NASA’s Perseverance Rover Completes First AI-Planned Drive on Mars - UNIVERSE

 

NASA’s Perseverance used its navigation cameras to capture its drive along the rim of Jezero Crater on Dec. 10, 2025. The navcam images were combined with rover data and placed into a 3D virtual environment, resulting in this reconstruction with virtual frames inserted about every 4 inches (0.1 meters) of drive progress.
NASA/JPL-Caltech

The team for the six-wheeled scientist used a vision-capable AI to create a safe route over the Red Planet’s surface without the input of human route planners. 

NASA’s Perseverance Mars rover has completed the first drives on another world that were planned by artificial intelligence. Executed on Dec. 8 and 10, and led by the agency’s Jet Propulsion Laboratory in Southern California, the demonstration used generative AI to create waypoints for Perseverance, a complex decision-making task typically performed manually by the mission’s human rover planners.  

“This demonstration shows how far our capabilities have advanced and broadens how we will explore other worlds,” said NASA Administrator Jared Isaacman. “Autonomous technologies like this can help missions to operate more efficiently, respond to challenging terrain, and increase science return as distance from Earth grows. It’s a strong example of teams applying new technology carefully and responsibly in real operations.” 

During the demonstration, the team leveraged a type of generative AI called vision-language models to analyze existing data from JPL’s surface mission dataset. The AI used the same imagery and data that human planners rely on to generate waypoints — fixed locations where the rover takes up a new set of instructions — so that Perseverance could safely navigate the challenging Martian terrain. 

The initiative was led out of JPL’s Rover Operations Center (ROC) in collaboration with Anthropic, using the company’s Claude AI models.  

This animation was created using data acquired during Perseverance’s Dec. 10, 2025, drive on Jezero Crater’s rim. Pale blue lines depict the track the rover’s wheels take. Black lines snaking out in front of the rover show the path options the rover is considering. The white terrain is a height map based on rover data. The blue circle that appears near the end of the animation is a waypoint.
NASA/JPL-Caltech

Progress for Mars, beyond 

Mars is on average about 140 million miles (225 million kilometers) away from Earth. This vast distance creates a significant communication lag, making real-time remote operation — or “joy-sticking” — of a rover impossible. Instead, for the past 28 years, over several missions, rover routes have been planned and executed by human “drivers,” who analyze the terrain and status data to sketch a route using waypoints, which are usually spaced no more than 330 feet (100 meters) apart to avoid any potential hazards. Then they send the plans via NASA’s Deep Space Network to the rover, which executes them. 

But for Perseverance’s drives on the 1,707 and 1,709 Martian days, or sols, of the mission, the team did something different: Generative AI provided the analysis of the high-resolution orbital imagery from the HiRISE (High Resolution Imaging Science Experiment) camera aboard NASA’s Mars Reconnaissance Orbiter and terrain-slope data from digital elevation models. After identifying critical terrain features — bedrock, outcrops, hazardous boulder fields, sand ripples, and the like — it generated a continuous path complete with waypoints.

To ensure the AI’s instructions were fully compatible with the rover’s flight software, the engineering team also processed the drive commands through JPL’s “digital twin” (virtual replica of the rover), verifying over 500,000 telemetry variables before sending commands to Mars. 

On Dec. 8, with generative AI waypoints in its memory, Perseverance drove 689 feet (210 meters). Two days later, it drove 807 feet (246 meters). 

“The fundamental elements of generative AI are showing a lot of promise in streamlining the pillars of autonomous navigation for off-planet driving: perception (seeing the rocks and ripples), localization (knowing where we are), and planning and control (deciding and executing the safest path),” said Vandi Verma, a space roboticist at JPL and a member of the Perseverance engineering team. “We are moving towards a day where generative AI and other smart tools will help our surface rovers handle kilometer-scale drives while minimizing operator workload, and flag interesting surface features for our science team by scouring huge volumes of rover images.”

“Imagine intelligent systems not only on the ground at Earth, but also in edge applications in our rovers, helicopters, drones, and other surface elements trained with the collective wisdom of our NASA engineers, scientists, and astronauts,” said Matt Wallace, manager of JPL’s Exploration Systems Office. “That is the game-changing technology we need to establish the infrastructure and systems required for a permanent human presence on the Moon and take the U.S. to Mars and beyond.”   

This annotated orbital image depicts the AI-planned (depicted 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 showing that generative AI could be incorporated into rover route planning.

NASA/JPL-Caltech/UofA

More about Perseverance 

Managed for NASA by Caltech, JPL is home to the Rover Operations Center (ROC). It also manages operations of the Perseverance rover on behalf of the agency’s Science Mission Directorate as part of NASA’s Mars Exploration Program portfolio.

For more information on the ROC, visit: https://www.jpl.nasa.gov/roc  

Source: NASA’s Perseverance Rover Completes First AI-Planned Drive on Mars  - NASA