Student enjoys NASA visiting researcher placement on AI

Eleni Bohacek speaks about her fully-funded visiting researcher placement with NASA on a Frontier development lab over the summer in Silicon Valley.

Congratulations to Eleni Bohacek, a PhD student supervised by Dr David Selviah in UCL EEE's Photonics Group who was selected to complete a fully-funded visiting researcher placement with NASA on a Frontier development lab over the summer in Silicon Valley. Eleni was applying Artificial Intelligence techniques to problems in planetary defence and space resources.

NASA Frontier Development Lab is an AI R&D accelerator with Intel as a key partner, that tackles knowledge gaps useful to the space program. The program is an intense 8-week concentrated study on topics not only important to NASA, but also to humanity’s future. This year, the Lab are focusing on challenges within the fields of Planetary Defence, Space Resources, and Space Weather.

 Eleni said “Lunar water and volatiles are a space resource that government and private space agencies are beginning to explore, however there are many technological barriers that need to be overcome before this is possible. My group was tasked with tackling one of these barriers using AI, and we chose to work on improved mapping of the Lunar poles, the location of the Moon’s water reservoirs.

We trained a convolutional neural network to identify craters on satellite images and elevation models of the lunar surface to a greater accuracy and at faster speeds than conventional image processing methods.

We were a diverse group of researchers with expertise ranging from geoinformatics, rover navigation and machine learning, and from me, computer vision and planetary science. This was my first experience working with AI in research: I helped to create the training and test data, training the model, and comparing this to the state of the art in crater detection. Along the way I gained experience in Python and Neon (a deep learning framework), and re-familiarised myself with ArcGIS and JMars.

The experience opened my eyes to the power of neural networks in solving science problems and I was inspired by the successes of the other FDL groups applying AI to solar physics, asteroids and comets. I will use the methods I learned for my PhD project, as machine learning provides a complimentary approach to my current methods, moreover this subject will be key to realising the challenge of autonomous rover navigation in space.”

Eleni is a student of the UCL-Cambridge Centre for Doctoral Training (CDT) in Integrated Photonic and Electronic Systems (IPES).