Showing papers on "Kernel adaptive filter published in 2011"

Kernel adaptive filtering with maximum correntropy criterion

Abstract: Kernel adaptive filters have drawn increasing attention due to their advantages such as universal nonlinear approximation with universal kernels, linearity and convexity in Reproducing Kernel Hilbert Space (RKHS). Among them, the kernel least mean square (KLMS) algorithm deserves particular attention because of its simplicity and sequential learning approach. Similar to most conventional adaptive filtering algorithms, the KLMS adopts the mean square error (MSE) as the adaptation cost. However, the mere second-order statistics is often not suitable for nonlinear and non-Gaussian situations. Therefore, various non-MSE criteria, which involve higher-order statistics, have received an increasing interest. Recently, the correntropy, as an alternative of MSE, has been successfully used in nonlinear and non-Gaussian signal processing and machine learning domains. This fact motivates us in this paper to develop a new kernel adaptive algorithm, called the kernel maximum correntropy (KMC), which combines the advantages of the KLMS and maximum correntropy criterion (MCC). We also study its convergence and self-regularization properties by using the energy conservation relation. The superior performance of the new algorithm has been demonstrated by simulation experiments in the noisy frequency doubling problem.

Distributed optimal consensus filter for target tracking in heterogeneous sensor networks

Abstract: This paper is concerned with the problem of filter design for target tracking over sensor networks. Different from most existing works on sensor networks, we consider the heterogeneous sensor networks with two types of sensors different on processing abilities (denoted as type-I and type-II sensors, respectively). However, questions of how to deal with the heterogeneity of sensors and how to design a filter for target tracking over such kind of networks remain largely unexplored. We propose in this paper a novel distributed consensus filter to solve the target tracking problem. Two criteria, namely, unbiasedness and optimality, are imposed for the filter design. The so-called sequential design scheme is then presented to tackle the heterogeneity of sensors. The minimum principle of Pontryagin is adopted for type-I sensors to optimize the estimation errors. As for type-II sensors, the Lagrange multiplier method coupled with the generalized inverse of matrices is then used for filter optimization. Furthermore, it is proven that convergence property is guaranteed for the proposed consensus filter in the presence of process and measurement noise. Simulation results have validated the performance of the proposed filter. It is also demonstrated that the heterogeneous sensor networks with the proposed filter outperform the homogenous counterparts in light of reduction in the network cost, with slight degradation of estimation performance.

Adaptive least squares support vector machines filter for hand tremor canceling in microsurgery

Abstract: One of the main problems for effective control of a minimally invasive surgery (MIS) is the imprecision that caused by hand tremor. In this paper, a novel adaptive filter, the least squares support vector machines adaptive filter (LS-SVMAF), is proposed to overcome this problem. Compared with traditional methods like multi layer perceptron (MLP), LS-SVM shows a superior performance of nonlinear modeling with small scale of data set or high dimensional input space. With the LS-SVMAF, we can model and predict the hand tremor more effectively and improve the precision and reliability in the master–slave robotic system for microsurgery. Simulation results demonstrate the effectiveness of the proposed filter and its superior performance over its competing rivals.

An Adaptive Kalman Filter for ECG Signal Enhancement

Abstract: The ongoing trend of ECG monitoring techniques to become more ambulatory and less obtrusive generally comes at the expense of decreased signal quality. To enhance this quality, consecutive ECG complexes can be averaged triggered on the heartbeat, exploiting the quasi-periodicity of the ECG. However, this averaging constitutes a tradeoff between improvement of the SNR and loss of clinically relevant physiological signal dynamics. Using a Bayesian framework, in this paper, a sequential averaging filter is developed that, in essence, adaptively varies the number of complexes included in the averaging based on the characteristics of the ECG signal. The filter has the form of an adaptive Kalman filter. The adaptive estimation of the process and measurement noise covariances is performed by maximizing the Bayesian evidence function of the sequential ECG estimation and by exploiting the spatial correlation between several simultaneously recorded ECG signals, respectively. The noise covariance estimates thus obtained render the filter capable of ascribing more weight to newly arriving data when these data contain morphological variability, and of reducing this weight in cases of no morphological variability. The filter is evaluated by applying it to a variety of ECG signals. To gauge the relevance of the adaptive noise-covariance estimation, the performance of the filter is compared to that of a Kalman filter with fixed, (a posteriori) optimized noise covariance. This comparison demonstrates that, without using a priori knowledge on signal characteristics, the filter with adaptive noise estimation performs similar to the filter with optimized fixed noise covariance, favoring the adaptive filter in cases where no a priori information is available or where signal characteristics are expected to fluctuate.

Adaptive Gaussian Sum Filter for Nonlinear Bayesian Estimation

Abstract: A nonlinear filter is developed by representing the state probability density function by a finite sum of Gaussian density kernels whose mean and covariance are propagated from one time-step to the next using linear system theory methods such as extended Kalman filter or unscented Kalman filter. The novelty in the proposed method is that the weights of the Gaussian kernels are updated at every time-step, by solving a convex optimization problem posed by requiring the Gaussian sum approximation to satisfy the Fokker-Planck-Kolmogorov equation for continuous-time dynamical systems and the Chapman-Kolmogorov equation for discrete-time dynamical systems. The numerical simulation results show that updating the weights of different mixture components during propagation mode of the filter not only provides us with better state estimates but also with a more accurate state probability density function.

New Strategies for Application of Adaptive Filters in Active Power Filters

Abstract: This paper describes new strategies to improve the transient response time of harmonic detection using adaptive filters applied to shunt active power filters. Two cases are presented and discussed, both using an adaptive notch filter, but one uses the least mean square algorithm to adjust the coefficients and the other uses the recursive least squares algorithm. The synchronization of the adaptive notch filter orthogonal input signals, which are generated by the Clarke transformation of the load currents, is achieved automatically without the need of a phase-locked loop. This procedure significantly reduces the real-time computation burden. Simulations using Matlab/Simulink are presented to clarify the algorithm, and practical implementation is performed using the DSP Texas Instruments TMS320F2812. The experimental results are presented and discussed.

Efficient Decoding With Steady-State Kalman Filter in Neural Interface Systems

Abstract: The Kalman filter is commonly used in neural interface systems to decode neural activity and estimate the desired movement kinematics. We analyze a low-complexity Kalman filter implementation in which the filter gain is approximated by its steady-state form, computed offline before real-time decoding commences. We evaluate its performance using human motor cortical spike train data obtained from an intracortical recording array as part of an ongoing pilot clinical trial. We demonstrate that the standard Kalman filter gain converges to within 95% of the steady-state filter gain in 1.5 ± 0.5 s (mean ±s.d.). The difference in the intended movement velocity decoded by the two filters vanishes within 5 s, with a correlation coefficient of 0.99 between the two decoded velocities over the session length. We also find that the steady-state Kalman filter reduces the computational load (algorithm execution time) for decoding the firing rates of 25±3 single units by a factor of 7.0±0.9. We expect that the gain in computational efficiency will be much higher in systems with larger neural ensembles. The steady-state filter can thus provide substantial runtime efficiency at little cost in terms of estimation accuracy. This far more efficient neural decoding approach will facilitate the practical implementation of future large-dimensional, multisignal neural interface systems.

Method and apparatus for reproducing audio signal by adaptively controlling filter coefficient

Abstract: A method and an apparatus for reproducing a sound signal are provided. The method includes generating an output sound signal to be transmitted to speakers by transmitting a first input sound signal through a filter; acquiring magnitude information of the output sound signal; determining frequency response parameters related to frequency responses of the filter based on the magnitude information; and adaptively adjusting coefficients of the filter based on the determined frequency response parameters.

A Moment Matching Ensemble Filter for Nonlinear Non-Gaussian Data Assimilation

Abstract: The ensemble Kalman filter is now an important component of ensemble forecasting. While using the linear relationship between the observation and state variables makes it applicable for large systems, relying on linearity introduces nonnegligible bias since the true distribution will never be Gaussian. This paper analyzes the bias of the ensemble Kalman filter from a statistical perspective and proposes a debiasing method called the nonlinear ensemble adjustment filter. This new filter transforms the forecast ensemble in a statistically principled manner so that the updated ensemble has the desired mean and variance. It is also easily localizable and, hence, potentially useful for large systems. Its performance is demonstrated and compared with other Kalman filter and particle filter variants through various experiments on the Lorenz-63 and Lorenz-96 systems. The results show that the new filter is stable and accurate for challenging situations such as nonlinear, high-dimensional systems with spar...

Cubature information filter and its applications

Abstract: This paper presents a new estimation algorithm called cubature information filtering for nonlinear systems. The proposed algorithm is developed from an extended information filter and a recently developed cubature Kalman filter. Unlike the extended Kalman filter, the proposed filter does not require the evaluation of Jacobians during state estimation. The efficacy of the proposed algorithm is demonstrated by simulation examples on frequency demodulation and localization problem and is compared with unscented information filtering.

A family of proportionate normalized subband adaptive filter algorithms

Abstract: In this paper, the concept of proportionate adaptation is extended to the normalized subband adaptive filter (NSAF), and seven proportionate normalized subband adaptive filter algorithms are established. The proposed algorithms are proportionate normalized subband adaptive filter (PNSAF), μ ‐law PNSAF (MPNSAF), improved PNSAF (IPNSAF), the improved IPNSAF (IIPNSAF), the set-membership IPNSAF (SM-IPNSAF), the selective partial update IPNSAF (SPU-IPNSAF), and SM-SPU-IPNSAF which are suitable for sparse system identification in network echo cancellation. When the impulse response of the echo path is sparse, the PNSAF has initial faster convergence than NSAF but slows down dramatically after initial convergence. The MPNSAF algorithm has fast convergence speed during the whole adaptation. The IPNSAF algorithm is suitable for both sparse and dispersive impulse responses. The SM-IPNSAF exhibits good performance with significant reduction in the overall computational complexity compared with the ordinary IPNSAF. In SPU-IPNSAF, the filter coefficients are partially updated rather than the entire filter at every adaptation. In SM-SPU-IPNSAF algorithm, the concepts of SM and SPU are combined which leads to a reduction in computational complexity. The simulation results show good performance of the proposed algorithms.

Adaptive PD+I Control of a Switch-Mode DC–DC Power Converter Using a Recursive FIR Predictor

Abstract: This paper presents an alternative technique for the adaptive control of power electronic converter circuits. Specific attention is given to the adaptive control of a dc-dc converter. The proposed technique is based on a simple adaptive filter method and uses a one-tap finite impulse response (FIR) prediction error filter (PEF). The method is computationally efficient and based around a dichotomous coordinate descent (DCD) algorithm. The DCD-recursive least squares (RLS) algorithm has been employed as the adaptive PEF to reduce the computational complexity of existing RLS algorithms for efficient hardware implementation. Results show that the DCD-RLS is able to improve the dynamic performance and convergence rate of the adaptive gains (filter taps) within the controller. In turn, this yields a significant improvement in the overall dynamic performance of the closed-loop control system, particularly in the event of abrupt parameter changes. The proposed controller uses an adaptive proportional-derivative+integral (PD +I) structure which, alongside the DCD algorithm, offers an effective substitute to a conventional proportional-integral-derivative (PID) controller. The nonadaptive integral controller (+I), introduced in the feedback loop, increases the excitation of the filter tap weight and ensures good regulation. The approach results in a fast adaptive controller with self-loop compensation. This is required to minimize the prediction error signal and, in turn, minimize the voltage error signal in the loop by automatically calculating the optimal pole locations. The prediction error signal is further minimized through a second-stage FIR filter (adaptation gain stage). This ensures that the adaptive gains converge to their optimal value. This paper presents detailed simulation analysis and experimental validation on a prototype synchronous dc-dc buck converter. The experimental results clearly demonstrate the superior dynamic performance and voltage regulation compared to conventional PID and adaptive LMS control schemes, with only a modest increase in the computational burden to the microprocessor.

Fast Bilateral Filter With Arbitrary Range and Domain Kernels

Abstract: In this paper, we present a fast implementation of the bilateral filter with arbitrary range and domain kernels. It is based on the histogram-based fast bilateral filter approximation that uses uniform box as the domain kernel. Instead of using a single box kernel, multiple box kernels are used and optimally combined to approximate an arbitrary domain kernel. The method achieves better approximation of the bilateral filter compared to the single box kernel version with little increase in computational complexity. We also derive the optimal kernel size when a single box kernel is used.

Intra smoothing filter for video coding

Abstract: This disclosure relates to techniques for reducing the amount of additional data encoded with a block encoded using intra-predictive coding. Particularly, the techniques provide apparatus and methods of applying a smoothing filter to prediction samples used in intra-predictive coding. For example, in fixed mode-dependent intra- predictive coding, a video encoder may determine the type of smoothing filter applied to prediction samples based on block size and intra-prediction mode combination associated with the current block, where the combination is used to look up a filter in a first filter table. In adaptive mode-dependent intra-predictive coding, the encoder uses two filters, one from the first filter table and another from a second filter table, applies both filters, and determines which yields better results. When the second filter table filter yields better results, the encoder encodes a filtering indication. When a filter from the first filter table is used, no filtering indication is encoded.

High-throughput pipelined realization of adaptive FIR filter based on distributed arithmetic

Abstract: In this paper, we propose an efficient pipelined architecture for high-speed adaptive filter based on distributed arithmetic (DA). We have shown that the sampling period could be substantially reduced by using carry-save accumulation instead of shift-accumulation for DA-based inner-product implementation for the computation of filter output. Unlike the existing design, the proposed design does not involve any lookup table (LUT). It involves half the number of registers compared to the existing DA-based design to store the sum of different combinations of input samples. The proposed design involves nearly 17% more hardware but offers nearly 7 times throughput and nearly 14 times less energy per sample, in average for filter orders N = 8, 16 and 32 over the existing DA-based design for adaptive filter.

Robust Ensemble Filtering and Its Relation to Covariance Inflation in the Ensemble Kalman Filter

Abstract: A robust ensemble filtering scheme based on the H∞ filtering theory is proposed. The optimal H∞ filter is derived by minimizing the supremum (or maximum) of a predefined cost function, a criterion different from the minimum variance used in the Kalman filter. By design, the H∞ filter is more robust than the Kalman filter, in the sense that the estimation error in the H∞ filter in general has a finite growth rate with respect to the uncertainties in assimilation, except for a special case that corresponds to the Kalman filter.The original form of the H∞ filter contains global constraints in time, which may be inconvenient for sequential data assimilation problems. Therefore a variant is introduced that solves some time-local constraints instead, and hence it is called the time-local H∞ filter (TLHF). By analogy to the ensemble Kalman filter (EnKF), the concept of ensemble time-local H∞ filter (EnTLHF) is also proposed. The general form of the EnTLHF is outlined, and some of its special cases are di...

Non-linear adaptive scheme for cancellation of transmit out of band emissions

Abstract: A method and apparatus for a non-linear adaptive scheme for transmit out of band emission cancellation is provided. Embodiments disclosed herein provide a method for removing unwanted transmitter emissions from a composite received signal. The method performs the steps of: extracting the I and Q samples from a modulator output; inputting the I and Q samples to a non-linear filter; applying weights to the non-linear filter outputs, combining the non-linear filter outputs to generate a broadband emission estimate; selecting a portion of a transmit emission in a desired portion of a receive band; subtracting an output of the non-linear filter from a composite signal; and feeding back a residual error to the non-linear filter; adapting the non-linear filter iteratively.

Adaptive bearings vibration modelling for diagnosis

Abstract: An adaptive algorithm for vibration signal modeling is proposed in the paper. The aim of the signal processing is to detect the impact signals (shocks) related to damages in rolling element bearings (REB). Damage in the REB may result in cyclic impulsive disturbance in the signal, however they are usually completely masked by the noise. Moreover, impulses may have amplitudes that vary in time due to changes transmission path, load and properties of the noise. Thus, the solution should be an adaptive one. The proposed approach is based on the normalized exact least-square time-variant lattice filter (adaptive Schur filter). It is characterized by an extremely fast start-up performance, an excellent convergence behavior, and a fast parameter tracking capability and make this approach interesting. The method is well-adapted for analysis of the non-stationary time-series, so it seems to be very promising for diagnostics of machines working in time varying load/speed conditions.

Evolutionary Computation on Programmable Robust IIR Filter Pole-Placement Design

Abstract: This paper explores the pole-placement design problem of a robust stable infinite-impulse-response (IIR) filter to attenuate or eliminate the undesired measurement noise and proposes a strategy based on an adaptive differential evolution (ADE) algorithm to design a filter. The results are compared to the results of other popular evolutionary algorithms, e.g., particle swarm optimization (PSO), genetic algorithm (GA), and improved genetic algorithm (IGA). The stability robustness for an IIR filter will be achieved by placing all poles inside a disk D(α, r) contained in the unit disk, in which α is the center, and r is the radius of the disk. This investigation first uses a robust stability criterion, called the D(α, r)-stability criterion, to ensure that digital filter poles lie inside a disk D(α, r). The proposed strategy checks the criterion during differential evolution (DE) and adaptively adjusts the DE parameters, depending on the current DE performance. This paper also introduces two design examples of a bandpass IIR filter and a low-pass IIR filter for the measurement of a speech signal. These examples show that the proposed strategy performance based on the proposed ADE is better than designs based on PSO, GA, and IGA. Finally, this paper implements an IIR filter on the field-programmable gate array (FPGA) chip to verify the designed filter performance in practical electronic devices and uses speech signals as an input signal to the FPGA chip to verify that the measurement noise of the speech signal is attenuated by the designed IIR filter.

Method and apparatus for performing intra-prediction using adaptive filter

Abstract: Provided is a method and apparatus for performing intra-prediction using an adaptive filter. The method for performing intra-prediction comprises the steps of: determining whether or not to apply a first filter for a reference pixel value on the basis of information of a neighboring block of a current block; applying the first filter for the reference pixel value when it is determined to apply the first filter; performing intra-prediction on the current block on the basis of the reference pixel value; determining whether or not to apply a second filter for a prediction value according to each prediction mode of the current block, which is predicted by the intra-prediction performance on the basis of the information of the neighboring block; and applying the second filter for the prediction value according to each prediction mode of the current block when it is determined to apply the second filter.

Sliding-Mode Filter Design for Linear Systems With Unmeasured States

Abstract: This paper addresses the mean-square and mean-module filtering problems for a linear system with Gaussian white noises. The obtained solutions contain a sliding-mode term, signum of the innovation process. It is shown that the designed sliding-mode mean-square filter generates the mean-square estimate, which has the same minimum estimation-error variance as the best estimate given by the classical Kalman-Bucy filter, although the gain matrices of both filters are different. The designed sliding-mode mean-module filter generates the mean-module estimate, which yields a better value of the mean-module criterion in comparison with the mean-square Kalman-Bucy filter. The theoretical result is complemented with an illustrative example verifying the performance of the designed filters. It is demonstrated that the estimates produced by the designed sliding-mode mean-square filter and the Kalman-Bucy filter yield the same estimation-error variance, and there is an advantage in favor of the designed sliding-mode mean-module filter.

Quantized filtering of linear stochastic systems

Abstract: >> In this paper we investigate a general multi-level quantized filter of linear stochastic systems. For a given multi-level quantization and under the Gaussian assumption on the predicted density, a quantized innovations filter that achieves the minimum mean square error is derived. The filter is given in terms of quantization thresholds and a simple modified Riccati difference equation. By optimizing the filtering error covariance with respect to quantization thresholds, the associated optimal thresholds and the corresponding filter are obtained. Furthermore, the convergence of the filter to the standard Kalman filter is established. We also discuss the design of a robust minimax quantized filter when the innovation covariance is not exactly known. Simulation and experimental results illustrate the effectiveness and advantages of the proposed quantized filter.

Multi-target tracking for measurement models with additive contributions

Abstract: Moment-based filters, such as the Probability Hypothesis Density (PHD) filter, are an attractive solution to multi-target tracking. However, an underlying assumption for the PHD filter is that each measurement is either caused by a single target or clutter. In this paper, we design a novel moment-based multi-target filter, the Additive Likelihood Moment (ALM) filter, where the measurements are affected by all targets. We focus on the cases where the likelihood can be expressed as a function of the sum of the individual target contributions. As an example, we consider radio tomographic tracking where the attenuation of the signal between a pair of sensors is the sum of attenuations caused by all targets. Our multi-target tracking algorithm is based on a particle approximation of our moment-based filter. Our simulations show that our algorithm has a lower estimation error than MCMC particle methods while achieving 80% savings in terms of computational time.

Adaptive Marginal Median Filter for Colour Images

Abstract: This paper describes a new filter for impulse noise reduction in colour images which is aimed at improving the noise reduction capability of the classical vector median filter. The filter is inspired by the application of a vector marginal median filtering process over a selected group of pixels in each filtering window. This selection, which is based on the vector median, along with the application of the marginal median operation constitutes an adaptive process that leads to a more robust filter design. Also, the proposed method is able to process colour images without introducing colour artifacts. Experimental results show that the images filtered with the proposed method contain less noisy pixels than those obtained through the vector median filter.

A high-speed FIR adaptive filter architecture using a modified delayed LMS algorithm

Abstract: In this paper, we present a modified delayed least means square (DLMS) adaptive algorithm to achieve lower adaptation-delay. Besides, we have proposed an efficient pipelined architecture for the implementation of this adaptive filter. We have shown that the proposed DLMS adaptive filter can be implemented by a pipelined inner-product computation unit for calculation of feedback error, and a pipelined weight-update unit consisting of N parallel multiply accumulators, for filter order N. From the synthesis results we find that the existing direct-form structure of [8] involves nearly 50% more area-delay product (ADP) and nearly 74% more energy per sample (EPS) than the proposed one, in average, for filter orders N = 8, 16 and 32. The best of the existing systolic structures [7], similarly, involves nearly 43% more ADP and nearly 35% higher EPS than the proposed one for the same filter orders.