A Machine Learning-Driven Optimization Strategy for Icy Regolith Extraction in the Shackleton–de Gerlache Ridge Crater Chain - SME Annual Meeting 2026

This study introduces a machine learning strategy to optimize icy regolith extraction from the Shackleton-de Gerlache Ridge using autonomous mining equipment. The model incorporates key parameters—distance, road gradient, vehicle load, speed, recharge logistics, and regolith water content—into a cost framework for excavation, hauling, and dumping. Regression models estimate energy and time requirements; classification predicts terrain navigability; clustering groups craters for efficiency; and K-Nearest Neighbors (KNN) determines optimal routing and scheduling. The system adapts to a six-month daylight cycle to maximize productivity and minimize costs, enabling intelligent planning for sustained lunar resource extraction under extreme environmental and logistical constraints.