We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

math.OC

Change to browse by:

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Mathematics > Optimization and Control

Title: The reproducing kernel Hilbert spaces underlying linear SDE Estimation, Kalman filtering and their relation to optimal control

Abstract: It is often said that control and estimation problems are in duality. Recently, in (Aubin-Frankowski,2021), we found new reproducing kernels in Linear-Quadratic optimal control by focusing on the Hilbert space of controlled trajectories, allowing for an elegant handling of state constraints and meeting points. We now extend this view to estimation problems where it is known that kernels are the covariances of stochastic processes. Here, the Markovian Gaussian processes stem from the linear stochastic differential equations describing the continuous-time dynamics and observations. Taking extensive care to require minimal invertibility requirements on the operators, we give novel explicit formulas for these covariances. We also determine their reproducing kernel Hilbert spaces, stressing the symmetries between a space of forward-time trajectories and a space of backward-time information vectors. We then recover the Kalman filter and smoother formulas. Extension to discrete-time observations or infinite-dimensional state, tough technical, would be straightforward.
Subjects: Optimization and Control (math.OC); Probability (math.PR)
MSC classes: 46E22, 60G35, 62M20
Cite as: arXiv:2208.07030 [math.OC]
  (or arXiv:2208.07030v1 [math.OC] for this version)

Submission history

From: Pierre-Cyril Aubin-Frankowski [view email]
[v1] Mon, 15 Aug 2022 07:06:06 GMT (27kb)

Link back to: arXiv, form interface, contact.