Luka Takki
Class of 2028Luxembourg, District of Luxembourg
About
Hello, my name is Luka and I am developing LIPAS (Lunar Impact and Predictive Analysis System), a project that utilizes machine learning to accurately forecast lunar space weather. I chose this focus because current predictive models used by agencies like NASA and ESA rely on complex, time-consuming mathematical simulations. As future lunar missions demand instantaneous forecasts, I am building LIPAS to bridge that gap with real-time predictions.Projects
- "LIPAS.: Building and Evaluating a Cross-Hazard Machine Learning Approach to Unified Lunar Forecasting" with mentor Kyle (Working project)
Project Portfolio
LIPAS.: Building and Evaluating a Cross-Hazard Machine Learning Approach to Unified Lunar Forecasting
Started Mar. 19, 2026
Abstract or project description
L.I.P.A.S. (Lunar Impact and Predictive Analysis System) is an innovative machine learning model designed to accurately predict and assess key environmental conditions on the Moon, including temperature, radiation, dust activity, micrometeorite impacts, solar storms, and minor moonquakes. By combining space weather data from NASA and NOAA APIs with historical data, machine learning, and probabilistic modeling, L.I.P.A.S. aims to deliver accurate short and medium term forecasts through an interactive, web-based dashboard for multiple lunar locations. By leveraging machine learning, L.I.P.A.S. not only integrates diverse hazard data but also adapts forecasts to evolving lunar conditions, offering practical decision support for mission planners.
In this research, I investigate whether integrating machine learning and probabilistic modeling into a single platform improves predictions and operational usability compared to existing single-purpose lunar environmental models. To do this, I am applying a two-part comparative framework, evaluating hazard coverage, data integration, temporal forecasting capability, update frequency, and usability, alongside a 14-day numerical accuracy study comparing L.I.P.A.S. against NASA flagship models across all six hazards.
Overall, this project demonstrates how combining scientific research, advanced coding, and publicly available datasets can produce a practical tool that contributes to safer and more predictable lunar exploration.