Classical Nonlinear Filtering Research Articles

Quantized genetic resampling particle filtering for vision-based ground moving target tracking

This paper concentrates on the quantized measurements fusion problem. Because of the non-Gaussian property of quantized measurements, classical nonlinear filters become not applicable, especially when the quantization noise is large. Motivated by the above problem, we propose a novel particle-based filtering named quantized genetic resampling particle filtering (QGRPF) in order to fuse quantized measurements optimally and improve estimation accuracy. Unlike the Gaussian assumption in traditional filers, the probability density function (PDF) of sensor measurement is modeled here as a discrete PDF that consists of a series of Dirac impulses. The particle filtering is employed to estimate the state, where the posterior of the state based on quantized measurements is approximated and updated by a set of weighted particles. As the likelihood function is a multi-dimensional integral of Gaussian density, which has no analytical solution, Genz's transformation and quasi-Monte Carlo method are combined to calculate the integral numerically. By integrating genetic resampling method into the particle filtering method, the proposed method can avoid particle degeneracy, in the meantime, guarantee particle diversity. The proposed QGRPF is demonstrated with an illustrative vision-based ground moving target (GMT) tracking example. Simulation results show the proposed approach is effective in fusing quantized measurements.

Bearings-only multi-target tracking using an improved labeled multi-Bernoulli filter

Most classical bearing-only target tracking algorithms model the measurement likelihood by one Gaussian distribution. The effectiveness of one Gaussian distribution model relies heavily on the accuracy of the predicted target position. However, due to the high nonlinearity of the bearing-only measurement, the predicted target position is mostly inaccurate before the target state observability is established. As a consequence, some classical nonlinear filters become not applicable for tracking bearing-only targets, especially when the measurements of multiple targets and clutter are present. The published bearings-only multiple-target tracking algorithms suffer from either the estimation inaccuracy or lack of track trajectories. Motivated by the problems mentioned above, we propose an improved labeled multi-Bernoulli filter for the goal of reducing estimation error under the premise that track trajectories are guaranteed. The proposed method divides the bearing measurement uncertainty into several measurement components that the measurement likelihood can be approximated by a Gaussian mixture. By assigning each track a unique label, the previous scan estimations and current scan measurements are associated and the track trajectories become available. Simulation results show that the proposed method considerably reduces estimation error. Further, various scenario parameters are investigated to validate the effectiveness of the proposed method.

Novel Fuzzy Filters for Noise Suppression from Digital Grey and Color Images

Images are rich information carriers and (such as medical images) are normally contaminated by additive and substitutive noise which makes the extraction of features (and clinical data analysis) difficult. Hence to enhance the image quality prior to post processing, image pre-processing operations such as de-noising with linear and non-linear filters have been applied traditionally. Recently nonlinear filtering techniques have been assumed a lot of significance as they are capable of suppressing the effects of substitutive (salt and pepper impulsive noise of low to high noise levels) and additive (Gaussian noise of low to medium noise levels) noise types and to preserve the important signal/image details such as edges and fine details and suppress the degradations occurring at the time of image/signal formation or transmission through nonlinear channels, during storage and retrieval. Broadly speaking, image filters exist in transform and spatial domains. Spatial domain nonlinear filters are more versatile than their counterparts, namely linear filters. Spatial domain nonlinear fuzzy classical filters are simply modification/extension of the classical median and moving average filtering approaches, offer several advantages over classical nonlinear filters, and using simple fuzzy rules it is easy to realize them. They are also capable of reasoning with vague and uncertain information. Work presented in this paper deals with nonlinear median based and linear average based fuzzy filters and aims at fulfilling three objectives, viz; (i) To systematically study the performance of classical nonlinear median and fuzzy median and average filters for the removal of impulse and Gaussian noise from gray and color images that have been corrupted from low to high values of noise and to present an experimental review to identify the best algorithm within the frame work of classical fuzzy median filters. (ii)To propose : (a) an impulse classifier based fuzzy switching median filter and (b) the design of a multi pass cascaded fuzzy filter for noise cancellation, and explore their applications to reduce noise in images with random and impulse characteristics. Finally to conclude the work a comparative study is done and the computational aspects are analyzed with the help of mean square error (MSE), peak signal to noise ratio (PSNR), and 2D correlation (COR) and some future solutions are proposed.

다중 UAV에서 측정된 거리차 정보를 이용한 선형 강인 표적추적 필터 설계

This paper addresses a new passive target tracking problem using the range differences measured by cooperative UAVs. In order to solve the range difference based passive target tracking problem within the framework of linear robust state estimation, the uncertain linear measurement model which contains the stochastic parameter uncertainty is derived by using the noisy range difference measurements. To cope with the performance degradation due to the stochastic parameter uncertainty, the recently developed non-conservative robust Kalman filtering technique [1] is applied. For the cruciform formation flying UAVs, the relationship between the target tracking performance and the measurement errors is quantitatively analyzed. The proposed filter has practical advantages over the classical nonlinear filters because, for its recursive linear structure, it can provide satisfactory convergence properties and is suitable for real-time multiple UAVs applications. Through the simulations, the usefulness of the proposed method is demonstrated.

Moving window-based double haar wavelet transform for image processing

Image denoising is a lively research field. The classical nonlinear filters used for image denoising, such as median filter, are based on a local analysis of the pixels within a moving window. Recently, the research of image denoising has been focused on the wavelet domain. Compared to the classical nonlinear filters, it is based on a global multiscale analysis of images. Apparently, the wavelet transform can be embedded in a moving window. Thus, a moving window-based local multiscale analysis is obtained. In this paper, based on the Haar wavelet, a class of nonorthogonal multi-channel filter bank with its corresponding wavelet shrinkage called Lee shrinkage is derived. As a special case of this filter bank, the double Haar wavelet transform is introduced. Examples show that it is suitable for a moving window-based local multiscale analysis used for image denoising, edge detection, and edge enhancement.