Kernel adaptive filter

About: Kernel adaptive filter is a research topic. Over the lifetime, 8771 publications have been published within this topic receiving 142711 citations.

Adaptive robust Kalman filter for relative navigation using global position system

Abstract: An adaptive robust Kalman filter algorithm is derived to account for both process noise and measurement noise uncertainty. The adaptive algorithm estimates process noise covariance based on the recursive minimisation of the difference between residual covariance matrix given by the filter and that calculated from time-averaging of the residual sequence generated by the filter at each time step. A recursive algorithm is proposed based on both Massachusetts Institute of Technology (MIT) rule and typical non-linear extended Kalman filter equations for minimising the difference. The measurement update using a robust technique to minimise a criterion function originated from Huber filter. The proposed adaptive robust Kalman filter has been successfully implemented in relative navigation using global position system for spacecraft formation flying in low earth orbit, with real-orbit perturbations and non-Gaussian random measurement errors. The numerical simulation results indicate that the proposed adaptive robust filter can provide better relative navigation performance in terms of accuracy and robustness as compared with previous filter algorithms.

Video encoding by filter selection

Abstract: A method and a device are described for selecting between multiple available filters in an encoder to provide a frame having a low error and distortion rate. For each full and sub pixel position, determining whether to use an alternative filter over the default filter during interpolation by estimating the rate distortion gain of using each filter and signaling to the decoder the optimal filter(s) applied to each full and sub-pixel position. In one embodiment, identifying a reference frame and a current frame, interpolating the reference frame using a default filter to create a default interpolated frame, interpolating the reference frame using an alternative filter to create an alternative interpolated frame, determining for each sub-pixel position whether to use the default filter or the alternative filter based on a minimal cost to generate a final reference frame.

Adaptive Kuwahara filter

Abstract: A new filter was created by improving the standard Kuwahara filter. It allows more efficient noise reduction without blurring the edges and image preparation for segmentation and further analyses operations. One of the biggest and most common restrictions encountered in filter algorithms is the need for a declarative definition of the filter window size or the number of iterations that an operation should be repeated. In the case of the proposed solution, we are dealing with automatic adaptation of the algorithm to the local environment of each pixel in the processed image.

Adaptive Fuzzy Filter and Its Application to Image Enhancement

Abstract: This chapter describes the design and evaluation of a novel adaptive fuzzy filter, and discusses its application to image enhancement. Most traditional edge detectors can perform well for uncorrupted images but are highly sensitive to impulse noise, so they can not work efficiently for blurred images. The proposed adaptive fuzzy filter consists of two major mechanisms: Adaptive Weighted Fuzzy Mean (AWFM) filter and Fuzzy Normed Inference System (FNIS) to realize the function of edge detection for smeared images. The membership functions of all fuzzy sets used in this filter can be adaptively determined for different images. Moreover, the adaptive fuzzy filter is capable of converting blurred edges to clear ones and suppressing noise at the same time. According to the experimental results, it works well in full range of random impulse noise probability and performs efficiently in the environment of mixed Gaussian impulse noise. This chapter also analytically evaluates the important properties of the filter to show its high performance in general cases.

A modified Sage-Husa adaptive Kalman filter for denoising Fiber Optic Gyroscope signal

Abstract: Fiber Optic Gyroscope (FOG) is a key component in Inertial Navigation System. The performance of FOG degrades due to different types of random errors in the measured signal. Although Kalman filter and its variants like Sage-Husa Kalman filters are being used to denoise the Gyroscope signal the performance of Kalman filter is limited by the initial values of measurement and process noise covariance matrix, and transition matrix. To address this problem, this paper uses modified Sage-Husa adaptive Kalman filter to denoise the FOG signal. In this work, the random error of fiber optic gyroscope is modeled using a first order auto regressive (AR) model and used the coefficients of the model to initialize the transition matrix of Sage-Husa Adaptive Kalman filter. Allan variance analysis is used to quantify the random errors of the measured and denoised signal. The performance of proposed algorithm is compared with conventional Kalman filter and the simulation results show that the modified Sage-Husa adaptive Kalman filter (SHAKF) algorithm outperforms the conventional Kalman filter technique while denoising FOG signal.