Project Description

The number of known near-Earth asteroids and potentially hazardous asteroids has exponentially increased over the last few decades. Asteroids have been the focus of several recent robotic space missions from NASA and internationally, and are more frequently being considered by companies as economically viable targets for in situ resource utilization. Whether for scientific discovery, planetary defense, or resource utilization, the tasks of effectively controlling the orbit of an asteroid and interacting with the asteroid body remain particularly challenging. This project investigates the application of deep learning algorithms to the topic of asteroid redirection, particularly, deep reinforcement learning concepts applied to low-thrust trajectory transfer optimization. Additionally, there are opportunities to investigate machine learning and computer vision approaches for asteroid characterization from granular models and image data for applications in robotics and spacecraft observation.

Qualifications

1. Experience with coding languages (e.g., MATLAB, Python, C++) an asset

2. Experience with orbital mechanics, robotics, computer vision, or granular modelling an asset

3. Experience with machine learning, reinforcement learning and deep learning an asset