What is it about?
Our work focuses on improving how routes are planned for lunar rovers, which is like figuring out the best path for a road trip on the Moon. Current methods for planning these routes are flawed because they use low-resolution elevation maps that lack detail and don’t account for uncertainty in terrain conditions, such as unclear or incomplete data about the Moon’s surface. This can lead to routes that aren’t reliable. Our approach uses artificial intelligence (AI) to analyze high-resolution lunar satellite imagery, creating detailed maps that not only show the terrain but also highlight uncertainties, like areas where the surface might be riskier than it appears. Our AI model predicts how safe different areas are for a rover to travel through, enabling smarter, uncertainty-aware path planning that produces more robust routes.
Featured Image
Photo by Nicolas Thomas on Unsplash
Why is it important?
This work is crucial because unreliable routes can jeopardize lunar missions. Current path planning methods can miss large hazards, such as craters or rocks tens of meters in size, or fail when rovers encounter unexpected terrain, which can lead to mission failure. By using AI to better understand terrain risks and uncertainties, our method creates safer, more dependable routes. Testing shows our AI models excel at predicting terrain risks and handling uncertainty, outperforming traditional methods. This means lunar rovers are more likely to navigate successfully, avoiding obstacles and ensuring missions achieve their goals, whether that’s exploring new areas or collecting vital scientific data.
Perspectives
With growing interest in lunar surface missions, more vehicles will navigate the Moon's rugged terrain. Currently, planning mission traverses is a time-consuming, labor-intensive process. It requires experts to carefully analyze each route segment while using outdated, suboptimal pathfinding algorithms. This inefficient approach is not sustainable for future missions. As the number of lunar traverses increases, we need a system like Google Maps to autonomously and accurately plan global routes for rovers. Our work lays the foundation for such a system, enabling efficient and reliable path planning to support the next era of lunar exploration.
Garrett Schueller
University of Texas at Austin
Read the Original
This page is a summary of: Analyzing Lunar Imagery Using Computer Vision Models to Improve Path Planning, July 2025, American Institute of Aeronautics and Astronautics (AIAA),
DOI: 10.2514/6.2025-4028.
You can read the full text:
Contributors
The following have contributed to this page
