Physical World Intelligence

Intelligence for the stars, the sea, and everything remote.

Intelligent systems that explore, construct, transform, and restore remote physical environments, from oceans and orbit to harsh built infrastructure.

Explore mission systemsHuman-supervised autonomy

Research themes

S

Stars

Autonomy and supervision for orbital, aerial, and long-distance systems where human presence is limited.

O

Sea

Persistent sensing and resilient physical action for ocean infrastructure, underwater operations, and coastal risk.

R

Remote Earth

Systems that restore and maintain places people cannot safely reach or stay in for long.

Rocket launch from Astro Lab reference
03extreme environment domains
24/7persistent remote operation
10+research questions in autonomy and care

From remote exploration to physical restoration.

Modeled after Astro Lab's stacked dark sections, these cards map the Stars and Sea research direction into concrete system programs.

Ocean intelligence

Long-duration autonomy for sea-scale environments

Robotic systems that can map, monitor, and maintain offshore and underwater infrastructure while keeping human operators in strategic control.

Orbital systems

Mission interfaces for delayed, distant action

Interfaces and decision layers that help teams supervise assets where latency, risk, and environmental uncertainty shape every action.

Resilient field work

Physical intelligence where people cannot safely remain

Sensing, manipulation, and recovery loops for harsh sites, disaster zones, and remote built environments.

Research Papers

Questions that guide the page.

01

How should autonomous systems communicate uncertainty when people are supervising high-risk physical work?

02

What interfaces let experts command systems at mission level without hiding operational detail?

03

How can robots recover gracefully when field environments break assumptions?

System layers for harsh physical worlds.

*

Perception

Multimodal sensing for water, weather, terrain, and low-visibility physical environments.

*

Autonomy

Planning systems that adapt when communication is partial and the world changes faster than expected.

*

Supervision

Human-in-the-loop control surfaces for mission-level intent, override, and trust calibration.

*

Recovery

Fault-aware behavior for systems that need to continue safely after damage, drift, or uncertainty.

Act on the physical world with people still at the center.

Shuai Li © 2026 . All rights reserved.

Shuai Li © 2026 . All rights reserved.