Next-gen Edge AI & Robotics (NEAR) Lab

Building robust autonomy
for the physical world.

Advancing the algorithms and systems that enable robots to perceive, reason, and act reliably in real-world environments.

Three pillars of Robust Physical AI

Pillar 01

Perception

Multi-modal, multi-robot perception for consistent world understanding across heterogeneous agents — fusing vision, depth, inertial, and proprioceptive signals into edge-deployable spatial and semantic representations.

Pillar 02

Coordination & Action

Hierarchical decision-making and multi-robot coordination for translating high-level intent into robust physical behavior under partial observability and communication constraints.

Pillar 03

Edge Intelligence

Hardware-aware learning and inference for robots with strict power, thermal, and latency budgets — from model compression to AI–systems co-design for on-device autonomy.