A study of Bayesian filtering methods applied for biomedical signal diagnosis

In this article, we address the issue of tracking a target in both straight-line and curvilinear motion. Common examples of tracking algorithms include the Kalman filter, the extended Kalman, the unscented Kalman, and the particle filter. In contrast to the Kalman filter, which can only be used to linear or linearized processes and measurement systems, particle filters may be applied to nonlinear ones. The particle filter can deal with noise that doesn't follow a Gaussian distribution, whereas the Kalman filter can only deal with uncertainty that follows a Gaussian distribution. Without making any assumptions about the distribution of mistakes, particle filters provide a Monte Carlo estimate of the probability density of a system. Here we take a look at the cutting edge of dynamic Bayesian linear and nonlinear models and Kalman filtering. As well as discussing some of the basic accomplishments, such as the derivation of the Kalman filtering equations, this article discusses recent breakthroughs in Kalman filter models and their extensions, including non-Gaussian state-space models. We investigate Bayesian approaches to parameter learning in state-space models, which often use Markov chain Monte Carlo and sequential Monte Carlo methods. Particle filtering and Bayesian particle learning are presented and reviewed, along with its application to state space models. And also, a new method based on ECG and MR artifact modeling using Bayesian filtering is presented.