We're proud to officially announce that Tangram Vision has been awarded a NASA Phase II SBIR to solve one of autonomous robotics' biggest challenges: keeping sensor systems calibrated without human intervention. Our work will advance autonomous lunar navigation for NASA's Artemis program and establish the foundation for a long-term autonomous lunar presence.

Design Partners Wanted: We’re looking for corporate design partners for this project. Read more below, then email us at info@tangramvision.com with your own use case and needs.

The Challenge

NASA's Artemis program aims to establish a permanent lunar base as a stepping stone to Mars. But this won't just be a human outpost—autonomous systems will handle surveying, discovery, and maintenance operations, often working alone for extended periods.

Consider the sensor suites that sophisticated autonomy companies deploy on Earth—complex camera arrays, LiDAR, and IMUs that require constant calibration and maintenance. Even with dedicated field engineers on call, keeping these systems perfectly aligned is challenging. Now imagine that same task on the Moon, where harsh conditions constantly affect sensor performance and no help is available.

The Unforgiving Lunar Surface

The lunar environment presents unique sensing challenges:

• Radiation bombardment: Without atmosphere or magnetic field protection, space radiation gradually degrades sensor components.

• Lunar dust: Fine particles enhance reflectance, increase contrast, and statically adhere to sensor surfaces, changing their optical properties over time.

• Extreme lighting: Minimal axial tilt creates persistent shadows alongside areas of intense, continuous sunlight — imagine trying to navigate from bright stadium lighting directly into a cave!

Shadows on the lunar surface. Credit: NASA's Scientific Visualization Studio

This creates a perfect storm for autonomous systems:

1. Immediate challenge: Traditional vision systems fail in extreme lunar lighting conditions.

2. Long-term challenge: The harsh environment gradually changes sensor behavior, making initial calibrations obsolete.

The solution requires both novel sensor use and the ability to recalibrate autonomously.

Our Approach

Our approach tackles both challenges simultaneously. We're developing an event-based stereo vision system combined with an IMU that breaks new ground in how robots perceive in harsh environments.

Benefit of Event Cameras

Event cameras offer three critical advantages for lunar operations:

• Extreme dynamic range (120dB vs 60dB for traditional cameras)—they can simultaneously see details in deep shadows and bright sunlight.

• Ultra-low power consumption—using just 10mW compared to 5-10W for traditional cameras, crucial for energy-limited space missions.

• High-speed performance—microsecond response times enable navigation during rapid lighting changes (aka moving fast across lunar terrain).

Our Phase I grant work was focused on proving out the use case of such a stereo system for lunar navigation. That work proved successful: when used in an odometry system, our stereo rig just might solve the issue of navigation in lunar lighting… assuming it remains calibrated.

AutoCal

This brings us to the real breakthrough in Phase II: AutoCal — our operator-free multi-modal calibration system.

During Phase I, we demonstrated calibration of event camera + IMU pairs using MetriCal, our production line calibration software. However, MetriCal relies on rigid targets in order to derive calibration figures and runs best in a controlled environment (think turntables or robot arms). MetriCal is best in its class on the factory floor, but autonomous vehicles on the Moon won’t be in a factory; they’ll be jet-setting along the lunar surface. They’ll need constant calibration verification and refinement during deployment to successfully operate.

That’s AutoCal: calibration that happens during normal operation, without targets, turntables, or human intervention. Starting with a MetriCal-verified calibration, AutoCal will monitor and tweak those calibration parameters in response to the changing operating conditions of the lunar surface. It uses environmental features and vehicle motion to continuously verify and adjust sensor alignment — essentially turning every drive into a calibration session.

Developing AutoCal is the main technical focus at Tangram Vision for the next 18 months (and maybe forever).

Become A Design Partner

The terrestrial opportunity is immediate. By solving calibration for the Moon's extreme conditions, we're creating technology that will transform Earth-based robotics. Think of AutoCal as "space-grade reliability" for any environment — from wind farms to mining operations to autonomous agriculture.

We're seeking design partners to help us develop AutoCal in any and all terrestrial environments. If your autonomous systems require continuous operation in interesting or challenging conditions, we want to work with you. Contact us at info@tangramvision.com with your company details and use case. Space is limited; in order to best focus our efforts, we plan to restrict the number of partners we work with for the foreseeable future.

—

We couldn’t be more proud to support NASA and the organization’s lunar exploration goals. This project represents everything Tangram Vision stands for: solving the hardest perception problems to make autonomous systems reliable everywhere they're needed. From the Moon to your backyard, we're building the future of autonomous perception.