PEARL

Research

PEARL explores physical AI and robotics for real-world autonomous systems.

Humanoid robot moving its arms while working at a table

Humanoid Robot AI

Combine perception, locomotion, manipulation, and whole-body coordination for autonomous humanoids.

Whole-body controlManipulation & planning
Animation connecting external knowledge to an LLM to generate answers with sources

Knowledge-Augmented Generative AI

Retrieve relevant knowledge and connect it to generative models to produce grounded answers.

Information retrievalGrounded generation
Diagram connecting context, an AI persona, and personalized dialogue

Persona AI

Study AI agents that maintain consistent personas and adapt their interactions to users and context.

Persona modelingPersonalized interaction

Advanced Air Vehicle AI

Use vision and sensors to perceive changing surroundings, plan flight paths, and make autonomous decisions.

PerceptionAutonomous flight
Street traffic with vehicles highlighted for autonomous driving perception

Self-Driving Car AI

Develop perception, localization, and planning for autonomous driving in real-world traffic.

Driving perceptionMotion planning

Underwater Robot AI

Study localization, detection, and navigation where visibility and communication are limited.

Underwater localizationDetection & navigation

Surface Robot AI

Enable unmanned surface vehicles to perceive waterways, plan routes, and coordinate with other robots.

Unmanned vesselsAutonomous navigation

Defense AI

Study perception, decision-making, and robust autonomy for robots in uncertain field environments.

Field perceptionRobust autonomy
Diagram connecting Physical AI, large language models, and robotics

Other Research Area

All fields related to Physical AI, LLMs, and robots.

Physical AILLMsRobotics