Students Turn to On-Board AI and Optical Sensors to Keep Nanosatellites From Colliding

Students at Germany’s Julius-Maximilians-Universität Würzburg (JMU) are to embark on a two-year project, dubbed KI-SENS, which aims to prevent nanosatellites from colliding in orbit using artificial intelligence.

The KI-SENS project centers around technology originally developed for the SONATE-2 satellite, a project of space technology professor Hakan Kayal and colleagues, which is designed for Earth observation work. SONATE-2 will join an ever-growing number of micro- and nanosatellites in orbit as research organizations and commercial ventures alike seek to capitalize on the “Space 2.0” revolution — but as their numbers grow, so too does the risk of two or more colliding and falling back to Earth.

A group of 20 students from the non-profit WüSpace association aims to tackle the problem using on-board artificial intelligence, though they’re not directly working as a team: instead, they’re operating independently in a project that aims to give them an overview of the entire process, from planning and project management to construction and testing — and, Kayal and colleagues hope, will result in technology to help keep satellites safe by detecting dangerous approaches and taking evasive action.

“We will develop intelligent, optical sensors, build a prototype and test it on the ground under realistic conditions,” research assistant Tobias Herbst explains. “On the ground,” though, means something different in space terms: the prototypes will go through testing using satellite dummies held aloft by drones, allowing the researchers to try out different approaches in-flight without the cost or risk of an actual orbital launch.

The project is funded for two years, with the first tests expected to take place next year; additional information on WüSpace’s work is available on the association’s website.