Science Daily July 22, 2022
Explosions produce infrasound waves capable of propagating globally, but the spatio-temporal variability of the atmosphere makes detecting events difficult. Machine learning (ML) is well-suited to identify the subtle and nonlinear patterns in explosion infrasound signals, but a previous lack of ground-truth data inhibited training of generalized models. A team of researchers in the US (University of Alaska, Air Force, University of Mississippi, Los Alamos National Laboratory) has developed a physics-based method that propagates infrasound sources through realistic atmospheres to create 28,000 synthetic events, which are used to train ML classifiers. A simple artificial neural network and modern temporal convolutional network discriminated synthetic events from background noise with >90% accuracy and, successfully identified most real-world explosion signals recorded during the Humming Road Runner experiment. ML models trained entirely on physics-based synthetics advanced explosion detection capabilities and made ML more viable to related fields lacking training data…read more. Open Access TECHNICAL ARTICLEÂ
Humming Roadrunner (HRR) regional infrasound stations. Credit: Geographical Research Letters, Volume49, Issue11, 16 June 2022, 2022GL097785Â