Learning-enabled control methods that adapt online while preserving stability, safety, and performance under uncertainty.
AIAS LabPurdue University Northwest
The research integrates rigorous nonlinear and adaptive control with online learning methods for uncertain systems. Emphasis is placed on adaptive optimal control, actor-critic reinforcement learning, neural function approximation, and safety mechanisms that preserve stability and constraint satisfaction during learning.
The work targets autonomous systems that must improve performance online while remaining reliable under modeling uncertainty, disturbances, and changing operating conditions.