Cadets modernize an ammo plant through robotics research

By Jorge Garcia Pointer View Staff WriterMay 18, 2022

Cadets modernize an ammo plant through robotics research
(From Left to right) Class of 2022 Cadets Skander Guizani, Ethan Rabb, Isaac Hagberg, Alexander Murphy and Steve Butts present their final Warehouse Autonomous Robots capstone research project to staff at the Iowa Army Ammunition Plant on May 5 in their lab at the U.S. Military Academy. (Photo Credit: Jorge Garcia) VIEW ORIGINAL

Purpose, Progression, Practicality

Before retiring in 2021, Bruce Jette, U.S. Military Academy Class of 1976 graduate, functioned as the assistant secretary of the Army (Acquisition, Logistics, and Technology).

The “industrial base” was one of the facets he worked on, which enabled the research, design, assembly and maintenance of weapon systems in ammunition plants to meet the U.S. Army’s directives.

During that time, Jette would regularly pay visits to ammo plants to get a comprehensive idea of what revisions needed to be applied to enhance the Army warfighter.

One of the plants he would regularly visit was the Iowa Army Ammunition Plant. Since its construction in the 1940s, the IAAAP has remained virtually unchanged in its design. Each visitation would hit Jette with a sense of nostalgia for the ingenuity and technological prowess achieved during the mid-20th century.

“He called the IAAAP his ‘time machine,”’ Lt. Col. Steve Crews, an assistant professor in the Department of Civil and Mechanical Engineering, said. “‘Step into my time machine,’ he would say.”

Despite this, Jette believed changes needed to be implemented to streamline productivity and increase efficiency during the manufacturing process.

With that in mind, he considered what the employees had to endure and the risk they would have to face daily to produce ammunition successfully.

“Imagine spending all day picking up and putting down ammunition rounds, with each round weighing 100 pounds ... It’s a tough proposition,” Jette said.

As a result, workers at the IAAAP have sustained injuries during the manufacturing process, reinforcing the need for innovative change to increase efficiency and, more importantly, to safeguard the well-being of workers.

“Many workers (at the IAAAP) complain about back problems and it’s also very laborious and time consuming lifting those heavy shells,” Jermain Magette, the project officer for the IAAAP, said.

Jette added that even with lifting chains and motors, it still takes a considerable amount of time to complete the task, and it still requires an employee to handle the ammo rounds physically.

And so, after studying the capabilities of the IAAAP and other ammo plants, Jette, along with colleagues and other subject matter experts, came to a fundamental realization: there was a dire need to modernize the working conditions of the facility and get people out of unsafe environments within the workspace.

With this intention, he was sure that the application of automated robots would be the primary way to get it done.

“A worker has to think, ‘could the round pop out of the chain,’ ‘could it fall to the ground,’ ‘could it swing against me and injure me with its weight?’ With robots, people will transition from handling explosives to handling the robots that handle the explosives,” Jette said.

The novel idea kickstarted operations at West Point as Crews spent most of his tenure at the academy teaching cadets, who had no prior knowledge, how to build robots.

“The concept I explained to them was, ‘I don’t want you spending too much effort building a robot. I want you to build a robot quickly, prototype it, and then I want you to focus on some actual research,ʼ” Crews said.

Throughout the process, Crews spent many semesters designing and then fine-tuning the lesson plan for cadets so that they could work on materials usually reserved for the graduate level.

“I wanted them to do algorithms, I wanted them to do SLAM (Simultaneous Localization and Mapping), I wanted them to do machine learning; I want them to apply these concepts and they’ve been able to perform all of these different elements within the lab, and it’s been fantastic,” Crews said.

Last year, in the beginning of their research, Class of 2022 Cadets Alexander Murphy, Ethan Rabb, Steven Butts, Skander Guizani, and Isaac Hagberg ventured to the IAAAP to acquire data and practical knowledge of how they operationalize in the plant. They also took notes on the architectural structure of the facility to model their laboratory at Mahan Hall, similar to the plant.

“We were allowed to see everything,” Butts said. “We were allowed to take notes and conceptualize everything we saw in our mind while being escorted by a government employee or the supervisor of whatever station we were at.”

They were allowed to observe, but they could not take photos or record videos of the warehouse due to operational security.

Once they had returned from Iowa, it was time to get in the lab, buckle down and work through two painstaking semesters of crafting robotic prototypes with safety at the forefront of their minds.

Applying the Pressure Where it Counts

It was April 18, and the cadets were underprepared.

They had just finished presenting their Warehouse Autonomous Robots (WAR) capstone project to officers and civilian professors at Mahan Hall.

As the onlookers exited the laboratory, Crews hit the cadets with a poignant reality check.

“The presentation did not go well,” he said. “However, you need to experience failures under stress with people watching you because otherwise, imagine all this happening with the Secretary of the Army or amongst generals?”

The WAR team spent two semesters developing an intricate vision and touch safety system meant to “stop, slow or alter robot movement,” Hagberg explained during a presentation.

First, they developed the robots’ mechanical structure, one with an omnidirectional base and a triangular frame with ‘mecanum’ wheels attached to the three ends of each module and a robotic arm attached to its surface to lift items and place them in certain areas.

The second robot served as a vision and touch safety prototype in the form of a robotic arm using HEBI modules capable of lifting the top four inches of an M795 round and moving them from one location to another.

The team plugged away at mathematical equations and coding sequences, teaching the robot arm how to identify the presence of humans relative to itself.

Leading up to their first presentation, the cadets thought they were ready. However, something in the code was amiss.

While the cadets properly explained what the robot arm could do, the code they had programmed into it did not allow it to perform tasks correctly.

When the robot arm didn’t move in short, aggressive spurts, a mistake in the coding sequence caused it not to move at all, and most importantly, it still couldn’t detect a human’s presence.

Safety, more so than anything else, was the selling point. If the robotic arm could not identify a human’s presence on the work floor while handling ammunition, the capstone project would be considered a pointless endeavor.

“They needed to fail their first presentation. They needed to feel ashamed of their work,” Crews said. “I was that cadet before. I came to present my project, and it just did not work. Then my advisor said, ‘we’re just gonna move on to the next thing,’ and they just skipped me. So, (the WAR team) needed to see that they don’t have it all together before Projects Day.”

For this reason, the cadets swung back into action, fine-tuning the robotic arm’s capabilities.

Moreover, correcting the errors came with improving the quality of their presentation.

Live demonstrations can come with their fair share of challenges. For example, products can break, malfunction or not work when prompted.

However, bringing someone outside of the team to test the product can be highly beneficial in understanding what needs to be tuned before a live demo, Crews said.

“Essentially, I had to put on the pressure and invite people to the demo. They were doing a good job briefing people, but their hardware didn’t actually work,” Crews said.

Each cadet had a specific piece to work on and incorporate into the project based on the sublet disciplines Crews had shown them earlier last semester.

Butts took on a managerial role serving as a point of contact for outside contacts, sponsors, handling schedules for presentations, and getting the resources the team members needed, among other duties. He also designed the gripper attached to the robot arm that lifts the top casing of the M795 round.

Rabb took aim at the mathematical quandaries that go into coding the robot’s movement.

During the process, Rabb encountered challenges turning all the theories he read about and studied on robotics into something tangible.

“There were challenges in learning the theory, and it was interesting seeing how it applies to this particular research and then implementing it and showing that it works and I didn’t try to take shortcuts in the code,” Rabb said. “Also, Lt. Col. Crews likes to find out our shortcuts, so there was no point in trying to take it to begin with.”

Ammo detection was Hagberg’s domain. Through computer vision and machine learning, he had to teach the robot how to locate and interface with the warhead of a round.

“Think of it as lining up your sight picture before you fire your rifle,” Hagberg said. “In this case, the first problem to solve is teaching the robot to accurately align itself with the warhead; then the second problem is moving to grab the individual round and pick it up.”

Murphy focused on programming the robot to intuit a sense of touch using torque feedback, the mathematical Jacobian method, and Denavit–Hartenberg parameters, which is a tool used in mechanical engineering during robotics development.

“I also spent my time helping Guizani with developing the robot’s vision (motion detection)and applying the sense of touch to meet the safety requirements needed to complete the project,” Murphy said.

Currently, warehouse robots work within the confines of a large yellow cage on an assembly line to avoid any minor or fatal human injuries in workspaces.

Yet, Guizani opted for a more innovative solution.

“I worked on an algorithm that would allow the robot to detect motion and gauge distances between itself relative to a human being,” Guizani said.

In addition, the robot will sense incidental physical contact and freeze immediately, preventing any possible injury to a human or damage to another piece of equipment.

The WAR team, though distraught from the critical feedback, slogged through intense work hours and sleepless nights. Weeks turned into days, and their deadline was fast approaching. The Corps of Engineers Ball would be their second attempt at making things right.

The Corps of Engineers Ball and Projects Day

Hagberg and Rabb watched the HEBI modules on the robot arm beam a red light indicating something was wrong.

“I think this is marginally worse than our presentation two days ago,” Hagberg sarcastically noted.

“No, it’s better because at least something gets to somewhere,” Rabb responded as they stood before their robot arm mounted on a mobile desk at Eisenhower Hall.

The WAR team members, Murphy, Rabb, Guizani and Hagberg, had set up their robot station and were in their full dress India whites, performing last-minute preparations for the Corps of Engineers Ball. They had 15 minutes left to organize an unfinished product before Lt. Gen. Scott Spellmon, the 55th Chief of Engineers and commanding general of the U.S. Army Corps of Engineers (USACE), came to their station.

“The general’s here. I saw him downstairs,” Crews said to the team. “When you show him this, how are you going to fill the void when something goes wrong during the live demo?”

Then, suddenly, as if on cue, the robot arm abruptly tremored as if it were going into a seizure as the WAR team attempted some last-minute tweaks. Crews continued observing.

“It’s almost as if (the modules) are pushing against each other,” he said. Crews pointed at the two modules conflicting with one another.

“They are fighting. They should always have the opposite command of one another,” Crews advised.

The War team and Crews watched the robot arm continue to tremor.

“We’ll fill the void with some stories about where it was and how proud we are that it’s gotten to this point,” Hagberg said as the floor filled with other cadets organizing their stations in full dress, ready to make an impression.

The minutes ticked away, and soon, the floor was packed with spectators, scientists, professors, engineers and cadets. Despite the robot arm’s mechanical setbacks, the WAR team coordinated the presentation in a way that kept onlookers engaged.

They piqued the people’s interest with the novel concepts of information they presented.

Every interaction kept the team on their toes and prepared them for the pivotal moment when Spellmon reached the WAR team’s station.

“In this case, you have to take what you get in the moment and work with what you got no matter how bad a situation is,” Murphy said.

He also believed the failures the team repeatedly encountered would only lead them closer to their eventual success.

“That’s the whole model of West Point,” Murphy added. “You will fail, but through hard work and determination, you can achieve success.”

The WAR team was midway into the Corps of Engineers Ball. Spellmon steadily made his way through a gaggle of people until he eventually landed on the WAR team.

Each explanation by a team member sparked another question by Spellmon. The novelty of the robotic arm, despite the malfunctions, kept the three-star general enthused.

Spellmon took the Xbox controller and attempted to control the robotic arm. Naturally, the robot did not handle well, but this did nothing to deter Spellmon’s interest in the project.

The presentation ended with Spellmon congratulating the cadets on their breakthroughs with the projects, but more improvements still needed to be made.

“So we had the Monday presentation in the lab, then we had the Corps of Engineers Ball presentation on Wednesday, and now we have Project Day coming up. It’s rehearsals on top of rehearsals,” Crew said.

The cadets put the Corps of Engineers Ball behind them as soon as it ended and immediately started toiling away at the final touches of the robotic arm.

The team had seven days left to debug and patch all the issues plaguing the project leading up to their third presentation on the 23rd annual Projects Day on April 28.

“On the 27th, everything should work,” Crews said to the WAR team. “We should carry it over to (Michie Stadium) and leave it overnight because you’re not going to have access to it anymore on the 27th. Then on the morning of the 28th, the (Secretary of the Army Christine Wormuth’s) helicopter will land, and, I’m telling you right now, your demo will be her first stop.”

It was three lessons ago leading up to Projects Day when Guizani finally hit his mark and made his most significant breakthrough. He along with the rest of the team, exhaled a sigh of relief when the robotic arm finally detected a human’s presence and came to a complete stop.

“It was the idea behind the code itself that led me to figuring it out,” Guizani said. “Once I understood the concept of what Lt. Col. Crews wanted me to do, I transformed the concept into a code. It became a process I applied piece by piece until I finally got it.”

As Projects Day kicked off early in the morning, everyone made their first stop at the WAR teamʼs section, and everything went off without a hitch.

Unfortunately, the Secretary of the Army did not show up due to a change of plans, but the show went on, and the team presented their capstone project successfully. From safety to ammo detection, all of the moving parts worked seamlessly.

“I just feel relieved,” Butts said. “I’m definitely satisfied with how far we got. We did a lot in a year. If you compare us to other counseling groups, a capstone team is usually working on one aspect of a project. We looked at the safety, movement, design and the gripping, a part from other elements.”

For Murphy, it was amazing to reach the success he had mentioned previously during the Corps of Engineers Ball.

“It feels amazing. Obviously, this was a very stressful project to work on. I’d say definitely one of the hardest, if not the hardest one in the Mechanical and Civil Engineering Department,” Murphy said. “We had a lot of long nights. We were working every day on this capstone. So it feels good that all of our hard work and dedication to this project ended up producing something that people are interested in, that might make a positive change in the Army.”

Jette was thoroughly impressed with the progress the cadets made throughout the academic year and commended them for their arduous efforts.

“They took on a number of pretty challenging problems to resolve,” Jette said. “For example, the locking hook that attaches to the round is safe, and it will not let go even if it loses power. That may seem inconsequential to some, but you’re asking cadets, undergraduates, to design and build a real device that actually works.”

Randy Doyle, the Environmental Coordinator on the IAAAP staff, worked with Crews and the cadets since the start of the project. With 39 years of military service before retiring, it was awe-inspiring for Doyle to watch the WAR team contribute to the advancement of the Army warfighter.

“I’ve been around a long time, and I just never comprehended that we in the Army have evolved to this capability,” Doyle said. “I mean, we were trained to be warfighters and now we’re training our Soldiers how to equip the warfighter.”