Rank constrained ML estimation of structured covariance matrices with applications in radar target detection

Vishal Monga, Muralidhar Rangaswamy

    Research output: Chapter in Book/Report/Conference proceedingConference contribution

    8 Citations (Scopus)

    Abstract

    We consider here the continually important problem of radar target detection in the presence of clutter, noise and jamming. Under complex Gaussian noise statistics, the optimal detection statistic relies on an inversion of the disturbance (clutter + noise and jamming) covariance matrix. The disturbance (and hence clutter) covariance must be estimated in practice from sample, i.e. training observations. Traditional maximum likelihood (ML) estimators are effective when training is abundant but lead to poor estimates and hence high detection error in the realistic regime of limited or small training. The problem is exacerbated by recent advances which have led to high dimensionality N of the observations arising from increased antenna elements (J) as well as higher temporal resolution (P time epochs and finally N = J.P). This work introduces physically inspired constraints into ML estimation. In particular, we exploit both the structure of the disturbance covariance and importantly the knowledge of the clutter rank to yield a new rank constrained maximum likelihood (RCML) estimator of clutter/disturbance covariance. Experimental validation on the KASSPER data set (where ground truth covariance is made available) shows that the proposed estimator vastly outperforms state-of-the art alternatives in the sense of: 1.) higher normalized signal to interference and noise ratio (SINR), and 2.) lower variance of target amplitude estimators that utilize disturbance covariance. Crucially the proposed RCML estimator can excel even for low training including the notoriously difficult regime of K ≤ N training samples.

    Original languageEnglish (US)
    Title of host publication2012 IEEE Radar Conference
    Subtitle of host publicationUbiquitous Radar, RADARCON 2012 - Conference Program
    Pages475-480
    Number of pages6
    DOIs
    StatePublished - Jul 30 2012
    Event2012 IEEE Radar Conference: Ubiquitous Radar, RADARCON 2012 - Atlanta, GA, United States
    Duration: May 7 2012May 11 2012

    Publication series

    NameIEEE National Radar Conference - Proceedings
    ISSN (Print)1097-5659

    Other

    Other2012 IEEE Radar Conference: Ubiquitous Radar, RADARCON 2012
    CountryUnited States
    CityAtlanta, GA
    Period5/7/125/11/12

    Fingerprint

    Maximum likelihood estimation
    Covariance matrix
    Target tracking
    Maximum likelihood
    Radar
    Jamming
    Statistics
    Radar clutter
    Gaussian noise (electronic)
    Error detection
    Antennas

    All Science Journal Classification (ASJC) codes

    • Electrical and Electronic Engineering

    Cite this

    Monga, V., & Rangaswamy, M. (2012). Rank constrained ML estimation of structured covariance matrices with applications in radar target detection. In 2012 IEEE Radar Conference: Ubiquitous Radar, RADARCON 2012 - Conference Program (pp. 475-480). [6212188] (IEEE National Radar Conference - Proceedings). https://doi.org/10.1109/RADAR.2012.6212188
    Monga, Vishal ; Rangaswamy, Muralidhar. / Rank constrained ML estimation of structured covariance matrices with applications in radar target detection. 2012 IEEE Radar Conference: Ubiquitous Radar, RADARCON 2012 - Conference Program. 2012. pp. 475-480 (IEEE National Radar Conference - Proceedings).
    @inproceedings{d249642ae792476a8f965e6676f8e85a,
    title = "Rank constrained ML estimation of structured covariance matrices with applications in radar target detection",
    abstract = "We consider here the continually important problem of radar target detection in the presence of clutter, noise and jamming. Under complex Gaussian noise statistics, the optimal detection statistic relies on an inversion of the disturbance (clutter + noise and jamming) covariance matrix. The disturbance (and hence clutter) covariance must be estimated in practice from sample, i.e. training observations. Traditional maximum likelihood (ML) estimators are effective when training is abundant but lead to poor estimates and hence high detection error in the realistic regime of limited or small training. The problem is exacerbated by recent advances which have led to high dimensionality N of the observations arising from increased antenna elements (J) as well as higher temporal resolution (P time epochs and finally N = J.P). This work introduces physically inspired constraints into ML estimation. In particular, we exploit both the structure of the disturbance covariance and importantly the knowledge of the clutter rank to yield a new rank constrained maximum likelihood (RCML) estimator of clutter/disturbance covariance. Experimental validation on the KASSPER data set (where ground truth covariance is made available) shows that the proposed estimator vastly outperforms state-of-the art alternatives in the sense of: 1.) higher normalized signal to interference and noise ratio (SINR), and 2.) lower variance of target amplitude estimators that utilize disturbance covariance. Crucially the proposed RCML estimator can excel even for low training including the notoriously difficult regime of K ≤ N training samples.",
    author = "Vishal Monga and Muralidhar Rangaswamy",
    year = "2012",
    month = "7",
    day = "30",
    doi = "10.1109/RADAR.2012.6212188",
    language = "English (US)",
    isbn = "9781467306584",
    series = "IEEE National Radar Conference - Proceedings",
    pages = "475--480",
    booktitle = "2012 IEEE Radar Conference",

    }

    Monga, V & Rangaswamy, M 2012, Rank constrained ML estimation of structured covariance matrices with applications in radar target detection. in 2012 IEEE Radar Conference: Ubiquitous Radar, RADARCON 2012 - Conference Program., 6212188, IEEE National Radar Conference - Proceedings, pp. 475-480, 2012 IEEE Radar Conference: Ubiquitous Radar, RADARCON 2012, Atlanta, GA, United States, 5/7/12. https://doi.org/10.1109/RADAR.2012.6212188

    Rank constrained ML estimation of structured covariance matrices with applications in radar target detection. / Monga, Vishal; Rangaswamy, Muralidhar.

    2012 IEEE Radar Conference: Ubiquitous Radar, RADARCON 2012 - Conference Program. 2012. p. 475-480 6212188 (IEEE National Radar Conference - Proceedings).

    Research output: Chapter in Book/Report/Conference proceedingConference contribution

    TY - GEN

    T1 - Rank constrained ML estimation of structured covariance matrices with applications in radar target detection

    AU - Monga, Vishal

    AU - Rangaswamy, Muralidhar

    PY - 2012/7/30

    Y1 - 2012/7/30

    N2 - We consider here the continually important problem of radar target detection in the presence of clutter, noise and jamming. Under complex Gaussian noise statistics, the optimal detection statistic relies on an inversion of the disturbance (clutter + noise and jamming) covariance matrix. The disturbance (and hence clutter) covariance must be estimated in practice from sample, i.e. training observations. Traditional maximum likelihood (ML) estimators are effective when training is abundant but lead to poor estimates and hence high detection error in the realistic regime of limited or small training. The problem is exacerbated by recent advances which have led to high dimensionality N of the observations arising from increased antenna elements (J) as well as higher temporal resolution (P time epochs and finally N = J.P). This work introduces physically inspired constraints into ML estimation. In particular, we exploit both the structure of the disturbance covariance and importantly the knowledge of the clutter rank to yield a new rank constrained maximum likelihood (RCML) estimator of clutter/disturbance covariance. Experimental validation on the KASSPER data set (where ground truth covariance is made available) shows that the proposed estimator vastly outperforms state-of-the art alternatives in the sense of: 1.) higher normalized signal to interference and noise ratio (SINR), and 2.) lower variance of target amplitude estimators that utilize disturbance covariance. Crucially the proposed RCML estimator can excel even for low training including the notoriously difficult regime of K ≤ N training samples.

    AB - We consider here the continually important problem of radar target detection in the presence of clutter, noise and jamming. Under complex Gaussian noise statistics, the optimal detection statistic relies on an inversion of the disturbance (clutter + noise and jamming) covariance matrix. The disturbance (and hence clutter) covariance must be estimated in practice from sample, i.e. training observations. Traditional maximum likelihood (ML) estimators are effective when training is abundant but lead to poor estimates and hence high detection error in the realistic regime of limited or small training. The problem is exacerbated by recent advances which have led to high dimensionality N of the observations arising from increased antenna elements (J) as well as higher temporal resolution (P time epochs and finally N = J.P). This work introduces physically inspired constraints into ML estimation. In particular, we exploit both the structure of the disturbance covariance and importantly the knowledge of the clutter rank to yield a new rank constrained maximum likelihood (RCML) estimator of clutter/disturbance covariance. Experimental validation on the KASSPER data set (where ground truth covariance is made available) shows that the proposed estimator vastly outperforms state-of-the art alternatives in the sense of: 1.) higher normalized signal to interference and noise ratio (SINR), and 2.) lower variance of target amplitude estimators that utilize disturbance covariance. Crucially the proposed RCML estimator can excel even for low training including the notoriously difficult regime of K ≤ N training samples.

    UR - http://www.scopus.com/inward/record.url?scp=84864208539&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=84864208539&partnerID=8YFLogxK

    U2 - 10.1109/RADAR.2012.6212188

    DO - 10.1109/RADAR.2012.6212188

    M3 - Conference contribution

    AN - SCOPUS:84864208539

    SN - 9781467306584

    T3 - IEEE National Radar Conference - Proceedings

    SP - 475

    EP - 480

    BT - 2012 IEEE Radar Conference

    ER -

    Monga V, Rangaswamy M. Rank constrained ML estimation of structured covariance matrices with applications in radar target detection. In 2012 IEEE Radar Conference: Ubiquitous Radar, RADARCON 2012 - Conference Program. 2012. p. 475-480. 6212188. (IEEE National Radar Conference - Proceedings). https://doi.org/10.1109/RADAR.2012.6212188