Particle Filter for Systems with Stationary State Processes
组织者
演讲者
孙泽钜
时间
2024年02月16日 21:30 至 22:00
地点
Online
摘要
In this talk, we will introduce a novel particle filtering algorithm for partially observed dynamical systems with stationary states. Instead of generating independent trajectories of state processes, we use variables in a long trajectory of the stationary state as particles in this algorithm, and the convergence of this algorithm is guaranteed by Birkhoff’s ergodic theorem. A suboptimal filtering algorithm is proposed for this kind of filtering system, and numerical experiments show that the computational cost of particle filter is sharply reduced, while the mean square error is only slightly increasing, in comparison with classical resampling particle filters.