Gaussian Process Latent Force Models for Learning and Stochastic Control of Physical Systems

Simo SärkkäMauricio A. ÁlvarezNeil D. Lawrence
IEEE Transactions on Automatic Control, IEEE 64(7):2953-2960, 2018.

Abstract

This paper is concerned with learning and stochastic control in physical systems that contain unknown input signals. These unknown signals are modeled as Gaussian processes (GP) with certain parameterized covariance structures. The resulting latent force models can be seen as hybrid models that contain a first-principle physical model part and a nonparametric GP model part. We briefly review the statistical inference and learning methods for this kind of models, introduce stochastic control methodology for these models, and provide new theoretical observability and controllability results for them.

Cite this Paper


BibTeX
@Article{Sarkka-lfmcontrol18, title = {Gaussian Process Latent Force Models for Learning and Stochastic Control of Physical Systems}, author = {Simo Särkkä and Mauricio A. Álvarez and Neil D. Lawrence}, journal = {IEEE Transactions on Automatic Control}, pages = {2953--2960}, year = {2018}, volume = {64}, number = {7}, publisher = {IEEE}, doi = {10.1109/TAC.2018.2874749}, abstract = {This paper is concerned with learning and stochastic control in physical systems that contain unknown input signals. These unknown signals are modeled as Gaussian processes (GP) with certain parameterized covariance structures. The resulting latent force models can be seen as hybrid models that contain a first-principle physical model part and a nonparametric GP model part. We briefly review the statistical inference and learning methods for this kind of models, introduce stochastic control methodology for these models, and provide new theoretical observability and controllability results for them. } }
Endnote
%0 Journal Article %T Gaussian Process Latent Force Models for Learning and Stochastic Control of Physical Systems %A Simo Särkkä %A Mauricio A. Álvarez %A Neil D. Lawrence %J IEEE Transactions on Automatic Control %D 2018 %F Sarkka-lfmcontrol18 %I IEEE %P 2953--2960 %R 10.1109/TAC.2018.2874749 %V 64 %N 7 %X This paper is concerned with learning and stochastic control in physical systems that contain unknown input signals. These unknown signals are modeled as Gaussian processes (GP) with certain parameterized covariance structures. The resulting latent force models can be seen as hybrid models that contain a first-principle physical model part and a nonparametric GP model part. We briefly review the statistical inference and learning methods for this kind of models, introduce stochastic control methodology for these models, and provide new theoretical observability and controllability results for them.
RIS
TY - JOUR TI - Gaussian Process Latent Force Models for Learning and Stochastic Control of Physical Systems AU - Simo Särkkä AU - Mauricio A. Álvarez AU - Neil D. Lawrence DA - 2018/10/08 ID - Sarkka-lfmcontrol18 PB - IEEE VL - 64 IS - 7 SP - 2953 EP - 2960 DO - 10.1109/TAC.2018.2874749 AB - This paper is concerned with learning and stochastic control in physical systems that contain unknown input signals. These unknown signals are modeled as Gaussian processes (GP) with certain parameterized covariance structures. The resulting latent force models can be seen as hybrid models that contain a first-principle physical model part and a nonparametric GP model part. We briefly review the statistical inference and learning methods for this kind of models, introduce stochastic control methodology for these models, and provide new theoretical observability and controllability results for them. ER -
APA
Särkkä, S., Álvarez, M.A. & Lawrence, N.D.. (2018). Gaussian Process Latent Force Models for Learning and Stochastic Control of Physical Systems. IEEE Transactions on Automatic Control 64(7):2953-2960 doi:10.1109/TAC.2018.2874749

Related Material