Numerical Study of a Hybrid Particle Filter
نویسندگان
چکیده
The particle filter method has been used in data assimilation problems for estimating states of nonlinear dynamic systems when both system errors and observation errors are present, and are possibly non-Gaussian. Weighted particles are used to represent the probability density function of the system states at any time. The particles are propagated through the system evolvement and the weights are updated. The method becomes very inefficient when the system dimension is high and the model is large and complicated. The notorious phenomenon of the method is the so called “particle degeneration” where the particles collapse until only one particle carries the majority of the weights. In this paper, we implement an improved hybrid particle filter method which aims to tackle the particle degeneration problem. The improved method introduces an auxiliary procedure which redistributes particles according to the information from the newest observation. The hybrid particle filter is tested on the Lorenz-63 model, the numerical analysis shows the method is effective for alleviating the particle degeneration problem.
منابع مشابه
HyPE: Hybrid Particle-Element Approach for Recursive Bayesian Searching and Tracking
This paper presents a hybrid particle-element approach, HyPE, suitable for recursive Bayesian searching-andtracking (SAT). The hybrid concept, to synthesize two recursive Bayesian estimation (RBE) methods to represent and maintain the belief about all states in a dynamic system, is distinct from the concept behind “mixed approaches”, such as Rao-Blackwellized particle filtering, which use diffe...
متن کاملUnscented Auxiliary Particle Filter Implementation of the Cardinalized Probability Hypothesis Density Filters
The probability hypothesis density (PHD) filter suffers from lack of precise estimation of the expected number of targets. The Cardinalized PHD (CPHD) recursion, as a generalization of the PHD recursion, remedies this flaw and simultaneously propagates the intensity function and the posterior cardinality distribution. While there are a few new approaches to enhance the Sequential Monte Carlo (S...
متن کاملAn Efficient Target Tracking Algorithm Based on Particle Filter and Genetic Algorithm
In this paper, we propose an efficient hybrid Particle Filter (PF) algorithm for video tracking by employing a genetic algorithm to solve the sample impoverishment problem. In the presented method, the object to be tracked is selected by a rectangular window inside which a few numbers of particles are scattered. The particles’ weights are calculated based on the similarity between feature vecto...
متن کاملBiomedical Image Denoising Based on Hybrid Optimization Algorithm and Sequential Filters
Background: Nowadays, image de-noising plays a very important role in medical analysis applications and pre-processing step. Many filters were designed for image processing, assuming a specific noise distribution, so the images which are acquired by different medical imaging modalities must be out of the noise. Objectives: This study has focused on the sequence filters which are selected ...
متن کاملThe Particle Filter and Extended Kalman Filter methods for the structural system identification considering various uncertainties
Structural system identification using recursive methods has been a research direction of increasing interest in recent decades. The two prominent methods, including the Extended Kalman Filter (EKF) and the Particle Filter (PF), also known as the Sequential Monte Carlo (SMC), are advantageous in this field. In this study, the system identification of a shake table test of a 4-story steel struct...
متن کامل