Thu, May 11|
AEG-DMV May 2023 Meeting with Potomac Geophysical Society (PGS)
How to Train Your Lander: Using Simple Machine Learning to Solve Complex Problems
Time & Location
May 11, 2023, 7:00 PM – 9:00 PM
North Bethesda, 11414 Rockville Pike, North Bethesda, MD 20852, USA
About the event
The costly power and data volume limitations of delivering seismic data back to Earth from planetary missions requires the development of algorithms for lander-side signal analysis for telemetry prioritization. This is difficult to explicitly program, especially if no prior seismic data are available from the planetary body. We demonstrate that Convolutional Neural Networks can be used to accurately catalog planetary seismicity without local training data by building binary noise/signal classifiers from a single Earth seismic station and applying the models to moonquakes from the Apollo Passive Seismic Experiment (PSE) and the Lunar Seismic Profiling Experiment (LSPE). Instead of assessing the seismic signals by their time series, we convert the data to spectrogram images, which are easier to generalize with only a fraction of data. Simple two- to five-layer convolution models are built using solely 580 seismic events and tested against a subset of 200 Grade-A events from the PSE and obtained station accuracy averages of 89–96%. The model with highest accuracy is used to create the first systematic seismic detection catalog for the LSPE continuous seismic data. We find two classes of seismic events: impulsive events generated by the lunar module left by the astronauts, and emergent events created by regolith processes that vary with temperature. This study proves that not only is landerside decision making possible using machine learning, but it can be achieved with simple algorithms and datasets if the problem is set up in the correct way.
Presenter: Francesco Civilini (NASA Goddard Space Flight Center)
Mr. Cvilini is a computational geophysicist and postdoc researcher at NASA Goddard focusing on algorithm development for planetary seismology working within the Solar System Exploration Data Services Office. He received his bachelor’s degree in geophysics from UC Santa Barbara and immediately began a Master’s degree, focusing on site-response seismology and seismic hazard. He then traveled to New Zealand to do a PhD at Victoria University of Wellington on ambient seismic noise tomography of volcanic areas. Since then he has completed postdocs at the US Geological Survey, NASA Marshall Space Flight Center, and Caltech.
Meeting & Presentation