SFT 2026-27 - Lunar Base Micrometeorite Inspection Robot(s)

NASA Reference Name: GLENN-P5-METEOR-2026

Executive Summary

Design, build, and program a robotic system that inspects a lunar base for micrometeorite impacts and structural damage. On a regular basis, and after each impact event, the robot(s) autonomously (or partially autonomously) survey the habitat, landing pad, and surrounding surfaces to detect and analyze damage — without hitting people or vehicles. Objective: inspect a 1-km-radius base (including a 500 m landing pad located 1 km from center) and survey all buildings. Deliverables span a software/AI & VR track (VR/GUI control interface, an AI computer-vision pipeline with code and docs, a digital habitat model, a longitudinal report, and a presentation) and a physical track (a physical rover, a habitat model with impact holes, and sensor integration).

Requested By

NASA Glenn Research Center (GRC) - NASA HUNCH. Ideator: Nancy R. Hall (GRC-MSI0), GRC HUNCH & XMIPT Project Manager. Author: Michael Hayes - NASA HUNCH.

Problem Statement

With no atmosphere to burn up debris, micrometeorites strike lunar habitats at ~20 km/s — risking cracks, slow leaks, loss of pressure, and astronaut safety. On a regular basis, and after each impact event, the robot(s) must autonomously inspect the lunar base to identify and analyze damage to structures, the landing pad, and surrounding surfaces.

Requirements Overview

Inspect >=110% of the base surface, including habitat and landing pad (SYS-001)

Autonomous plus remote-controlled operation (SYS-002)

Real / near-real-time data, latency <2 s (SYS-003, HW-007)

Modular sensors: RGB >=1080p, IR, LiDAR (HW-002)

Detect pits >=0.1 mm diameter; measure depth to 0.05 mm (HW-003)

Traverse flat, curved, and vertical surfaces to 90 degrees (ME-003/005/006)

AI detects and classifies impacts vs. a baseline (SW-003)

Assess severity and decompression risk (SW-005); VR/GUI control (SW-001); battery >=120 min (HW-005)

Major Constraints

Lunar environment: ~1/6 g with simulated lunar lighting (deep shadows and dark areas), with large temperature variation across a day and when crossing shadows; regolith covers most areas and some building surfaces (for radiation control)

Curved and vertical habitat surfaces demand reliable adhesion

Must detect very small impacts — pits as small as 0.1 mm in diameter

Telemetry and control over a 2-km inspection area with <2 s latency

Key Challenges

Telling a brand-new impact apart from normal surface features

Keeping adhesion while climbing curved and vertical surfaces (sides and roof)

Measuring hole length, width, and depth accurately in 3D

Estimating decompression risk from detected damage

Reliable low-latency streaming plus rover localization across a 2-km area

Cards

1 — Other Points / Comments

Combined project: Design & Prototype plus Software/AI & VR

Two operational modes: Safety Mode (real-time alarms) and Research Mode (longitudinal study)

The source requirements doc had conflicting battery figures; the 120-min value (HW-005) is used as authoritative

Motivated by a recent Artemis II observation of micrometeorite flashes on the lunar surface

Aligns with NASA Artemis and Moon-to-Mars habitat-safety goals

2 — Examples of Excellence

The rover autonomously traverses the full habitat exterior (sides and roof) and finds every simulated impact hole

The AI correctly flags new impacts vs. the baseline with high confidence

The system measures hole size and depth and estimates decompression risk

Grounded in real NASA/ESA work: Perseverance (HazCams/Navcams), VIPER (10 cm feature detection), Astrobee, and ESA METERON teleoperation

3 — Examples of Innovation

VR teleoperation with optional haptic feedback (ESA METERON-inspired)

Deep-learning semantic segmentation for anomaly detection (NASA MMOD analogues)

Longitudinal / time-series surface-degradation analysis (Research Mode)

Automatic severity ranking and prioritized impact reports

Digital-twin habitat for simulation before physical testing

Suggestions for High School Students

The software/AI track may design a virtual 3D habitat instead of a physical one (HAB-005)

A physical habitat model needs at least 4 holes of different sizes — 1, 5, 10, and 20 mm (HAB-002) for testing

An Earth-scale classroom/lab demo is acceptable; lunar gravity can be modeled conceptually

The AI may be rule / threshold-based for beginner teams; use ML for advanced teams

Robots can do the survey, while a local processor handles visualization and detection

Please review more info below:

Requirements Spec

Other Points / Comments

Examples of Excellence

Samples of Innovation

User Stories