Filtering for fast mean-reverting processes

Research output: Contribution to journalArticle

Abstract

We consider nonlinear filtering applications to target tracking based on a vector of multi-scaled models where some of the processes are rapidly mean reverting to their local equilibria. We focus attention on target tracking problems because multiple scaled models with fast mean-reversion (FMR) are a simple way to model latency in the response of tracking systems. The main results of this paper show that nonlinear filtering algorithms for multi-scale models with FMR states can be simplified significantly by exploiting the FMR structures, which leads to a simplified Baum-Welch recursion that is of reduced dimension. We implement the simplified algorithms with numerical simulations and discuss their efficiency and robustness.

Original languageEnglish (US)
Pages (from-to)155-176
Number of pages22
JournalAsymptotic Analysis
Volume70
Issue number3-4
DOIs
StatePublished - 2010

Fingerprint

Mean Reversion
Process Mean
Nonlinear Filtering
Filtering
Target Tracking
Multiscale Model
Local Equilibrium
Tracking System
Recursion
Latency
Model
Robustness
Numerical Simulation

Keywords

  • fast mean reversion
  • Kramers-Smoluchowski
  • nonlinear filtering
  • tracking

ASJC Scopus subject areas

  • Mathematics(all)

Cite this

Filtering for fast mean-reverting processes. / Papanicolaou, Andrew.

In: Asymptotic Analysis, Vol. 70, No. 3-4, 2010, p. 155-176.

Research output: Contribution to journalArticle

@article{cebec37572fd4fae8c4c8f1501e4d17a,
title = "Filtering for fast mean-reverting processes",
abstract = "We consider nonlinear filtering applications to target tracking based on a vector of multi-scaled models where some of the processes are rapidly mean reverting to their local equilibria. We focus attention on target tracking problems because multiple scaled models with fast mean-reversion (FMR) are a simple way to model latency in the response of tracking systems. The main results of this paper show that nonlinear filtering algorithms for multi-scale models with FMR states can be simplified significantly by exploiting the FMR structures, which leads to a simplified Baum-Welch recursion that is of reduced dimension. We implement the simplified algorithms with numerical simulations and discuss their efficiency and robustness.",
keywords = "fast mean reversion, Kramers-Smoluchowski, nonlinear filtering, tracking",
author = "Andrew Papanicolaou",
year = "2010",
doi = "10.3233/ASY-2010-1011",
language = "English (US)",
volume = "70",
pages = "155--176",
journal = "Asymptotic Analysis",
issn = "0921-7134",
publisher = "IOS Press",
number = "3-4",

}

TY - JOUR

T1 - Filtering for fast mean-reverting processes

AU - Papanicolaou, Andrew

PY - 2010

Y1 - 2010

N2 - We consider nonlinear filtering applications to target tracking based on a vector of multi-scaled models where some of the processes are rapidly mean reverting to their local equilibria. We focus attention on target tracking problems because multiple scaled models with fast mean-reversion (FMR) are a simple way to model latency in the response of tracking systems. The main results of this paper show that nonlinear filtering algorithms for multi-scale models with FMR states can be simplified significantly by exploiting the FMR structures, which leads to a simplified Baum-Welch recursion that is of reduced dimension. We implement the simplified algorithms with numerical simulations and discuss their efficiency and robustness.

AB - We consider nonlinear filtering applications to target tracking based on a vector of multi-scaled models where some of the processes are rapidly mean reverting to their local equilibria. We focus attention on target tracking problems because multiple scaled models with fast mean-reversion (FMR) are a simple way to model latency in the response of tracking systems. The main results of this paper show that nonlinear filtering algorithms for multi-scale models with FMR states can be simplified significantly by exploiting the FMR structures, which leads to a simplified Baum-Welch recursion that is of reduced dimension. We implement the simplified algorithms with numerical simulations and discuss their efficiency and robustness.

KW - fast mean reversion

KW - Kramers-Smoluchowski

KW - nonlinear filtering

KW - tracking

UR - http://www.scopus.com/inward/record.url?scp=78649606928&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=78649606928&partnerID=8YFLogxK

U2 - 10.3233/ASY-2010-1011

DO - 10.3233/ASY-2010-1011

M3 - Article

VL - 70

SP - 155

EP - 176

JO - Asymptotic Analysis

JF - Asymptotic Analysis

SN - 0921-7134

IS - 3-4

ER -