Graph-based sensor fusion for classification of transient acoustic signals

Umamahesh Srinivas, Nasser M. Nasrabadi, Vishal Monga

    Research output: Contribution to journalArticle

    2 Citations (Scopus)

    Abstract

    Advances in acoustic sensing have enabled the simultaneous acquisition of multiple measurements of the same physical event via co-located acoustic sensors. We exploit the inherent correlation among such multiple measurements for acoustic signal classification, to identify the launch/impact of munition (i.e., rockets, mortars). Specifically, we propose a probabilistic graphical model framework that can explicitly learn the class conditional correlations between the cepstral features extracted from these different measurements. Additionally, we employ symbolic dynamic filtering-based features, which offer improvements over the traditional cepstral features in terms of robustness to signal distortions. Experiments on real acoustic data sets show that our proposed algorithm outperforms conventional classifiers as well as the recently proposed joint sparsity models for multisensor acoustic classification. Additionally our proposed algorithm is less sensitive to insufficiency in training samples compared to competing approaches.

    Original languageEnglish (US)
    Article number6846349
    Pages (from-to)576-587
    Number of pages12
    JournalIEEE Transactions on Cybernetics
    Volume45
    Issue number3
    DOIs
    StatePublished - Mar 1 2015

    Fingerprint

    Fusion reactions
    Acoustics
    Sensors
    Signal distortion
    Rockets
    Mortar
    Classifiers
    Experiments

    All Science Journal Classification (ASJC) codes

    • Software
    • Control and Systems Engineering
    • Information Systems
    • Human-Computer Interaction
    • Computer Science Applications
    • Electrical and Electronic Engineering

    Cite this

    Srinivas, Umamahesh ; Nasrabadi, Nasser M. ; Monga, Vishal. / Graph-based sensor fusion for classification of transient acoustic signals. In: IEEE Transactions on Cybernetics. 2015 ; Vol. 45, No. 3. pp. 576-587.
    @article{7c99b9f847b0434ea80d2e6adefdc096,
    title = "Graph-based sensor fusion for classification of transient acoustic signals",
    abstract = "Advances in acoustic sensing have enabled the simultaneous acquisition of multiple measurements of the same physical event via co-located acoustic sensors. We exploit the inherent correlation among such multiple measurements for acoustic signal classification, to identify the launch/impact of munition (i.e., rockets, mortars). Specifically, we propose a probabilistic graphical model framework that can explicitly learn the class conditional correlations between the cepstral features extracted from these different measurements. Additionally, we employ symbolic dynamic filtering-based features, which offer improvements over the traditional cepstral features in terms of robustness to signal distortions. Experiments on real acoustic data sets show that our proposed algorithm outperforms conventional classifiers as well as the recently proposed joint sparsity models for multisensor acoustic classification. Additionally our proposed algorithm is less sensitive to insufficiency in training samples compared to competing approaches.",
    author = "Umamahesh Srinivas and Nasrabadi, {Nasser M.} and Vishal Monga",
    year = "2015",
    month = "3",
    day = "1",
    doi = "10.1109/TCYB.2014.2331284",
    language = "English (US)",
    volume = "45",
    pages = "576--587",
    journal = "IEEE Transactions on Cybernetics",
    issn = "2168-2267",
    publisher = "IEEE Advancing Technology for Humanity",
    number = "3",

    }

    Graph-based sensor fusion for classification of transient acoustic signals. / Srinivas, Umamahesh; Nasrabadi, Nasser M.; Monga, Vishal.

    In: IEEE Transactions on Cybernetics, Vol. 45, No. 3, 6846349, 01.03.2015, p. 576-587.

    Research output: Contribution to journalArticle

    TY - JOUR

    T1 - Graph-based sensor fusion for classification of transient acoustic signals

    AU - Srinivas, Umamahesh

    AU - Nasrabadi, Nasser M.

    AU - Monga, Vishal

    PY - 2015/3/1

    Y1 - 2015/3/1

    N2 - Advances in acoustic sensing have enabled the simultaneous acquisition of multiple measurements of the same physical event via co-located acoustic sensors. We exploit the inherent correlation among such multiple measurements for acoustic signal classification, to identify the launch/impact of munition (i.e., rockets, mortars). Specifically, we propose a probabilistic graphical model framework that can explicitly learn the class conditional correlations between the cepstral features extracted from these different measurements. Additionally, we employ symbolic dynamic filtering-based features, which offer improvements over the traditional cepstral features in terms of robustness to signal distortions. Experiments on real acoustic data sets show that our proposed algorithm outperforms conventional classifiers as well as the recently proposed joint sparsity models for multisensor acoustic classification. Additionally our proposed algorithm is less sensitive to insufficiency in training samples compared to competing approaches.

    AB - Advances in acoustic sensing have enabled the simultaneous acquisition of multiple measurements of the same physical event via co-located acoustic sensors. We exploit the inherent correlation among such multiple measurements for acoustic signal classification, to identify the launch/impact of munition (i.e., rockets, mortars). Specifically, we propose a probabilistic graphical model framework that can explicitly learn the class conditional correlations between the cepstral features extracted from these different measurements. Additionally, we employ symbolic dynamic filtering-based features, which offer improvements over the traditional cepstral features in terms of robustness to signal distortions. Experiments on real acoustic data sets show that our proposed algorithm outperforms conventional classifiers as well as the recently proposed joint sparsity models for multisensor acoustic classification. Additionally our proposed algorithm is less sensitive to insufficiency in training samples compared to competing approaches.

    UR - http://www.scopus.com/inward/record.url?scp=85027936433&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=85027936433&partnerID=8YFLogxK

    U2 - 10.1109/TCYB.2014.2331284

    DO - 10.1109/TCYB.2014.2331284

    M3 - Article

    AN - SCOPUS:85027936433

    VL - 45

    SP - 576

    EP - 587

    JO - IEEE Transactions on Cybernetics

    JF - IEEE Transactions on Cybernetics

    SN - 2168-2267

    IS - 3

    M1 - 6846349

    ER -