Nuclear norm minimization for blind subspace identification (N2BSID)

Dexter Scobee; Lillian Ratliff; Roy Dong; Henrik Ohlsson; Michel Verhaegen; S. Shankar Sastry

doi:10.1109/CDC.2015.7402521

Nuclear norm minimization for blind subspace identification (N2BSID)

Dexter Scobee, Lillian Ratliff, Roy Dong, Henrik Ohlsson, Michel Verhaegen, S. Shankar Sastry

Team Raf Van de Plas

Research output: Chapter in Book/Conference proceedings/Edited volume › Conference contribution › Scientific › peer-review

5 Citations (Scopus)

Abstract

In many practical applications of system identification, it is not feasible to measure both the inputs applied to the system as well as the output. In such situations, it is desirable to estimate both the inputs and the dynamics of the system simultaneously; this is known as the blind identification problem. In this paper, we provide a novel extension of subspace methods to the blind identification of multiple-input multiple-output linear systems. We assume that our inputs lie in a known subspace, and we are able to formulate the identification problem as rank constrained optimization, which admits a convex relaxation. We show the efficacy of this formulation with a numerical example.

Original language	English
Title of host publication	Proceedings of the 54th IEEE Conference on Decision and Control (CDC 2015)
Editors	Yoshito Ohta, Mitsuji Sampei, Shigemasa Takai
Place of Publication	Piscataway, NJ, USA
Publisher	IEEE
Pages	2127-2132
ISBN (Electronic)	978-1-4799-7886-1
DOIs	https://doi.org/10.1109/CDC.2015.7402521
Publication status	Published - 2015
Event	54th IEEE Conference on Decision and Control, CDC 2015 - Osaka, Japan Duration: 15 Dec 2015 → 18 Dec 2015 Conference number: 54

Conference

Conference	54th IEEE Conference on Decision and Control, CDC 2015
Abbreviated title	CDC 2015
Country/Territory	Japan
City	Osaka
Period	15/12/15 → 18/12/15

Keywords

Optimization
Mathematical model
Finite impulse response filters
Minimization
Maximum likelihood estimation
MIMO
Numerical models

Access to Document

10.1109/CDC.2015.7402521

Cite this

@inproceedings{356e4623da9b41d4ab7c8a9b037b284f,

title = "Nuclear norm minimization for blind subspace identification (N2BSID)",

abstract = "In many practical applications of system identification, it is not feasible to measure both the inputs applied to the system as well as the output. In such situations, it is desirable to estimate both the inputs and the dynamics of the system simultaneously; this is known as the blind identification problem. In this paper, we provide a novel extension of subspace methods to the blind identification of multiple-input multiple-output linear systems. We assume that our inputs lie in a known subspace, and we are able to formulate the identification problem as rank constrained optimization, which admits a convex relaxation. We show the efficacy of this formulation with a numerical example.",

keywords = "Optimization, Mathematical model, Finite impulse response filters, Minimization, Maximum likelihood estimation, MIMO, Numerical models",

author = "Dexter Scobee and Lillian Ratliff and Roy Dong and Henrik Ohlsson and Michel Verhaegen and Sastry, {S. Shankar}",

year = "2015",

doi = "10.1109/CDC.2015.7402521",

language = "English",

pages = "2127--2132",

editor = "Yoshito Ohta and Mitsuji Sampei and Shigemasa Takai",

booktitle = "Proceedings of the 54th IEEE Conference on Decision and Control (CDC 2015)",

publisher = "IEEE",

address = "United States",

note = "54th IEEE Conference on Decision and Control, CDC 2015, CDC 2015 ; Conference date: 15-12-2015 Through 18-12-2015",

}

Scobee, D, Ratliff, L, Dong, R, Ohlsson, H, Verhaegen, M & Sastry, SS 2015, Nuclear norm minimization for blind subspace identification (N2BSID). in Y Ohta, M Sampei & S Takai (eds), Proceedings of the 54th IEEE Conference on Decision and Control (CDC 2015). IEEE, Piscataway, NJ, USA, pp. 2127-2132, 54th IEEE Conference on Decision and Control, CDC 2015, Osaka, Japan, 15/12/15. https://doi.org/10.1109/CDC.2015.7402521

Nuclear norm minimization for blind subspace identification (N2BSID). / Scobee, Dexter; Ratliff, Lillian; Dong, Roy et al.
Proceedings of the 54th IEEE Conference on Decision and Control (CDC 2015). ed. / Yoshito Ohta; Mitsuji Sampei; Shigemasa Takai. Piscataway, NJ, USA: IEEE, 2015. p. 2127-2132.

Research output: Chapter in Book/Conference proceedings/Edited volume › Conference contribution › Scientific › peer-review

TY - GEN

T1 - Nuclear norm minimization for blind subspace identification (N2BSID)

AU - Scobee, Dexter

AU - Ratliff, Lillian

AU - Dong, Roy

AU - Ohlsson, Henrik

AU - Verhaegen, Michel

AU - Sastry, S. Shankar

N1 - Conference code: 54

PY - 2015

Y1 - 2015

N2 - In many practical applications of system identification, it is not feasible to measure both the inputs applied to the system as well as the output. In such situations, it is desirable to estimate both the inputs and the dynamics of the system simultaneously; this is known as the blind identification problem. In this paper, we provide a novel extension of subspace methods to the blind identification of multiple-input multiple-output linear systems. We assume that our inputs lie in a known subspace, and we are able to formulate the identification problem as rank constrained optimization, which admits a convex relaxation. We show the efficacy of this formulation with a numerical example.

AB - In many practical applications of system identification, it is not feasible to measure both the inputs applied to the system as well as the output. In such situations, it is desirable to estimate both the inputs and the dynamics of the system simultaneously; this is known as the blind identification problem. In this paper, we provide a novel extension of subspace methods to the blind identification of multiple-input multiple-output linear systems. We assume that our inputs lie in a known subspace, and we are able to formulate the identification problem as rank constrained optimization, which admits a convex relaxation. We show the efficacy of this formulation with a numerical example.

KW - Optimization

KW - Mathematical model

KW - Finite impulse response filters

KW - Minimization

KW - Maximum likelihood estimation

KW - MIMO

KW - Numerical models

UR - http://www.scopus.com/inward/record.url?scp=84961999215&partnerID=8YFLogxK

U2 - 10.1109/CDC.2015.7402521

DO - 10.1109/CDC.2015.7402521

M3 - Conference contribution

AN - SCOPUS:84961999215

SP - 2127

EP - 2132

BT - Proceedings of the 54th IEEE Conference on Decision and Control (CDC 2015)

A2 - Ohta, Yoshito

A2 - Sampei, Mitsuji

A2 - Takai, Shigemasa

PB - IEEE

CY - Piscataway, NJ, USA

T2 - 54th IEEE Conference on Decision and Control, CDC 2015

Y2 - 15 December 2015 through 18 December 2015

ER -

Nuclear norm minimization for blind subspace identification (N2BSID)

Abstract

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this