Showing papers on "Kernel adaptive filter published in 2015"

An LCL - LC Filter for Grid-Connected Converter: Topology, Parameter, and Analysis

Abstract: In order to further cut down the cost of filter for grid-connected pulsewidth modulation (PWM) converter under the more and more stringent grid code, a new kind of high-order filter, named LCL - LC filter, is presented in this paper. The resonant frequency characteristics of the filter are analyzed, and a parameter design method on the base of the characteristics is also proposed in the paper. The proposed parameter design method can easily make full use of the existing research results about the traditional LCL filter parameter design. And then a parameter robustness analysis method based on four-dimensional graphics is proposed to analyze parameter robustness of the presented filter. Compared with the traditional one, the proposed analysis method can analyze the filter performance under variations of several parameters at a time without any iteration. The comparative analysis and discussion considering the LCL filter, the trap filter, and the LCL - LC filter, are presented and verified through the experiments on a 5 kW grid-connected converter prototype. Experiment results demonstrate the accuracy of theoretical analysis and prove that the presented filter has a better performance than two others.

The Guided Bilateral Filter: When the Joint/Cross Bilateral Filter Becomes Robust

Abstract: The bilateral filter and its variants, such as the joint/cross bilateral filter, are well-known edge-preserving image smoothing tools used in many applications. The reason of this success is its simple definition and the possibility of many adaptations. The bilateral filter is known to be related to robust estimation. This link is lost by the ad hoc introduction of the guide image in the joint/cross bilateral filter. We here propose a new way to derive the joint/cross bilateral filter as a particular case of a more generic filter, which we name the guided bilateral filter. This new filter is iterative, generic, inherits the robustness properties of the robust bilateral filter, and uses a guide image. The link with robust estimation allows us to relate the filter parameters with the statistics of input images. A scheme based on graduated nonconvexity is proposed, which allows converging to an interesting local minimum even when the cost function is nonconvex. With this scheme, the guided bilateral filter can handle non-Gaussian noise on the image to be filtered. A complementary scheme is also proposed to handle non-Gaussian noise on the guide image even if both are strongly correlated. This allows the guided bilateral filter to handle situations with more noise than the joint/cross bilateral filter can work with and leads to high peak signal-to-noise ratio values as shown experimentally.

An Adaptive Spatial Filter for User-Independent Single Trial Detection of Event-Related Potentials

Abstract: Goal: Current brain–computer interfaces (BCIs) are usually based on various, often supervised, signal processing methods. The disadvantage of supervised methods is the requirement to calibrate them with recently acquired subject-specific training data. Here, we present a novel algorithm for dimensionality reduction (spatial filter), that is ideally suited for single-trial detection of event-related potentials (ERPs) and can be adapted online to a new subject to minimize or avoid calibration time. Methods: The algorithm is based on the well-known xDAWN filter, but uses generalized eigendecomposition to allow an incremental training by recursive least squares (RLS) updates of the filter coefficients. We analyze the effectiveness of the spatial filter in different transfer scenarios and combinations with adaptive classifiers. Results: The results show that it can compensate changes due to switching between different users, and therefore allows to reuse training data that has been previously recorded from other subjects. Conclusions: The presented approach allows to reduce or completely avoid a calibration phase and to instantly use the BCI system with only a minor decrease of performance. Significance: The novel filter can adapt a precomputed spatial filter to a new subject and make a BCI system user independent.

Constant Time Median and Bilateral Filtering

Abstract: This paper formulates both the median filter and bilateral filter as a cost volume aggregation problem whose computational complexity is independent of the filter kernel size. Unlike most of the previous works, the proposed framework results in a general bilateral filter that can have arbitrary spatial$$^{1}$$1 and arbitrary range filter kernels. This bilateral filter takes about 3.5 s to exactly filter a one megapixel 8-bit grayscale image on a 3.2 GHz Intel Core i7 CPU. In practice, the intensity/range and spatial domain can be downsampled to improve the efficiency. This compression can maintain very high accuracy (e.g., 40 dB) but over $$100\times $$100? faster.

Accurate Methods for Approximate Bayesian Computation Filtering

Abstract: The Approximate Bayesian Computation (ABC) filter extends the particle filtering methodology to general state-space models in which the density of the observation conditional on the state is intractable. We provide an exact upper bound for the mean squared error of the ABC filter, and derive sufficient conditions on the bandwidth and kernel under which the ABC filter converges to the target distribution as the number of particles goes to infinity. The optimal convergence rate decreases with the dimension of the observation space but is invariant to the complexity of the state space. We show that the adaptive bandwidth commonly used in the ABC literature can lead to an inconsistent filter. We develop a plug-in bandwidth guaranteeing convergence at the optimal rate, and demonstrate the powerful estimation, model selection, and forecasting performance of the resulting filter in a variety of examples.

Design of adaptive H ∞ filter for implementing on state-of-charge estimation based on battery state-of-charge-varying modelling

Abstract: This study suggests a new method for modelling lithium-ion battery types and state-of-charge (SOC) estimation using adaptive H∞ filter (AHF). First, a universal linear model with some free parameters is considered for dynamical behaviour of the battery. The battery voltage and SOC are used as states of the model. Then for every period in the charge/discharge process the free parameters of the model are identified. Each period of process is associated with a specific SOC value, hence the parameters can be regarded as functions of SOC in the entire process. The functions are determined based on polynomial approximation and least squares method. The proposed SOC-varying model is incorporated in AHF for SOC estimation. Moreover, a new method for adjusting the tuning parameters of the filter is suggested. The proposed method is verified by experimental tests on a lithium-ion battery and is compared with adaptive extended Kalman filter and square-root unscented Kalman filter

Double-talk detection for acoustic echo cancellation

Abstract: An acoustic echo canceller includes an adaptive filter and a double-talk detector. The adaptive filter includes a linear filter and a coefficient calculator. The linear filter has a transfer function that is controlled by a set of variable filter coefficients and that is configured to cancel an estimate of echo in a microphone signal to provide an output signal. The coefficient calculator is configured to update the set of variable filter coefficients based on a variable adaptation rate. The double-talk detector is configured to calculate changes in the energy of the variable filter coefficients (between updates of the coefficients). The acoustic echo canceller is configured to adjust the variable adaptation rate based on whether the energy of the variable filter coefficients is determined to be either oscillating or steadily changing (increasing or decreasing).

Poisson's equation in nonlinear filtering

Abstract: The aim of this paper is to provide a variational interpretation of the nonlinear filter in continuous time. A time-stepping procedure is introduced, consisting of successive minimization problems in the space of probability densities. The weak form of the nonlinear filter is derived via analysis of the first-order optimality conditions for these problems. The derivation shows the nonlinear filter dynamics may be regarded as a gradient flow, or a steepest descent, for a certain energy functional with respect to the Kullback--Leibler divergence. The second part of the paper is concerned with derivation of the feedback particle filter algorithm, based again on the analysis of the first variation. The algorithm is shown to be exact. That is, the posterior distribution of the particle matches exactly the true posterior, provided the filter is initialized with the true prior.

Adaptive Learning in Cartesian Product of Reproducing Kernel Hilbert Spaces

Abstract: We propose a novel adaptive learning algorithm based on iterative orthogonal projections in the Cartesian product of multiple reproducing kernel Hilbert spaces (RKHSs). The objective is to estimate or track nonlinear functions that are supposed to contain multiple components such as i) linear and nonlinear components and ii) high- and low- frequency components. In this case, the use of multiple RKHSs permits a compact representation of multicomponent functions. The proposed algorithm is where two different methods of the author meet: multikernel adaptive filtering and the algorithm of hyperplane projection along affine subspace (HYPASS). In a particular case, the “sum” space of the RKHSs is isomorphic, under a straightforward correspondence, to the product space, and hence the proposed algorithm can also be regarded as an iterative projection method in the sum space. The efficacy of the proposed algorithm is shown by numerical examples.

A fuzzy logic feedback filter design tuned with PSO for L 1 adaptive controller

Abstract: No simple way of tuning L1 adaptive controller feedback filter exists.Propose a Fuzzy-logic based approach for on-line tuning of the filter.Particle Swarm Optimization (PSO) is used to optimize the filter.Class of a strictly proper low pass filters with fixed structure is considered.Simulation demonstrate simplicity excellent performance and robustness. L 1 adaptive controller has been recognized for having a structure that allows decoupling between robustness and adaption owing to the introduction of a low pass filter with adjustable gain in the feedback loop. The trade-off between performance, fast adaptation and robustness, is the main criteria when selecting the structure or the coefficients of the filter. Several off-line methods with varying levels of complexity exist to help finding bounds or initial values for these coefficients. Such values may require further refinement using trial-and-error procedures upon implementation. Subsequently, these approaches suggest that once implemented these values are kept fixed leading to sub-optimal performance in both speed of adaptation and robustness. In this paper, a new practical approach based on fuzzy rules for online continuous tuning of these coefficients is proposed. The fuzzy controller is optimally tuned using Particle Swarm Optimization (PSO) taking into accounts both the tracking error and the controller output signal range. The simulation of several examples of systems with moderate to severe nonlinearities demonstrate that the proposed approach offers improved control performance when benchmarked to L 1 adaptive controller with fixed filter coefficients.

Robustness and Performance of Adaptive Suppression of Unknown Periodic Disturbances

Abstract: In recent years, a class of adaptive schemes has been developed for suppressing periodic disturbance signals with unknown frequencies, phases, and amplitudes. The stability and robustness of these schemes with respect to inevitable unmodeled dynamics and noise disturbances in the absence of persistently exciting signals has not been established despite successful simulation results and implementations. The purpose of this technical note is to propose a robust adaptive scheme for rejection of unknown periodic components of the disturbance and analyze its stability and performance properties. First, we consider the ideal case (non-adaptive) when complete information about the characteristics of the disturbance is available. We show that the rejection of periodic terms may lead to amplification of output noise and in some cases lead to a worse output performance. The way to avoid such undesirable noise amplification is to increase the size of the feedback control filter in order to have the flexibility to achieve rejection of the periodic disturbance terms while minimizing the effect of the noise on the output. The increased filter order leads to an over-parameterized scheme where persistence of excitation is no longer possible, and this shortcoming makes the use of robust adaptation essential. With this important insight in mind, the coefficients of the feedback filter whose size is over parameterized are adapted using a robust adaptive law. We show analytically that the proposed robust adaptive control scheme guarantees stability, performance and robustness with respect to unmodeled dynamics and bounded broadband noise disturbances. We use numerical simulations to demonstrate the results.

An adaptive fading Kalman filter based on Mahalanobis distance

Abstract: An adaptive Kalman filter with fading factor is derived to address the modeling errors. In any recursion of the proposed filter, a judging index is defined as the square of the Mahalanobis distance...

Adaptive dualISO HDR reconstruction

Abstract: With the development of modern image sensors enabling flexible image acquisition, single shot high dynamic range (HDR) imaging is becoming increasingly popular. In this work, we capture single shot HDR images using an imaging sensor with spatially varying gain/ISO. This allows all incoming photons to be used in the imaging. Previous methods on single shot HDR capture use spatially varying neutral density (ND) filters which lead to wasting incoming light. The main technical contribution in this work is an extension of previous HDR reconstruction approaches for single shot HDR imaging based on local polynomial approximations (Kronander et al., Unified HDR reconstruction from raw CFA data, 2013; Hajisharif et al., HDR reconstruction for alternating gain (ISO) sensor readout, 2014). Using a sensor noise model, these works deploy a statistically informed filtering operation to reconstruct HDR pixel values. However, instead of using a fixed filter size, we introduce two novel algorithms for adaptive filter kernel selection. Unlike a previous work, using adaptive filter kernels (Signal Process Image Commun 29(2):203–215, 2014), our algorithms are based on analyzing the model fit and the expected statistical deviation of the estimate based on the sensor noise model. Using an iterative procedure, we can then adapt the filter kernel according to the image structure and the statistical image noise. Experimental results show that the proposed filter de-noises the noisy image carefully while well preserving the important image features such as edges and corners, outperforming previous methods. To demonstrate the robustness of our approach, we have exploited input images from raw sensor data using a commercial off-the-shelf camera. To further analyze our algorithm, we have also implemented a camera simulator to evaluate different gain patterns and noise properties of the sensor.

Image denoising using optimally weighted bilateral filters: A sure and fast approach

Abstract: The bilateral filter is known to be quite effective in denoising images corrupted with small dosages of additive Gaussian noise. The denoising performance of the filter, however, is known to degrade quickly with the increase in noise level. Several adaptations of the filter have been proposed in the literature to address this shortcoming, but often at a substantial computational overhead. In this paper, we report a simple pre-processing step that can substantially improve the denoising performance of the bilateral filter, at almost no additional cost. The modified filter is designed to be robust at large noise levels, and often tends to perform poorly below a certain noise threshold. To get the best of the original and the modified filter, we propose to combine them in a weighted fashion, where the weights are chosen to minimize (a surrogate of) the oracle mean-squared-error (MSE). The optimally-weighted filter is thus guaranteed to perform better than either of the component filters in terms of the MSE, at all noise levels. We also provide a fast algorithm for the weighted filtering. Visual and quantitative denoising results on standard test images are reported which demonstrate that the improvement over the original filter is significant both visually and in terms of PSNR. Moreover, the denoising performance of the optimally-weighted bilateral filter is competitive with the computation-intensive non-local means filter.

Singular Value Decomposition-based Robust Cubature Kalman Filtering for an Integrated GPS/SINS Navigation System

Abstract: A new nonlinear robust filter is proposed in this paper to deal with the outliers of an integrated Global Positioning System/Strapdown Inertial Navigation System (GPS/SINS) navigation system. The influence of different design parameters for an H∞ cubature Kalman filter is analysed. It is found that when the design parameter is small, the robustness of the filter is stronger. However, the design parameter is easily out of step in the Riccati equation and the filter easily diverges. In this respect, a singular value decomposition algorithm is employed to replace the Cholesky decomposition in the robust cubature Kalman filter. With large conditions for the design parameter, the new filter is more robust. The test results demonstrate that the proposed filter algorithm is more reliable and effective in dealing with the outliers in the data sets produced by the integrated GPS/SINS system.

Global Color Sparseness and a Local Statistics Prior for Fast Bilateral Filtering

Abstract: The property of smoothing while preserving edges makes the bilateral filter a very popular image processing tool. However, its non-linear nature results in a computationally costly operation. Various works propose fast approximations to the bilateral filter. However, the majority does not generalize to vector input as is the case with color images. We propose a fast approximation to the bilateral filter for color images. The filter is based on two ideas. First, the number of colors, which occur in a single natural image, is limited. We exploit this color sparseness to rewrite the initial non-linear bilateral filter as a number of linear filter operations. Second, we impose a statistical prior to the image values that are locally present within the filter window. We show that this statistical prior leads to a closed-form solution of the bilateral filter. Finally, we combine both ideas into a single fast and accurate bilateral filter for color images. Experimental results show that our bilateral filter based on the local prior yields an extremely fast bilateral filter approximation, but with limited accuracy, which has potential application in real-time video filtering. Our bilateral filter, which combines color sparseness and local statistics, yields a fast and accurate bilateral filter approximation and obtains the state-of-the-art results.

Prototype Filter Design Approaches for Near Perfect Reconstruction Cosine Modulated Filter Banks - A Review

Abstract: Cosine Modulated Filter Banks (CMFB) are very popular among the different maximally decimated filter banks due to their design ease and simplicity in implementation and the property that all the coefficients of all the filters are real. All the analysis and synthesis filters are derived from one or two prototype filters. Hence, recently, the design of the prototype filter in a CMFB has become a subject of interest in the field of multirate signal processing. Perfect Reconstruction (PR) filter banks are those which can produce at the output, a weighted delayed version of the input. But in most of the applications a near perfect reconstruction (NPR) is sufficient. This can reduce the computational complexity. Different approaches developed for the efficient and optimal design of the prototype filter in a NPR orthogonal CMFB are studied, classified and summarized in this paper. In today's applications, less space and low power consumption are very essential. When the different design approaches for the design of the prototype filter in CMFB are compared, it is observed that the one using frequency response masking(FRM) and meta-heuristic optimization techniques gives better performance in terms of implementation complexity, which in turn can lead to reduced chip size and power consumption. It is hoped that this review will be highly beneficial to the researchers working in the area of multirate signal processing. At the end, we also propose some novel design approaches for the design of low complexity prototype filter using FRM technique.

A study of recursive least squares (RLS) adaptive filter algorithm in noise removal from ECG signals

Abstract: Electrocardiogram (ECG) is a diagnostic procedure that measures and records the electrical activity of heart in detail. By reviewing an ECG report, one's condition of heart can be evaluated. But ECG signals are often affected and altered by the presence of various noises that degrade the accuracy of an ECG signal and thus misrepresents the recorded data. To filter out these noises conventional digital filters have been used for decades. Yet noise cancellation with finite and determined coefficients has often been unsuccessful due to the non-stationary nature of ECG signal. Adaptive filters adapt their filter coefficients with the continuous change of signal using adaptive algorithms, providing the optimum noise removal features for non-stationary signals like ECG. In this study, the adaptive filter algorithm, RLS has been used in cancellation of various noises in ECG signals. We have also performed noise removal using LMS adaptive filter algorithm to compare the performance of RLS algorithm. We have used MATLAB® to simulate different noise signals and process the noises. The ECG signals used here have been taken from the PhysioNet ECG-ID database. The simulation results depict that RLS algorithm renders a much better performance in removing noises from the ECG signals than LMS algorithm.

A novel neural network-based adaptive control for a class of uncertain nonlinear systems in strict-feedback form

Abstract: In this paper, a novel robust adaptive tracking control approach is presented for a class of strict-feedback single-input single-output nonlinear systems. By employing radial basis function neural network to account for system uncertainties, the proposed scheme is developed by combining “command filter” and “minimal learning parameter” techniques. The main advantages of the proposed controller are that: (1) the problem of “explosion of complexity” inherent in the conventional backstepping method is avoided; (2) the problem of “dimensionality curse” is solved, and only one adaptive parameter needs to be updated online. These advantages result in a much simpler adaptive control algorithm, which is convenient to implement in applications. In addition, stability analysis shows that uniform ultimate boundedness of the solution of the closed-loop system can be guaranteed. Simulation results demonstrate the effectiveness of the proposed scheme.

Gaussian Sum Reapproximation for Use in a Nonlinear Filter

Abstract: A new method has been developed to approximate one Gaussian sum by another. This algorithm is being developed as part of an effort to generalize the concept of a particle filter. In a traditional particle filter, the underlying probability density function is described by particles: Dirac delta functions with infinitesimal covariances. This paper develops an important component of a more general filter, which uses a Gaussian sum with “fattened” finite-covariance “blobs” (i.e., Gaussian components), which replace infinitesimal particles. The goal of such a filter is to save computational effort by using many fewer Gaussian components than particles. Most of the techniques necessary for this type of filter exist. The one missing technique is a resampling algorithm that bounds the covariance of each Gaussian component while accurately reproducing the original probability distribution. The covariance bounds keep the blobs from becoming too “fat” to ensure low truncation error in extended Kalman filter or unsc...

Kernel Least Mean Square with Single Feedback

Abstract: In this letter, a novel kernel adaptive filtering algorithm, namely the kernel least mean square with single feedback (SF-KLMS) algorithm, is proposed. In SF-KLMS, only a single delayed output is used to update the weights in a recurrent fashion. The use of past information accelerates the convergence rate significantly. Compared with the kernel adaptive filter using multiple feedback, SF-KLMS has a more compact and efficient structure. Simulations in the context of time-series prediction and nonlinear regression show that SF-KLMS outperforms not only the kernel adaptive filter with multiple feedback but also the kernel adaptive filter without feedback in terms of convergence rate and mean square error.

The multiple model labeled multi-Bernoulli filter

Abstract: In many applications, multi-object tracking algorithms are either required to handle different types of objects or rapidly maneuvering objects. In both cases, the usage of multiple motion models is essential to obtain excellent tracking results. In the field of random finite set based tracking algorithms, the Multiple Model Probability Hypothesis Density (MM-PHD) filter has recently been applied to tackle this problem. However, the MM-PHD filter requires error-prone post-processing to obtain target tracks and its cardinality estimate is fluctuating. The Labeled Multi-Bernoulli (LMB) filter is an accurate and computationally efficient approximation of the multi-object Bayes filter which provides target tracks. In applications using only a single motion model, LMB filter has been shown to significantly outperform the PHD filter. In this contribution, the Multiple Model Labeled Multi-Bernoulli (MM-LMB) filter is proposed. The MM-LMB filter is applied to scenarios with rapidly maneuvering objects and its performance is compared to the single model LMB filter using simulated data.