From Benes Filter to General Nonlinear Filters
组织者
演讲者
孙泽钜
时间
2023年04月10日 14:30 至 15:30
地点
Online
摘要
As a finite-dimensional nonlinear filter, Benes filter allows people to calculate the conditional probability density function exactly and recursively. In the meanwhile, there are many existing methods to analyze the asymptotic behavior of Benes filter. For general nonlinear filters, the conditional probability density function cannot be calculated exactly. However, it is still possible to analyze their asymptotic behavior with the methods used in Benes filter. In this talk, I will briefly review the derivation of the exact formula of Benes filter, and introduce the idea of generalizing the methods in the asymptotic analysis of Benes filter to general nonlinear filters.