About the event
Presented by Dr. Anthony Fragoso, Staff Scientist at the Graduate Aerospace Laboratories at the California Institute of Technology
Field robots must consider the statistics of their environments to operate safely in adverse conditions. Conventional approaches to autonomy can be engineered to provably accommodate general operating geometries, but rely on fragile perception heuristics that are rarely satisfied in practice. Deep learning, on the other hand, excels at handling complex image and sensor statistics but can only reliably use data similar to its training set. Environment-driven autonomy aims to resolve this contradiction by considering conserved quantities and engineered advantages whenever possible, and restricting deep learning only to the extent needed to faithfully capture the statistics of a robot’s environment. In this talk I will discuss the use of environment-driven autonomy for trustworthy robotic navigation in highly challenging aerial and automotive scenarios.
Dr. Anthony Fragoso is Staff Scientist at the Graduate Aerospace Laboratories at the California Institute of Technology. His work focuses on field robotics, particularly with aerial and automotive applications. He received a Ph.D. in Aeronautics from Caltech in 2018, under the supervision of Prof. Richard Murray. His thesis research was conducted as an associate at the NASA Jet Propulsion Laboratory and investigated vision-based motion planning of micro air-vehicles in GPS-denied environments. Anthony also received a Bachelor of Science in Mathematics and Physics from Yale University in 2013.