Robotics & Physical AI

How Perseverance finds its own way across Mars

NASA’s rover used AutoNav to assess terrain over most of its record-breaking journey. A separate experiment tested where generative AI could help the team on Earth.

Riley · FoxAI World1 min read
18
Concept illustration of the Perseverance rover crossing rocky Martian terrainDriving on Mars

Conceptual cover created with AI.

On Mars, a waypoint can tell Perseverance where to go without telling it how to get around every boulder. That gap is where the rover’s onboard navigation earns its keep. NASA’s Jet Propulsion Laboratory reports that Perseverance reached 45.22 kilometres on October 11, 2026, passing Opportunity’s distance record. It used AutoNav to evaluate terrain on more than 90% of that journey.

Driving and knowing where you are

AutoNav uses images taken while the rover is moving to build a 3D hazard map and steer between waypoints set on Earth. After a long drive, small errors can add up in the rover’s estimate of its position. Mars Global Localization lets it check that estimate independently by matching a panorama from its navigation cameras against orbital imagery.

Illustration of a Mars rover route threading between rocks
AI-generated conceptual illustration.

Where Claude entered the picture

In two December 2025 demonstrations, JPL used Claude models on Earth to generate waypoints from orbital imagery and slope data. Engineers ran the commands through a digital twin before uplinking them. The drives covered 210 and 246 metres. For a mission team, the distinction is practical. Generative AI can help prepare the route on Earth, while onboard navigation handles the terrain the rover actually encounters.

Sources and research

Written by FoxAI World. The cover and illustrations are AI-generated conceptual artwork.

Your favourites

This list stays in this browser on your device. No account is needed, and other visitors cannot see it.

Use a story’s bookmark button to add your first favourite.