NASA’s Perseverance Rover Finishes its First Autonomous, AI-Directed Drive

Design Sem Nome 2026 02 05T133325.948
NASA’s Perseverance rover has been exploring Mars for almost five years, yet the agency continues to push its capabilities. Recently, NASA announced that Perseverance completed its first drive planned entirely by artificial intelligence.
The Perseverance Rover taking a selfie on a rock named Cheyava Falls. | Image Credits: NASA/JPL-Caltech/MSSS

NASA’s Perseverance rover has been exploring Mars for almost five years, yet the agency continues to push its capabilities. Recently, NASA announced that Perseverance completed its first drive planned entirely by artificial intelligence.

For the demo, NASA used vision-language models (VLMs) to set rover waypoints, a task normally done by humans. The demo ran from Dec. 8 to 10 and was led by the NASA Jet Propulsion Laboratory (JPL) in Southern California.

“This demonstration highlights how much our capabilities have progressed and expands how we can explore other worlds,” said NASA Administrator Jared Isaacman. “Autonomous technologies like this help missions operate more efficiently, navigate challenging terrain, and boost science returns as the distance from Earth increases. It’s an excellent example of teams applying new technology carefully and responsibly in real operations.”

The VLMs examined data from JPL’s surface mission dataset, using the same images and information that human planners depend on to create waypoints—specific locations where the rover receives new instructions.

The project was managed by JPL’s Rover Operations Center (ROC) in partnership with Anthropic, utilizing the company’s Claude AI models.

NASA Puts AI in the Driver’s Seat

Mars lies about 140 million miles (225 million km) from Earth on average, creating a significant communication delay that makes real-time remote control impossible.

For the past 28 years, human “drivers” have planned and executed mission routes. They analyze terrain and rover data, then map out paths using waypoints.

A key constraint is that waypoints must be no more than 330 ft (100 m) apart to prevent hazards. Once finalized, the drivers send the route via NASA’s Deep Space Network, and the rover executes it.

Now, with Perseverance, NASA is taking a new approach. Generative AI studied HiRISE orbital images and terrain slope data from digital elevation models.

The AI identified key terrain features—such as bedrock, outcrops, hazardous boulder fields, and sand ripples—and generated a continuous path complete with waypoints.

To ensure the AI’s instructions worked seamlessly with the rover’s flight software, engineers tested the drive commands using JPL’s digital twin, which verified over 500,000 telemetry variables before NASA transmitted the commands to Mars.

On Dec. 8, using generative AI–planned waypoints, Perseverance drove 689 ft (210 m), followed by 807 ft (246 m) two days later.

“The core elements of generative AI show great promise for streamlining off-planet autonomous navigation—perception (identifying rocks and ripples), localization (knowing our position), and planning and control (choosing and following the safest path),” said Vandi Verma, a JPL space roboticist on the Perseverance team.

“We’re moving toward a future where generative AI and other smart tools will enable surface rovers to cover kilometer-scale drives, reduce operator workload, and highlight interesting features for scientists by analyzing vast amounts of rover imagery,” she added.

What Lies Ahead for Perseverance?

Perseverancerover 768x573
This orbital image depicts the AI-planned (in magenta) and actual (orange) routes the Perseverance Mars rover took during its Dec. 10, 2025, drive at Jezero Crater. The drive was the second of two demonstrations incorporating generative AI into rover route planning. | Image Credits: NASA/JPL-Caltech

NASA hopes the technology tested with Perseverance can benefit a wide range of applications.

“Imagine intelligent systems not just on Earth, but also deployed on our rovers, helicopters, drones, and other surface assets, trained with the combined expertise of NASA engineers, scientists, and astronauts,” said Matt Wallace, manager of JPL’s Exploration Systems Office. “This technology is key to building the systems for a permanent Moon presence and U.S. missions to Mars and beyond.”

Since landing, Perseverance has been collecting rock samples for NASA’s Mars Sample Return campaign. However, the timeline for the campaign—originally set for a 2027 launch—is now uncertain.

In May 2025, the Trump administration proposed canceling the MSR program in NASA’s 2026 budget. Last month, Congress approved a budget that does not fund the MSR, effectively ending the mission.

The MSR campaign was a collaborative effort between NASA and the European Space Agency (ESA), and it remains uncertain how—or if—the ESA will continue the project without NASA’s involvement.


Read the original article on: The Robot Report

Read more: ISS Crew Safely Returns to Earth after a Medical Evacuation

Scroll to Top