Summaries - Office of Research & Innovation
Research Summaries
Back Cognitive Threat Emitter Recognition of Behaviors with Deep Neural Networks
Fiscal Year | 2018 |
Division | Research & Sponsored Programs |
Department | Naval Research Program |
Investigator(s) |
Schwamm, Noboru E.
Das, Arijit Rowe, Neil C. |
Sponsor | NPS Naval Research Program (Navy) |
Summary |
Our previous work has examined and classified radar signals. UxS signals are more varied because of the wide range of tasks they perform. But we do have clues from the size and time pattern of transmissions to and from known UxS platforms as well as their routes and behavior. These permit us to recognize new kinds of UxS platforms. A natural question is whether current research in cognitive pattern recognition can be used to see patterns in this data, of which there can be a considerable amount since UxS platforms can perform complex tasks. Various kinds of "deep" neural networks can be used to try to find latent structure in large amounts of data, and we will explore their application to this problem. Input to such networks can include not just the emissions and imaging of adversary UxS systems but their behavior in time and space. For instance, tracks can indicate whether a mission is reconnaissance, sentry activity, or part of an offensive mission. But this requires putting together a collection of data about it and doing more substantial reasoning than just classifying a signal based on its frequency and modulation characteristics. Aggregation of data from multiple observations will help improve accuracy of these classifications. This will be particularly important with decoy emissions and tracks, a concern because of the flexibility of programming of UxS platforms. We will analyze some simulated data to start, and then some real data. The simulated data will be created by using mission planner, a communications model, and a model of random effects in the physical world that require adjustments to the plans. The real data we will obtain from the sponsor. |
Keywords | |
Publications | Publications, theses (not shown) and data repositories will be added to the portal record when information is available in FAIRS and brought back to the portal |
Data | Publications, theses (not shown) and data repositories will be added to the portal record when information is available in FAIRS and brought back to the portal |