Deep learning will help future Mars rovers go farther, faster, and do more science

NASA JPL are developing autonomous capabilities that could allow future Mars rovers to go farther, faster and do more science. Training machine learning models on the Maverick2 supercomputer at the Texas Advanced Computing Center, their team developed and optimized models for Drive-By Science and Energy-Optimal Autonomous Navigation. The team presented results of their work at the IEEE Aerospace Conference in March 2020. The project was a finalist for the NASA Software Award.