Neuro-symbolic Robot Planning & Learning
Combining symbolic reasoning, formal task structures, and learning-based models to enable robots to understand high-level goals, reason over long-horizon tasks, and generate interpretable plans.
02 / RESEARCH DIRECTIONSRIGOR × INTELLIGENCE
From possibility
to provability.
Robots should do more than perform. They should understand the goal—and give us reason to trust the outcome.
Trustworthy Robot Learning & Embodied Intelligence
Enabling robots to learn interpretable physical skills that remain safe, robust, and reliable in real-world environments.
Trustworthy & Assured Autonomy
Developing foundations for specification, verification, certification, and correct-by-construction design for autonomous systems.
Task and Motion Planning
Integrating high-level task reasoning with geometric and motion planning to generate feasible, collision-free robot behaviors that satisfy complex task specifications and physical constraints.
Scalable Multi-robot Systems
Developing methods for task allocation, coordination, and planning under complex high-level objectives.
Real-World Applications
Grounding trustworthy autonomy in smart manufacturing, human-robot collaboration, mobile manipulation, autonomous inspection, warehouse automation, service robotics, and field robot teams.