Data Strategy for Autonomous Systems - Network Optional Warfare
Data Strategy for Autonomous Systems Field Experimentation (FX), Simulation, and Analysis
Abstract. Data collection and analysis techniques for robot experiments are haphazard and incomplete. Building best-practice workflows for data and metadata from unmanned systems can leverage both field experimentation (FX) and simulation to support archival data re-use and repeatable analysis. Reusable end-to-end data workflows are needed. Building on multiple open standards, open-source tools, and authoritative data formats, ongoing NPS CRUSER work is focused on applying Data Format Description Language (DFDL) techniques to archival recording/playback of mission orders, recorded telemetry and sensor streams. This necessary capability may enable an even-larger context, namely a Data Strategy for Unmanned Systems field experimentation (FX), modeling and simulation (M&S) supporting Live-Virtual-Constructive (LVC) synthesis, data repositories, and repeatable analysis. A full end-to-end toolchain built using open capabilities has the potential to address these important needs. This memorandum describes component technologies that together might establish such an information infrastructure. Failure to implement a shared data strategy blocks necessary interoperability of human-machine teams.
- Data Strategy for Autonomous Systems Field Experimentation (FX), Simulation, and Analysis, Technical Memorandum, Naval Postgraduate School (NPS), Don Brutzman, Curt Blais, and Terry Norbraten, updated 25 January 2023. (quad chart, technical report, and slideset)
Related work:
- Ethical Control of Autonomous Systems: Repeatable Mission Evaluation through Unmanned Systems Data Strategy, CRUSER video presentation, Naval Postgraduate School (NPS), Curt Blais, Kristen Fletcher, Terry Norbraten and Don Brutzman, 10 December 2021. This work proposes robot test and evaluation using both virtual and actual unit testing. A consistent Data Strategy for recording and comparing robot performance is essential for such acceptance testing to be meaningful. (website)