Artificial Intelligence Robotics and Control Systems Laboratory (AIRoCS)

Develop methods that combine formal verification, control barrier functions, and learning algorithms to ensure safety certificates for robots (dual-arm manipulators, swarm, soft robots). This direction pushes toward trustworthy autonomy by embedding guarantees directly into learning-based controllers.

Explore swarm coordination and collective learning across heterogeneous robot fleets. Applications include precision agriculture, multi-arm manipulation, and remote testbeds (e.g., space repair). The emphasis is on compositional analysis, communication-constrained knowledge transfer, and scalability of distributed intelligence.

Investigate how tactile sensing, force feedback, and vision can be fused for robust perception in manipulation tasks. This research direction aims to achieve data-efficient skill acquisition, adaptive reflexes, and high-precision motor control, especially in unstructured or human-centered environments.

Use advanced mathematical tools—Grassmannian geometry, tensor-train decompositions, and physics-informed neural fields—to develop compact, generalizable skill representations. This allows robots to extrapolate beyond training data, handle singularities, and adapt across different embodiments while maintaining efficiency.

Design humanoid-like robotic avatars and immersive testbeds (e.g., 6G-enabled surgical robots, telepresence systems) that blend quantum control, mechatronics, and safe learning. The focus lies in achieving naturalistic interaction, remote embodiment, and reliable shared autonomy between human operators and robotic systems.