News that the Army is developing a fleet of autonomous ground robots to assist soldiers on complex and dangerous missions is nothing new.
However, the way these robots will learn how to maneuver and interact with soldiers is pretty revolutionary.
According to a recent report on TaskandPurpose.com, the goal of researchers from the U.S. Army Combat Capabilities Development Command’s Army Research Laboratory, who are working in collaboration with the University of Southern California’s Institute for Creative Technologies, is to develop a platform that allows for conversational dialogue between soldiers and autonomous systems.
The thought is that this ability will place less strain on soldiers having to remember precise commands under challenging situations.
So, for example, instead of having to remember the exact command for telling the robot to reach a certain point by a certain time, the Joint Understanding and Dialogue Interface (JUDI) allows the robot to understand what is meant by “get to the f@#%&! rally point before this whole situation turns to s%#!.”
And because JUDI is a platform that runs both ways, the robot might very well respond with “don’t f#%&! worry about me, just get your sorry a$$ there on time.” Or, it could ask questions to receive more clarification on what is needed.
In addition to adjusting to the colorful language often employed by our nation’s finest, this back-and forth allows JUDI to learn more about the operating environment, so terminology and commands can be the same when talking to a person or a machine.
JUDI is as much about understanding the situation and intent of the dialogue as it is the actual words being spoken. This allows the robot to adjust to the situation, unlike voice-activated assistants like Siri or Alexa.
While it’s unclear when JUDI will come online, its initial applications will probably focus on search and rescue, with longer-term uses potentially including reconnaissance functions or autonomous fighting vehicles comparable to the M1 Abrams tank.