NASA’s Perseverance rover achieved a significant milestone in space exploration in December 2025, completing a 400-meter traverse across Mars guided by artificial intelligence. This marks the first time an AI has independently planned a route for a rover on another planet, representing a major step forward in the integration of AI into deep-space missions.
The AI, known as Claude and developed by Anthropic, was utilized by engineers at NASA’s Jet Propulsion Laboratory (JPL) to navigate the rover through a rocky area within Jezero Crater. This advancement is particularly important as communication delays between Earth and Mars – approximately 20 minutes each way – prevent real-time control of the rover, necessitating more autonomous navigation capabilities.
AI-Driven Route Planning on the Martian Surface
Between December 8 and 10, 2025, Perseverance followed a path entirely designed by Claude. Traditionally, route planning involves human operators meticulously analyzing satellite imagery and onboard camera data to chart a safe course, avoiding obstacles like slippery terrain or areas where the rover could become stuck. The use of AI significantly accelerated this process.
Prior to the drive, NASA provided Claude Code, an Anthropic programming agent, with comprehensive mission data, including the rover’s history, Martian terrain conditions, and safety parameters. The AI then constructed the route incrementally, connecting waypoints approximately 10 meters apart.
How Claude Planned the Traverse
Claude processed images from NASA’s HiRISE camera and digital elevation models to understand the rocky Martian landscape. It then translated the planned route into a rover markup language, a specific command language that Perseverance can execute. This capability could potentially streamline future missions and allow for more efficient exploration of other planets.
Rigorous Verification and Adjustments by NASA Engineers
Before implementation, the AI-generated route underwent extensive simulations at JPL, evaluating over 500,000 telemetry variables to ensure the rover’s safety and security. Engineers made only minor adjustments, as the surface imagery available to them was sometimes more detailed than the data Claude had access to during route design. Perseverance successfully completed the route without issue.
Impact on Mission Efficiency and AI Implementation
Utilizing Claude reduced route planning time by as much as 50%, freeing up operators to focus on scientific data collection and broader exploration planning. NASA views this as a significant step toward more productive and consistent missions.
NASA Administrator Jared Isaacman emphasized that the use of autonomous technologies like Claude can maximize exploration results even with limited human resources. This is particularly relevant as NASA faces budget constraints and workforce reductions amid increasing demands for space exploration.
The Evolution of AI in Space Exploration
Just under a year prior, Claude’s navigational abilities were limited. Its successful navigation of the challenging Martian terrain demonstrates rapid progress in AI capabilities. This success paves the way for continued collaboration between NASA and AI developers to optimize future planetary missions.
NASA considers autonomous AI systems crucial for exploring increasingly distant and challenging regions of the solar system. Claude’s achievement on Mars signifies that generative AI is no longer simply an experiment, but a practical component of the future of space exploration.
Baca selengkapnya di: www.beritasatu.com
