Artificial Intelligence Could Help Neutralize Enemy Bombs

By Jon Harper
A ground robot controlled by a Marine EOD technician investigates a potential improvised explosive device.

Photo: Defense Dept.

As improvised explosive devices proliferate and evolve, the U.S. military’s explosive ordnance disposal community hopes to leverage artificial intelligence to defeat threats and protect technicians.

The use of IEDs by militant groups is growing more sophisticated and widespread. The Pentagon’s Joint Improvised-Threat Defeat Organization must therefore acquire better technologies that can sense, detect and neutralize enemy bombs, said Army Lt. Gen. Michael Shields, director of JIDO.

“Future initiatives and solutions must take advantage of new and predictive algorithms strengthened through deep machine learning, artificial intelligence, the integration of neural networks, autonomous systems, and manned and unmanned teaming,” he said during his keynote address at the Global Explosive Ordnance Disposal Symposium and Exhibition in North Bethesda, Maryland, which was hosted by the National Defense Industrial Association and the EOD Warrior Foundation.

Shields’ vision meshes well with the Pentagon’s so-called third offset strategy, which aims to use artificial intelligence — among others technologies — in innovative ways to help the U.S. military maintain its warfighting edge.

“This is really about bringing some of the principles and concepts behind the … third offset to the EOD community through robotic systems,” Rich Thissell, a contractor science and engineering expert for JIDO, said during a presentation at the symposium.

For example, the technology could play a major role in intelligence, surveillance and reconnaissance, Shields said.

The EOD community hopes to use algorithms and computer vision to enable machines to perform tasks like object recognition and classification.

Unmanned aerial systems or ground robots equipped with the technology could potentially identify aluminum powder, passive infrared sensors and other bomb-related components as the vehicles maneuver through hazardous areas, he said.

“Imagine a scenario where we’re able to take [an autonomous] platform, send it into a building … looking for that house-borne IED or booby-trapped” items without having to put troops in harm’s way, he said.

The Islamic State terrorist group has placed a variety of traps in homes and buildings in urban areas like Mosul, Iraq. They include trip wires, passive infrared sensors and pressure plates, officials noted.

One way to potentially deal with that threat would be to deploy a small aerial drone or ground robot inside a building and have it map and image the contents. The technology for mapping a single level has already been demonstrated, said JIDO scientist George Pappas. Swarm technology could someday be used to map multiple levels of a building or even an entire city block while searching for IEDs, officials said.

Having autonomous platforms that can go into a hazardous environment, recognize threats and report back “will give you pretty good situational awareness,” Jon Young, J-8 requirements division chief at JIDO, said during a speech at the symposium.

The services are brainstorming operating concepts.

Col. David Schmitt, EOD branch chief, Army G-38, said: “One of the things we’re looking at is … are there ways of leveraging other technologies such as autonomy in order to help the operator get the robot or whatever the remote platform is down on target and get a picture of the device?

“And then are there ways to leverage either artificial intelligence or other … advanced computing-type options in order to analyze the data that comes out of that?” he asked during a panel discussion.

The goal is to reduce EOD technicians’ time on target, he said.

Capt. Scott Kraft, commanding officer at the Naval Surface Warfare Center Indian Head technology division in Maryland, said artificial intelligence and big data analytics could potentially help technicians more quickly recognize exactly what type of bomb they are dealing with and choose the best option for neutralizing it. The vast amount of data collected during the past 16 years of war could be exploited to make faster decisions in combat situations, he said.

“What we need to do is bring those data sets together in what I call a common virtual architecture so that algorithms can work across those data sets in useful ways to aid decision-making for the EOD technician posed with a challenging set of circumstances,” he said in an interview.

The technology could tap into “all the collective knowledge that we’ve gained and we have in our midst” to help technicians accomplish their mission, he added.

Meanwhile, the Navy wants to take advantage of advances in autonomous systems for mine countermeasure missions. The sea service is looking to industry for help in this regard, said Rear Adm. Brian Brakke, commander of Navy Expeditionary Combat Command.

Unmanned underwater vehicles now have sufficient detection capabilities, he said. But personnel have to retrieve the data collected by UUVs before they can analyze what they found, he noted. The service wants to be able to defeat threats more quickly after they are detected.

“There’s an entire detect-to-engage sequence that we go through from the time from identifying the mine to neutralizing the mine,” Brakke said.

The Navy needs systems with a real-time or a near real-time capability for automated target recognition and through-water data transmission, he said. The technology could “get that information out of the UUV and up to us faster to be able to do that mission analysis,” he added.

Additionally, swarm technology could be useful for amphibious assaults, he noted.

“We also want to figure out how to navigate through that surf and onto the beach and how [unmanned surface vehicles] could help us with that,” Brakke said.

The systems would need to be expendable, he said.

“Think back to the tanker wars” of the 1980s in the Persian Gulf, he said. “Think of industry saying, ‘I’m going to use a dual-hulled ship to just clear, and if it strikes a mine it strikes a mine’ — so a guinea pig-type mentality.”

A swarm of unmanned systems could help the Navy eliminate mines and clear the way for Marines to invade from the sea, he added.

Drones and artificial intelligence could also facilitate the clearance and repair of airfields that come under attack, officials said.

Edwin Oshiba, Air Force deputy director of civil engineers, said time is of the essence in high-end warfare against near-peer competitors, where the challenge of clearing damaged airfields of unexploded ordnance and getting the strips back up and running will become more daunting.

“We’ve typically talked about doing that for repair … of craters if you will — maybe tens to hundreds of pieces of ordnance,” he said. “We’re going to be in a place I would say fairly soon where we’re not talking about tens and hundreds, we’re talking about hundreds and thousands. And we’re talking about being able to do that in minutes and hours, not days and weeks. We have got to be able to get operational much faster.”
Brakke said artificial intelligence could speed up the process.

The Navy wants unmanned systems equipped with deep machine-learning capabilities that could “run across that field, scan it and tell me what type of ordnance I have left, what condition it’s in, what craters do I have, how can they be filled, and then map ... [out] a plan” for resolving the problem as quickly as possible, he said.

AI could also help EOD forces defeat electronic warfare threats by detecting sources of transmission and interference, officials said.

“The electromagnetic spectrum is now the new high ground on the battlefield,” Young said. U.S. troops “have to have situational awareness of it, what’s happening and why, and if we don’t we’re going to be at a disadvantage.”

Signals interference can impede the operations of robots and other EOD tools.

“If you’ve been to theater lately … you’ve heard about a lot of the counter-UAS systems along with all the jammers, along with all the electronic warfare systems,” Young said.

“It becomes very complex. So we want to try to simplify that” for operators that aren’t EW experts, Young said.

Schmitt noted that the Army has a renewed focus on the need for soldiers to be able to operate effectively in contested or congested electronic environments.

“Is there a way to analyze and scan the RF spectrum and then deconflict in real time with our crew systems and our comms systems so that we can better manage the available spectrum in order to get the job done?” he asked.

Doing so could enable operators to use standoff technology instead of manual approaches to detect and neutralize explosives, he added.

Although artificial intelligence holds great promise, technical challenges must be overcome before the EOD community can fully leverage it, Thissell noted.

The goal is to have real-time detection, recognition and identification of objects of interest by autonomous ground- and aerial-robotic systems in outdoor, indoor and underground locations, he said.

That requires an integrated system that uses autonomous platform navigation, sensor controls and computer vision exploitation simultaneously, he said. Each of those tasks is difficult to master. Doing all three at the same time is “a really hard problem,” he said.

Autonomous vehicles must have depth perception and object sensing capabilities for collision avoidance. Additionally, such systems require a lot of software, he added.
Thissell envisioned a family of platforms with different payload capacities supporting plug-in sensors, data links with open frameworks, software development kits and application programming interfaces.

Open system architectures will be required to integrate these technologies, he noted. Innovation will be coming from “a lot of different companies, a lot of different sources,” he added.

Advances in technology have made object detection capabilities more reliable, Thissell said. “The revolution in object detection algorithms brought about by convolutional neural nets and deep neural nets — it has really increased the probability of detection.”

However, computers must be able to recognize an object before they can classify it. “We need to create a labeled data set for training these algorithms [to identify] lots of different objects,” he said. Doing so is very manpower intensive, he noted.

Despite the recent advances in artificial intelligence, machines are currently unable to perform the full EOD mission set, officials said.

Having fully autonomous systems that can do the job entirely on their own “is a ways off for us,” Shields said. Right now, JIDO is focused on developing a roadmap and enabling future architectures to integrate deep machine-learning algorithms to increase the efficiency of intelligence analysts and technicians, he added.

Meanwhile, officials at Indian Head recently initiated a project to develop “a digital EOD” operating concept, Kraft said. “It can and will be transformative if we do it smartly.”

The effort is still in its early stages, he noted. “It really has to be fed and informed by industry and by JIDO and by the intelligence community. But we’re trying to get to a start point that we can all rally around and focus on and maturate,” he said.


Topics: Robotics, Robotics and Autonomous Systems

Comments (0)

Retype the CAPTCHA code from the image
Change the CAPTCHA codeSpeak the CAPTCHA code
Please enter the text displayed in the image.