Goal-Driven Autonomy for Trusted Autonomous Systems
Abstract
Autonomous agents, both hardware and software, have narrow bands of capability that fail to perform in complex situations that include changing environmental conditions and uncertainty. We will focus this research on autonomous agents and platforms capable of deploying in such situations for extended periods with limited communications. Our technical approach consists of a two-fold process: (1) to recognize problems and opportunities when the world diverges from what is normal; (2) to distinguish between competing hypotheses that explain these changes of circumstance. Results are expected to improve the robustness of trusted agents under a wider set of circumstances than is currently possible.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Sep 11, 2017
- Source ID
- FA23861714063
Entities
People
- Michael M. Cox
Organizations
- Air Force Office of Scientific Research
- United States Air Force