In a Bayesian Approach the Kalman filter can be regarded as recursive Bayesian estimator and be described as Bayesian dynamic network. The linear dynamic system discretized in the time domain follows a first order hidden Markov process where uncertainties in the system model and the easurement model are assumed to be Gaussian and modeled as uncorrelated white noise processes. As the assumption of linearity and Gaussianity is often violated a Bayesian approach of an extended skewed Kalman filter is derived which allows to consider a nonlinear dynamic system excited by a process with skew-normal probability distributions.     «

In a Bayesian Approach the Kalman filter can be regarded as recursive Bayesian estimator and be described as Bayesian dynamic network. The linear dynamic system discretized in the time domain follows a first order hidden Markov process where uncertainties in the system model and the easurement model are assumed to be Gaussian and modeled as uncorrelated white noise processes. As the assumption of linearity and Gaussianity is often violated a Bayesian approach of an extended skewed Kalman filte...     »