Things look pretty tough right now for the Terminator.
Forget “I’ll be back”… some robots are struggling to even get there in the first place.
Take Alphabet Inc. (GOOGL), for example, which recently put its Boston Dynamics robotics division up for sale.
Having only acquired the company at the end of 2013, it’s not like Boston Dynamics isn’t putting in the effort or making substantial progress. Far from it, in fact. (Take a look at some clips on YouTube if you don’t believe me.)
Rather, it’s rumored that since much of Boston Dynamics’ work is with DARPA (Defense Advanced Research Projects Agency), Alphabet has become impatient with working at the speed of the Department of Defense.
The company was also concerned that if it became associated with making autonomous killing machines, people might be uneasy about buying its other products.
But the biggest barrier to widespread adoption of military robots at the moment is that some of them aren’t true robots at all.
Problems include excessive weight, low battery life, lack of battlefield durability, and requiring one or more full-time operators. So, while a device such as a bomb disposal robot may keep troops at a safe distance, it still requires human operators to be relatively close to the robot.
It’s easy to see how that might actually require more manpower, not less, as troops would need to defend the operator from danger.
But the U.S. Marines are looking to change that.
Multiple Robots… One Operator
They’re currently working on a program called the Unmanned Tactical Autonomous Control and Collaboration (UTACC).
This uses multiple robots and artificial intelligence that enables robots to work together to provide reconnaissance for the Marines.
This is a crucial tool in terms of intelligence and saving lives.
Currently, when Marines enter a hostile area, they typically have detailed maps of the area and at least some intelligence about where threats may occur, but humans still need to be present to actually detect those threats. With UTACC, multiple robots would accompany a squadron, both on the ground and overhead, to provide a threat assessment.
And instead of each robot needing its own human operator, there would just be a single operator for multiple robots.
That person wouldn’t have to micromanage each robot, either. He or she would simply tell the robots what to do and they’d know how to do it. In the meantime, the actual “intelligence” of the system could be hundreds of miles away on a ship.
So how would this work in practicality?
Well, an order to the robots might be, “Check the streets for potential threats – snipers, IEDs, arms caches.”
The robots, pre-equipped with a map and working together, would move to carry out the task. And if one of the ground robots can’t see a certain area, it would autonomously request that an aerial drone fills in any missing details. The computers onboard the ship coordinate with the robots, allowing the robots themselves to carry a smaller amount of computing power.
That’s the exact scenario that the Marines tested two weeks ago at the Marine Corps Base in Quantico.
Its software system is called DRAGON (Distributed Real-time Autonomously Guided Operations Engine) and was put together by the robotics lab at Carnegie Mellon University.
- The Marines instructed the ground robot to find a certain target, which it did.
- Then the Marines hid the target.
- Unable to find it, the ground robot and computer system instructed the drone to help. With the assistance of the drone, the robot again found the target.
If it were a real target, the robot’s operator could then have directly ordered a missile strike.
There’s Still One Thing That Robots Won’t Do
The Marines are testing other autonomous robots, too.
For example, they’re looking at an unmanned helicopter that can fly into hazardous areas and drop cargo to Marines on the ground.
They’re also exploring robots that can quickly load and unload cargo from aircraft, allowing each plane to make more deliveries during its mission window. With each project, the idea is that the robots have to be fully or mostly autonomous. Robotic systems that don’t lower the workload for humans are no longer of much interest.
One thing you won’t see robots doing anytime soon is autonomously fire weapons. Even the most advanced unmanned aerial vehicles still require a human to decide whether (and when) to pull the trigger.
This is intentional. There’s a broad consensus that even as technology advances rapidly, battlefield robots are nowhere near the level of sophistication required to make life-or-death decisions.
So for at least the next couple of decades, consider yourself safe from a renegade army of Terminators out to destroy us. Instead, robots are being used to save the lives of our troops.
To living and investing in the future,