Robust Adaptive Autonomy in Contested Environments
Abstract
The research undertaken in this YIP proposal seeks to develop theoretical underpinnings and practical algorithms for Robust Adaptive Autonomy in Contested Environments (RAACE) for mixed manned-unmanned aerial teams. The specific objectives of this YIP proposal are:Create a new computationally and spatiotemporally scalable class of BNP and BNP-inspired predictive models and algorithms for nonstationary stochastic processes with both smooth and discrete transitions. Create Nonstationary model based reinforcement learning (N-MBRL) algorithms for risk-averse adaptive decision making in contested and time-varying environments. Develop algorithms to enable autonomous agents to infer latent human sub-goals directly from human demonstrations and feedback in contested environments. Validate the developed algorithms through data-simulations, multi-agent simulations, and flight experimentation involving humans and machines working together.
Document Details
- Document Type
- Technical Report
- Publication Date
- Sep 28, 2021
- Accession Number
- AD1154209
Entities
People
- Girish Chowdhary
Organizations
- Oklahoma State University–Stillwater