Timely classification of aircraft is important for small unmanned aerial system (sUAS) technologies, such as onboard collision avoidance systems, and aerial perimeter security for prisons and sports venues. This work uses velocity-based metrics to classify multi-rotor sUAS, fixed wings UAS, and general aviation planes using two classification methods: Support Vector Machines (SVM), and Likelihood Ratio (LR) tests. We found that a 96% classification accuracy is achieved when either classifier is trained using average speed derived from flight controller data or radar data and tested with one second of radar data. Further, we show that LR tests perform similarly to SVM for single metric classification. In addition, we present two novel metrics for classifying aircraft: log variance of absolute change in speed, and log variance of relative change in speed. Finally, we discuss challenges associated with training classifiers with flight controller data but testing on radar data.


    Zugriff

    Zugriff über TIB

    Verfügbarkeit in meiner Bibliothek prüfen


    Exportieren, teilen und zitieren



    Titel :

    Classifying Aircraft using Velocity Data with Support Vector Machines and Likelihood Ratio Tests


    Beteiligte:
    L. T. Dihel (Autor:in) / C. V. Dolph (Autor:in) / H. T. Holbrook (Autor:in) / S. Roy (Autor:in)

    Erscheinungsdatum :

    2022


    Format / Umfang :

    14 pages


    Medientyp :

    Report


    Format :

    Keine Angabe


    Sprache :

    Englisch





    Classifying Aircraft Using Velocity Data with Support Vector Machines and Likelihood Ratio Tests

    Dihel, Logan / Dolph, Chester / Holbrook, Henry T. et al. | TIBKAT | 2023


    Classifying Aircraft using Velocity Data with Support Vector Machines and Likelihood Ratio Tests

    Dihel, Logan / Dolph, Chester / Holbrook, Henry T. et al. | AIAA | 2023



    Classifying Acoustic Emission Data in Structural Health Monitoring Using Support Vector Machines

    Esterline, Albert / Krishnamurthy, Kiran / Sundaresan, Mannur et al. | AIAA | 2010