SoarTech Uses DARPA-Funded Work to Improve US Army Training

SoarTech, as part of a team led by Amentum, demonstrated software for use in military training that determines hits and misses for simulated direct and indirect fires. The capability leverages computer vision capabilities originally prototyped by lead scientist, Ben Purman, under a DARPA STTR Phase II contract conducted by SoarTech with partner University of Michigan. It is an example of SoarTech transitioning academic research to the military’s toughest problems. STTRs require 30 to 40 percent effort from the university, while SBIRs can have a university partnership, but it is not required.

Drs. Jeff Craighead and Ross Hoehn, scientists from SoarTech, participated in the Army’s Synthetic Training Environment, Live Training Systems (STE-LTS) Soldier Touch Point 3 evaluation exercise at Fort Hood, Texas, from 31 May to 16 June 2022.

“The prototype SoarTech partnered to create represents an innovation that will provide warfighters with greater accuracy in simulated weapons fire, making for more realistic training than previous, similar devices,” Craighead said.

Using technology from the partnering organizations, SoarTech's prototype solution is built into a weapon-mounted optical sensor made by FN America. The sensor is about the size of a handgun-mounted flashlight, which can attach to various small arms, such as an M-4, M-249 or AT-4.

Cornet Technology provides a man-packable LTE/5G cell that increases network coverage and enables edge computing capabilities, which are used to run part of the adjudication stack that judges the hits and misses of simulated fire. The LTE and 5G communications relay adjudication messages between players and white cell operations.

A fidelity-enhancing technique employed by SoarTech’s adjudication system includes a computer-vision component that identifies and segments objects that the weapon-mounted optic observes when a player shoots. The system attempts to identify humans, vehicles, buildings, and other objects, which are all classified into types and materials used by the damage model that calculates ballistics information, such as firing position, round type, and weapon orientation.

These factors help determine if and where a target was hit, how much damage the object or player took, and how it affected their operation – for example, whether the hit completely disabled the target. The adjudication system notifies players of the hit or wound, and it disables subsequent shots from weapons in the event of a “kill.”

“We conducted testing during day and night operations and the system performed as expected for the current state of development,” Craighead said. “Soldier Touch Point 3 gave us a great opportunity to review and identify areas where the solution needs enhancement. Getting direct feedback at these evaluation events from the Soldiers in the field is a very efficient way to adjust and help us improve Army training.”

 The next Soldier Touch Point event is scheduled for September in Fort Bragg, North Carolina. The Amentum team includes SoarTech, FN America, and Cornet Technology.