Without a map, picking apart a robot’s brain leaves only viscera and circuits. Understanding how drones store information, and how to access that information, even if the machine has crashed or attempted to erase its own memories, means one less obstacle to tracking down a drone pilot.
To prepare for improved drone brain extraction, the National Institute of Standards and Technology announced in June that it has a library of forensic images of drones, a library it expects to include 30 commercial models by December.
The project started in May 2017, with VTO labs of Colorado awarded a contract from the Department of Homeland Security to produce the forensic images, or complete data extraction from the drone. That meant purchasing three drones of every type for the project, and then flying them to build a baseline of data. Steve Watson, chief technology officer at VTO, then extracted that data three ways: from an intact drone, from the circuit board and cameras of a disassembled drone, and from the chips directly of third drone.
“The forensic images contain all the ones and zeroes we recovered from each model,” Watson told NIST.
Formatted to match existing forensic standards, and paired with step-by-step guidelines as to how the data was removed, the images provide a map into the brains of recovered drone circuitry. Training on this data will let investigators learn the pathways to the relevant details needed to reconstruct a drone’s flight and actions after the fact. That’s valuable for law enforcement, and it may even have implications for understanding drone usage on the battlefield.
Commercial drones, like DJI phantoms, are cheap enough and useful enough that they’ve been spotted in battlefields from Iraq to Ukraine. The remote nature of the drones means pilots often have a window to escape if the situation looks untenable, but the abandoned drone contains information that can inform defenses and point to security flaws. It could also, possibly, lead back to where would-be assassins launched their drone-borne attack, letting investigators pull additional information from nearby cameras or question additional witnesses.
Extracting the data from damaged drones and even the chips themselves is crucial, since in many cases the drone will be best handled by a bomb squad or EOD team first. Finding relevant data, even on a machine built to explode, could lead back to bomb makers, or at least their accomplices.
Even if the Pentagon persists with a ban on commercial drones by its own forces, understanding how those drones work and store data will be an essential part of war to come.
Kelsey Atherton blogs about military technology for C4ISRNET, Fifth Domain, Defense News, and Military Times. He previously wrote for Popular Science, and also created, solicited, and edited content for a group blog on political science fiction and international security.