Autonomous Operations · Mining + Space

Autonomy for the most extreme environments

Lunarian builds the autonomy and intelligence that lets robots work where people can't — delivering zero-harm safety and uninterrupted productivity in the world's most hazardous terrain, from the post-blast mine heading to the lunar subsurface.

Explore the System → Request Briefing
Scroll
01 — MISSION

Out of
Harm's Way

Enabling safe, effective extraction and exploration — on, and off, Earth.

People should never have to enter the most dangerous places to keep them productive. Lunarian develops the autonomy that sends robots in instead — while operators stay in the control room: zero harm to personnel, no loss of production, in terrain that no human should have to set foot in.

/ 01

Zero Harm

Robots take the hazardous window — toxic gas, instability, the unknown — so no person has to. Safety stops being a trade-off against schedule.

/ 02

Uninterrupted Productivity

Autonomous re-entry and inspection recover the hours lost to manual clearance. The operation keeps moving while conditions are still being verified.

/ 03

Extreme Reach

One autonomy stack, many environments — from active blast zones and deep mine headings to lunar lava tubes. Hardware adapts; the intelligence carries over.

Autonomous quadruped operating in an underground cavern
UNIT TS-04
GAIT ADAPTIVE
LINK NONE · ONBOARD
▸ Domain-Specific

Built for the Mine

We're not building a universal robot brain. Our autonomy is engineered around the mine process itself — onboard decision-making and compute tuned to the specific tasks that matter underground, predictable and self-directed rather than over-built to think like a human. That same domain specificity is what carries the stack to lunar exploration and lava-tube mapping.

Post-blast underground heading, dust still settling
02 — THE SYSTEM

Progressive
Re-Entry

MVP SHIPPING · PHASE 01

Continuous Gas During the Vent Cycle

An autonomous quadruped enters the heading immediately after a blast and runs continuous gas readings through the entire ventilation cycle — recovering shift productivity and removing personnel from the hazardous window.

  • AContinuous O₂ / CO / CH₄ readings, live to surface
  • BGPS-denied navigation deep into the heading
  • CRecovered re-entry window — earlier clearance
PHASE 2 ROADMAP · PHASE 02

Decision-Ready Metrics, Same Window

The same pass delivers more. Pre-processed, in-shift metrics — not raw data — layered onto the gas profile through our data abstraction stack.

  • AFace mapping + fault & GSI metrics
  • BBolt audit + support verification
  • CMuck-pile scanning + volume estimates
ZERO
Teleoperation — tap start,
the robot runs itself
ON-SITE
Stored in-heading, no deploy
trip in or out of the mine
ANY
Platform — stack rides
your hardware spec
IN-SHIFT
Decision-ready data,
not raw point clouds
High-resolution LiDAR point cloud of an underground heading
SCAN 1.32M PTS
RES 4MM
MODE SLAM · LOCK
▸ Every Surface, Every Fracture

Captured as Data

Multi-modal sensing resolves the heading down to the millimeter — geometry, lithology, and atmosphere fused into one model the operator can act on before anyone sets foot inside.

Autonomous quadruped advancing through dust toward obscured miners
03 — EMERGING

Search
& Rescue

When a mine fails, the first hours decide outcomes — and the entry is exactly where no rescuer should be. Encouraged by mining-industry partners, we're extending Lunarian autonomy to send robots in first: to map the damage, locate personnel, and stream a live picture of the hazard back to the surface — turning a blind, deadly entry into an informed rescue.

/ 01

Rapid Damage Assessment

Survey collapse, water ingress, and structural failure in minutes — before a rescue team commits to the entry.

/ 02

Locate Personnel

Thermal and visual sensing search for trapped or unconscious workers through smoke, dust, and zero visibility.

/ 03

Live Hazard Picture

Gas, heat, and geometry streamed continuously to incident command — so every decision is made on current ground truth.

04 — TECHNOLOGY

An Autonomy Stack
Built for the Unknown

01

Multi-Modal Sensor Fusion

Paired RGB, thermal-infrared, and gas sensing fused in real time into one coherent model of the environment — seeing through dust, dark, and zero visibility.

02 PATENT PENDING

Edge-Deployed Vision-Language Models

Onboard VLA / VLM models reason about the scene with no link to the surface — validated against models such as LLaVA in GPS-denied terrain.

03

Multi-Height Environmental Characterization

Coordinated legged and aerial agents characterize the heading from floor to back, muck pile to roof — a full volumetric picture, not a single sweep.

04 PATENT PENDING

Heterogeneous Robot Autonomy

One mission, many platforms. Quadrupeds and UAVs operate as a coordinated team, so the stack rides whatever hardware the environment — or the customer — demands.

05 — FIELD PROOF

Mojave Desert
Lava Tubes

MOJAVE CAMPAIGN · LIVE SCAN
▸ Lava Tube Campaign · Lunar Analog

We deployed a legged quadruped and UAV system in the Mojave Desert lava tubes — a leading terrestrial analog for lunar subsurface exploration. The campaign captured paired RGB and thermal-infrared data across complex, featureless terrain and established baseline performance for our onboard vision models operating fully GPS-denied.

Platforms
Quadruped + UAV
Sensing
RGB · Thermal IR
Navigation
GPS-Denied
Analog
Lunar Subsurface
06 — TEAM

Operators &
Advisors

Shaheen Bharwani CEO · 01
Shaheen Bharwani
CEO

Go-to-market strategy, customer development, autonomous navigation, and vision models.

Middlebury College · SCI-Arc M.Arch
Frank Chen CTO · 02
Frank Chen
CTO

Robot platform integration, custom sensor hardware, ROS systems, and path planning.

SAIC BFA · SCI-Arc M.Arch
Dr. Keenan Albee ADVISOR · 03
Dr. Keenan Albee
Technical Advisor

Motion planning under uncertainty, multi-agent autonomy, space robotics, and GPS-denied navigation.

MIT PhD · NASA JPL · USC Viterbi ASTE
Dr. Feifei Qian ADVISOR · 04
Dr. Feifei Qian
Technical Advisor

Legged locomotion, robot-terrain interactions, bio-inspired mobility, and proprioceptive terrain sensing.

Georgia Tech PhD · NSF CAREER · USC Viterbi ECE
Karl Malitz ADVISOR · 05
Karl Malitz
Industry Advisor

Innovation, technical strategy, mine operations, and industry partnerships.

UNSW · 20 Years at Rio Tinto
Partner With Lunarian

Request a
Briefing

We work with mining operators, space agencies, and research partners deploying autonomy in the world's most demanding environments. Tell us about your terrain.

Contact Us →