Shashwat Raj
Computer systems engineering
Hometown: Tempe, Arizona, United States
Graduation date: Spring 2027
Additional details: First-generation college student
FURI | Fall 2024
Optimizing Earth Science Observations: Developing Reinforcement Learning Techniques for Autonomously Determining Priority Observations in a Dynamic Environment
Dynamic Targeting is an extensively researched concept used by satellites to point the instruments in the optimal direction to get the maximum scientific yield. This research aims to enable satellites to dynamically determine valid observation locations in real time. Utilizing state features like geolocation, sun-angle relationship, ground status as well as convective precipitation, the research proposes a reinforcement learning technique using PyTorch to train Deep Q-Learning Network (DQN) agent and creating a reinforcement learning environment that emulates the satellite’s movement and thereafter predict the potential observation points. The project involves data preparation and feature engineering of the large simulated environmental data in NASA GEOS-5 dataset, model selection and training, testing and tuning of the model, and evaluation through comparison with existing approaches. The research will be considered a success when the results accurately correspond with the simulated data on convective precipitation storms, proving a greater accuracy than random/pre-decided targeting techniques or other predictive supervised learning models. The ultimate goal of the research is to autonomously enable satellites in prioritizing areas with high atmospheric activity, resulting in enhanced environmental monitoring and providing deeper insights to Earth’s atmospheric dynamics.
Mentor: Paul Grogan