Charles Stark Draper Laboratory, Inc. is designing an autonomous software suite for small, handheld unmanned aircraft systems (UAS) weighing about half a pound or less–systems that may fly 20 minutes round trip to search and identify threats to soldiers so that they are able to enter buildings and other complex, dangerous environments better prepared.
Draper is working on the project under a three-year, U.S. Army contract awarded in 2018 and potentially worth $3.2 million. Funded by the Army Research Development and Engineering Command, the Army Materiel Command (AMC) contract is part of the Army’s Soldier Borne Sensor Program.
Army officials may get a demonstration of the capability on a surrogate drone at Draper this September.
“What we’re seeing now are a greater proliferation of smaller UAS getting pushed down to the squad and platoon level to enable organic ISR [intelligence, surveillance, and reconnaissance] capabilities, but a lot of these systems really require a lot of involvement on behalf of the soldier to operate them so they’re still pretty manually intensive,” said Jon Cash, a senior program manager at Draper.
“What we’ve been asked to do is develop the software and algorithms to make these small systems a little bit smarter so they can start to do more things on their own and begin to free up the soldier from having to intensively fly these and operate them,” he said.
The software is intended to be platform agnostic and scaleable to larger platforms.
Handheld drones appear to be increasingly important to Army operations.
In the spring of last year, soldiers from the U.S. Army 82nd Airborne Division’s 3rd Brigade Combat Team headed to Afghanistan received Black Hornet Personal Reconnaissance Systems (PRS) — drones designed to provide platoon and squad-level intelligence on enemy units
.
Oregon-based FLIR Systems [FLIR] makes the Black Hornet. In January last year, the company received a nearly $40 million Army award for the micro drones, which FLIR Systems says provide the “non-specialist dismounted soldier with immediate covert situational awareness” and “game-changing EO [electro-optical] and IR technology” that “bridges the gap between aerial and ground-based sensors, with the same SA [situational awareness] as a larger UAV and threat location capabilities of UGVs [unmanned ground vehicles].”
In May, FLIR received a follow-on order of nearly $21 million for the drones under the Army’s Soldier Borne Sensor program.
Designing systems for such small UAS poses a challenge.
“The challenges we face are trying to come up with efficient, yet effective, software tools and algorithms that are able to be run on the platform itself,” Cash said. “As you get into bigger systems, you have the luxury of increased processor power, increased batteries, being able to add additional sensors like LIDAR or radar. But once you get down to the handheld size, what you’re really relying on is a simple, monocular camera and very little processing power onboard, having to work on a space that’s not much bigger than a cellphone. That’s where this effort is unique and game changing, trying to bring advanced autonomy to that class of UAS.”
The Draper software is to allow 3D mapping and leverages artificial intelligence/machine learning for object recognition.
“We have a library of images from chairs to tables to lamps to potted plants, boxes, and we’ve trained this algorithm to recognize these images,” Cash said. “We have a library of a couple hundred images. What that enables the system to do is, as it’s flying, it can classify an object, how far away it is, and the dimensions of it. The combination of 3D mapping and object classification and recognition are some of our key tools for being able to allow the drone to interpret its environment. Then we bring that into a mission planning node. As it’s feeding information about its environment, it’s also updating its planned path, its flight plan, in real time to be able to determine how it’s going to navigate from Point A to Point B.”
The Draper software for the small UAS employs a “novel” user interface that is able to provide a soldier with real-time information about the environment observed by the UAS, objects of interest in that environment and the drone’s operational status without distracting the soldier or overloading the soldier with information, Cash said.
Draper is looking into the possibilities for collaborative autonomy so that one soldier can operate multiple, small UAS.