLS filtering approach for baseline wander removal in ECG signal acquisition

Electrocardiogram (ECG) signals often get contaminated during physical activity, due to baseline disturbances. Removal of baseline interference also leads to variation in the timing of ECG waves. This paper proposes a dual phase dependent-recursive least square (DPD-RLS) adaptive filter that removes the baseline interference during ECG signal acquisition. This approach initially detects the upper envelope followed by the subtraction of the upper envelope with the original signal. The lower envelope is then detected from the resultant signal. Two phases are extracted from the upper envelope signal and the lower envelope signal that compensates for the phase change effects during the baseline wander removal process. A DPD-RLS adaptive filter structure is proposed that compensates for the phase changes near the P, Q, and T onset of the ECG signal. The evaluation was done using the clean ECG signals from the MIT-BIH arrhythmia dataset and baseline wander signals from MIT-BIH noise stress dataset signals using the metrics such as output signal to noise ratio (SNRo), correlation coefficient (γ), and percent root mean square difference (PRD) for the different input signal to noise ratios (SNRi). The proposed scheme provides an average output SNR, the correlation coefficient of 14.07dB and 0.980 respectively at SNRi=5dB when evaluated on different ECG records. Experimental results reveal that the proposed baseline wander removal scheme outperforms the traditional schemes.

A Low-Cost Multistage Cascaded Adaptive Filter Configuration for Noise Reduction in Phonocardiogram Signal.

Phonocardiogram (PCG), the graphic recording of heart signals, is analyzed to determine the cardiac mechanical function. In the recording of PCG signals, the major problem encountered is the corruption by surrounding noise signals. The noise-corrupted signal cannot be analyzed and used for advanced processing. Therefore, there is a need to denoise these signals before being employed for further processing. Adaptive Noise Cancellers are best suited for signal denoising applications and can efficiently recover the corrupted PCG signal. This paper introduces an optimal adaptive filter structure using a Sign Error LMS algorithm to estimate a noise-free signal with high accuracy. In the proposed filter structure, a noisy signal is passed through a multistage cascaded adaptive filter structure. The number of stages to be cascaded and the step size for each stage are adjusted automatically. The proposed Variable Stage Cascaded Sign Error LMS (SELMS) adaptive filter model is tested for denoising the fetal PCG signal taken from the SUFHS database and corrupted by Gaussian and colored pink noise signals of different input SNR levels. The proposed filter model is also tested for pathological PCG signals in the presence of Gaussian noise. The simulation results prove that the proposed filter model performs remarkably well and provides 8–10 dB higher SNR values in a Gaussian noise environment and 2-3 dB higher SNR values in the presence of colored noise than the existing cascaded LMS filter models. The MSE values are improved by 75–80% in the case of Gaussian noise. Further, the correlation between the clean signal and its estimate after denoising is more than 0.99. The PSNR values are improved by 7 dB in a Gaussian noise environment and 1-2 dB in the presence of pink noise. The advantage of using the SELMS adaptive filter in the proposed filter model is that it offers a cost-effective hardware implementation of Adaptive Noise Canceller with high accuracy.

A Novel PSO-Based Adaptive Filter Structure with Switching Selection Criteria for Active Noise Control

In recent years, active noise control (ANC) systems have been widely used in advanced electronic appliances. Nowadays, several authors use gradient-optimization algorithms since they can be easily implemented in these devices. However, these algorithms need to estimate the secondary path in advance. As consequence, this factor can limit its use in real-ANC applications since the secondary path can undergo significant variations over time. To solve this problem, we propose an ANC system with switching filter selection based on particle swarm optimization (PSO) algorithms. Specifically, we use two sets of populations of particles with different acceleration coefficients and inertia weights to create an advanced structure in which the first PSO algorithm guarantees a high convergence speed while the use of the second PSO algorithm allows to achieve a high-level noise reduction. The results demonstrate that the proposed algorithm exhibits better convergence properties compared with previously reported solutions.

Design of High-speed Delay-FXLMS Hardware Architecture Based on FPGA

In order to improve the convergence and clock speed of DFxLMS adaptive filter, a hardware architecture of fine-grained retiming DFxLMS (HS-TF-RDFXLMS) filter in the form of hardware sharing transpose is proposed. Firstly, the architecture adopts delay decomposition algorithm to solve the problem that the convergence of filter decreases due to the increase of delay and output lag. Secondly, on the premise that the algorithm performance remains unchanged, the adaptive filter module and the secondary path module are transposed to further reduce the critical path to improve the clock speed of the system. The number of registers is reduced by optimizing circuit sub-module. Finally, the area/speed tradeoff of TF-RDFXLMS filter is realized by hardware sharing on the basis of constant critical path. Experimental results show that the convergence speed of the algorithm is 3.5 times that of DFxLMS algorithm, and the critical path is shortened by ([log2N]+1)TADD. The circuit structure of adaptive filter designed in this paper is realized by Xilinx Artix7 FPGA platform. The clock speed of HS-TF-RDFXLMS filter is reduced by 4.386% compared with TF-RDFXLMS filter. However, the resources of LUT and FF are saved by 10.964% and 28.322% respectively. The power consumption is 150.73 mW. This improves the performance of the system.

Adaptive Attitude Estimation Using a Hybrid Model-Learning Approach

Attitude determination using the smartphone’s inertial sensors poses a major challenge due to the sensor low-performance grade and variate nature of the walking pedestrian. In this paper, data-driven techniques are employed to address that challenge. To that end, a hybrid deep learning and model based solution for attitude estimation is proposed. Here, classical model based equations are applied to form an adaptive complementary filter structure. Instead of using constant or model based adaptive weights, the accelerometer weights in each axis are determined by a unique neural network. The performance of the proposed hybrid approach is evaluated relative to popular model based approaches using experimental data.

Analog Self-Interference Cancellation in Dual-Polarization Full-Duplex MIMO Systems

Full-duplex (FD) technology combined with dual-polarization (DP) multiple-input multiple-output (MIMO) systems is attractive to improve spectral efficiency and to enhance link capacity. Cancelling self-interference (SI) in such DPFD MIMO systems using beamforming techniques is very challenging due to a significant difference of the co-polarization and cross-polarization SI channels. In this letter, an analog adaptive filter structure is proposed to mitigate both co-polarization and cross-polarization SIs in DPFD MIMO systems. Stationary analysis is applied to evaluate the performance of the proposed structure. Simulation results show that about 45 dB to 55 dB of SI cancellation can be achieved regardless of the isolation differences between cross-polarization and co-polarization channels.

Learning Characteristics of a new Cascaded Second Order Adaptive Filter.(Dept.E)

This paper introduces a new structure of adaptive filter called Rcos filter. Analysis of the filter is considered in details. A comparative study of performance of the filter with different structures is also considered. It is found that the adaptive Rcos cascaded second order structure outperforms the ladder, lattice, and cascaded second order structures. The new structure find its applications in linear prediction of speech signals.

Online Censoring Based Weighted-Frequency Fourier Linear Combiner for Estimation of Pathological Hand Tremors

An online censoring (OC) based weighted-frequency Fourier linear combiner (OC-WFLC) adaptive filtering structure is proposed to reduce data processing costs in the estimation of pathological hand tremor (PHT) measurements. The proposed OC-WFLC is combined with the Fourier linear combiner (FLC) to effectively separate the PHT and voluntary movement from the hand tremor signal. The OC-WFLC is shown to adaptively extract the most informative frequency information, that is readily employed within the FLC to adaptively decompose the measurement signal into its PHT and voluntary movement components. The utilization of the OC strategy in the proposed framework is shown to significantly reduce data processing costs without adverse effects on the performance. Simulation results on real-world PHT data demonstrate the ability of the proposed OC-WFLC to yield a dramatic reduction of the processing time, a prerequisite for real-time rehabilitative, wearable, and assistive technology designed for PHT patients.

Optimal nonlinear fi ltering of stochastic processes in rescue radar

Subject and Purpose. Smoke, fog, avalanches, debris of collapsed structures and other optically opaque obstacles in both natural and man-made disasters make optical sensors useless for detecting victims. Electromagnetic waves of the decimeter range penetrate well almost all obstacles, reflect from the trapped people and return to the radar receiver. Due to the breathing and heartbeat, the human-reflected sounding signals get the Doppler phase modulation, which is an information signal. These information signals and their properties provide the subject matter for the present research with the aim to create optimal methods and algorithms of random event processing for the prompt location of survivors by rescuers. Method and Methodology. The method of stochastic analysis of the fluctuation Doppler spectra of reflected sounding signals shows that the information signals have properties of conditional Markov processes. Results. The problem of optimal nonlinear filtering of conditional Markov processes entering the radar signal processing unit has been examined closely. An optimal adaptive filter has been proposed to reduce the masking effect of interferences caused by non-stationary noises and sounding signal reflections from stationary objects. The optimality criterion is the minimum mean square error function whose current value is evaluated in real time during the filtering process as the statistics is accumulated. The filter coefficients are calculated by the recurrent, steepest descent algorithm. The real-time work is carried out through the use of fast Fourier transform algorithms. Conclusion. The structure of the optimal adaptive filter to be built into the radar signal processing unit has been developed. Real radar signals have shown that the optimal filtering during the signal processing in systems designed for detecting live people by their breathing and heartbeat facilitates the interpretation of the observed signals. Some spectra of real signals generated by human breathing and heartbeat are presented.

Double-talk robust acoustic echo canceller based on CNN filter

Conventional acoustic echo cancellation works by using an adaptive algorithm to identify the impulse response of the echo path. In this paper, we use the CNN neural network filter to remove the echo signal from the microphone input signal, so that only the speech signal is transmitted to the far-end. Using the neural network filter, weights are well converged by the general speech signal. Especially it shows the ability to perform stable operation without divergence even in the double-talk state, in which both parties speak simultaneously. As a result of simulation, this system showed superior performance and stable operation compared to the echo canceller of the adaptive filter structure.

Double-talk robust acoustic echo canceller based on CNN filter

Conventional acoustic echo cancellation works by using an adaptive algorithm to identify the impulse response of the echo path. In this paper, we use the CNN neural network filter to remove the echo signal from the microphone input signal, so that only the speech signal is transmitted to the far-end. Using the neural network filter, weights are well converged by the general speech signal. Especially it shows the ability to perform stable operation without divergence even in the double-talk state, in which both parties speak simultaneously. As a result of simulation, this system showed superior performance and stable operation compared to the echo canceller of the adaptive filter structure.

Beam-Based Analog Self-Interference Cancellation in Full-Duplex MIMO Systems

Self-interference (SI) cancellation for full-duplex (FD) multiple input multiple output (MIMO) systems is challenging due to both hardware and signal processing complexity. In this paper, a beam-based adaptive filter structure with analog least mean square (ALMS) loops is proposed to significantly reduce the complexity of SI cancellation for FD MIMO systems. With this structure, the number of adaptive filters required for SI cancellation scales linearly with the number of transmit beams rather than quadratically with the number of antennas. Furthermore, to avoid additional transmit chains used to up-convert the beam signals to generate reference signals for the ALMS loops, a novel method is proposed to select the optimized reference signals from all transmitted signals. In addition, our stationary analysis shows that the proposed structure for FD MIMO systems outperforms the ALMS loop employed for an FD single input single output system. Simulations are conducted to confirm the theoretical analyses.

Widely Linear Quaternion-Valued Least-Mean Kurtosis Algorithm

A widely linear quaternion-valued least-mean kurtosis (WL-QLMK) algorithm is introduced for adaptive filtering of quaternion-valued circular and noncircular signals. In the design, kurtosis-based cost function is first defined in the quaternion domain by integrating the widely linear model, and augmented statistics, and then minimized using the recently developed generalized Hamilton-real (GHR) calculus. In this way, the novel WL-QLMK algorithm is obtained for training quaternion-valued adaptive filter structures. Furthermore, its steady-state performance is theoretically analyzed to determine the bounds of the step size, which provides a theoretical justification for simulations. The simulation results over both benchmark system identification scenarios, and one-step-ahead predictions of real-world 4D pathological resting tremors show that the proposed WL-QLMK algorithm, by virtue of its newly defined cost function, significantly enhances the performance compared to the recently developed quaternion-valued algorithms, especially for noncircular signals.