NASA’s Perseverance rover completes first AI-planned drive on Mars

NASA's Perseverance rover has completed the first drives on another planet that were planned by artificial intelligence, a milestone achieved on December 8 and 10. The demonstration, led by NASA's Jet Propulsion Laboratory (JPL), showcased the use of generative AI to select safe waypoints for the rover, a task typically handled by human experts on Earth.

The AI system, developed in collaboration with Anthropic using the company's Claude AI models, analyzed high-resolution images and terrain data from NASA's Mars Reconnaissance Orbiter. This data included orbital images from the HiRISE camera and digital elevation models, which helped the AI identify surface features such as bedrock, boulder fields, and sand ripples. The AI then generated a continuous path with necessary waypoints for the rover to follow safely.

Before sending the commands to Mars, engineers tested the AI-generated plan using JPL's digital twin, a virtual replica of the rover. This process checked over 500,000 telemetry variables to ensure the plan would work with Perseverance's flight software. On December 8, the rover traveled 689 feet (210 meters) using the AI-generated plan, and two days later, it drove an additional 807 feet (246 meters).

NASA Administrator Jared Isaacman highlighted the significance of the achievement, stating that autonomous technologies like this could improve mission efficiency, help rovers navigate challenging terrain, and increase scientific returns as distances from Earth grow. Vandi Verma, a space roboticist at JPL, noted that generative AI is showing promise in streamlining autonomous navigation, including perception, localization, and path planning. The success of this demonstration could pave the way for future missions that rely on intelligent systems to handle large-scale drives and identify scientifically interesting features on planetary surfaces.