Space and Exploration Technology

On 27^th May 2020, Giulia Viavattene was invited to present her PhD research in the series of colloquia "Machine Learning in Science" (a network of UofG researchers working with Machine Learning). She presented "Neural Networks in designing multiple near-Earth asteroid missions".

The series of Machine Learning in Science gathers researchers with different backgrounds for monthly presentations on the topic of machine learning. The series represents a unique opportunity to share the knowledge of Machine Learning, enhance the research and encourage interdisciplinary collaborations.

During her PhD research, Giulia is designing an Artificial Neural Network (ANN) to quickly estimate the cost and duration of low-thrust transfers between near-Earth asteroids. Integrating the ANN within the sequence search algorithm enables the identification of optimal candidates of sequences of asteroids to visit in a single mission. Giulia’s proposed methodology with ANN made the algorithm two orders of magnitude faster than other methods used in previous works, while maintaining a good level of accuracy of the solution.