Current autonomous systems need to actively guide it to navigate in the dark (headlights, lidar, or some other emitting sensor), but these active signatures can also be detected by a distant adversary. DARPA has selected four industry and university research teams for the Invisible Headlights program, which aims to determine whether self-driving cars can navigate in complete darkness using only passive sensors.

Invisible Headlights’ technology technology can expand the environments and types of tasks in which autonomous assets operate – allowing them to operate at night, underground, arctic and foggy.

What’s more, a deeper understanding of the information contained in ambient thermal radiation will inform future applications utilizing infrared light.

The Invisible Headlights program aims to eliminate this vulnerability by discovering how to convert ambient thermal light present in all environments into a completely passive 3D sensor that is accurate and fast enough to support autonomous navigation. Researchers will explore and quantify the information contained in these environmental releases, and then create new sensors and algorithms to exploit this information.

The research team will address these challenges:

+ The Arete Associates team in Northridge, Calif., will perform virtual analysis to understand the ambient spectral and polarized infrared environment and develop low-contrast 3D vision algorithms.

+ The Kitware, Inc. team in Clifton Park, New York will use a custom multi-band hyperspectral infrared camera combined with artificial intelligence and machine learning algorithms to estimate local 3D scene structure and navigation semantics.

+ A team led by the Massachusetts Institute of Technology in Cambridge, MA is working to expand superconducting nanowire single-photon detector (SNSPD) technology into very low-noise infrared sensing arrays.

+ A Purdue University team in West Lafayette, Indiana, is developing new ultrafast, spin-based sensors and 3D vision methods to exploit the properties of the surrounding thermal environment.

“These teams are looking for innovative ways to use infrared spectroscopy,” said Joe Altepeter, program manager for the Invisible Headlights program in the Defense Department’s DARPA Office of Defense Science. “They are exploring ways to capture more scene information using new equipment, improved algorithms and increased measurement diversity.”

In addition, a government team led by the U.S. Army C5ISR Center Night Vision and Electronic Sensors Committee (NVESD) is accelerating the research team’s work by assembling an unprecedented set of ground-based hyperspectral and polarimetric images with correlated 3D ground truth.

In the first phase, project executives were tasked with investigating whether thermal radiation contained enough information to enable autonomous driving in very dark conditions. In Phase 2, the team was tasked with designing and testing sensors and algorithms to demonstrate that a real system could measure enough information for 3D vision. The final phase will build and test passive demonstration systems that will be compared to active sensors in field tests.

Invisible Headlights technology can expand the environments and types of tasks in which autonomous assets operate – enabling them to operate at night, underground, arctic and foggy. What’s more, a deeper understanding of the information contained in ambient thermal radiation will inform future applications utilizing infrared light.