AI for Autonomous Systems

Autonomous systems—ranging from self-driving vehicles to intelligent laboratories—are redefining how machines perceive, learn, and act. This lecture examines how Artificial Intelligence, particularly reinforcement and transfer learning, enables adaptive autonomy in complex physical and cyber environments.

Dr. Kumar introduces the conceptual evolution of autonomy—from rule-based control to self-learning agents—and the algorithms that power decision-making under uncertainty. Drawing examples from autonomous vehicles, robotics, and distributed cyber-physical systems, the talk details how deep reinforcement learning, graph neural networks, and multimodal fusion are revolutionizing perception and planning. Case studies from his Department of Energy MEDAL project illustrate AI’s role in automating scientific experimentation through self-optimizing workflows.

The lecture also explores the cybersecurity and ethical dimensions of autonomy, including trust calibration, human–AI collaboration, and safety assurance. By integrating research insights from AI, and control theory, the talk provides a cohesive view of how autonomy will shape the next generation of intelligent infrastructure and scientific discovery.