Physics-Enhanced Deep Learning for Fast-than-Real-Time Prediction of Dynamic Behavior of Space Robotic Systems

Abstract

The U.S. Space Force (USSF) desires safe AI conduct and space operation battlespace awareness principles such that technology determines a viable and ethical solution to challenging scenarios. This research will build a solid foundation for further development of emerging AI technologies (e.g., PEDL) to advance the DoD capabilities for reliable ISAM missions. This project will develop a Physics-Enhanced Deep Learning (PEDL) methodology for faster-than-real-time (near-instant) prediction of the dynamic behavior of a complex spacecraft-robot system in response to operation-control actions for challenging In-space Servicing, Assembly, and Manufacturing (ISAM) operations. The operation-control actions can be teleoperation commands from a human operator or autonomous commands from an intrinsic safety response or extrinsic automated decision maker. The motivation is for the near-instant prediction to enhance operational situation awareness to mitigate effects of non-safe situations in highly-cluttered environments and human teleoperated spacecraft-robot systems in the presence of long latency.

Document Details

Document Type
DoD Grant Award
Publication Date
Mar 06, 2024
Source ID
FA95502310440

Entities

People

  • Ou Ma

Organizations

  • Air Force Office of Scientific Research
  • United States Air Force
  • University of Cincinnati

Tags

Readers

  • Aerospace Engineering.
  • Neural Network Machine Learning.
  • Team-Based Human-Centered Cognitive Task Decision Making and Information Performance.

Technology Areas

  • AI & ML
  • AI & ML - Autonomous Systems
  • AI & ML - DoD AI Strategy
  • Autonomy
  • Autonomy - Autonomous System Control
  • Autonomy - UAVs
  • Space
  • Space - Spacecraft Maneuvers