5 Connections Between Space Force and Autonomous Vehicles, Explained by a Veteran
John Seminatore’s career reads like the background of a character from a sci-fi movie.
A lead technical program manager for autonomy at Argo AI, the global self-driving technology company headquartered in Pittsburgh, Seminatore’s job involves conceiving, coordinating and managing the work behind-the-scenes that allows Argo’s cars to make safe driving decisions on their own and ultimately, transport people and goods safely and comfortably around cities without human guidance.
Even before the New York State native joined Argo in early 2018, Seminatore held fascinating jobs, helping design autonomous floor cleaning robots for airports at Carnegie Robotics, working at Virginia Tech on humanoid robots to compete in the Defense Advanced Research Projects Agency (DARPA)’s competitions, and from 2004 to 2009, serving in the United States Air Force’s Space Command, the division focused on out-of-this-world technology that would later be spun out as its own branch, now called Space Force.
It’s Seminatore’s work on technology at Space Force that we here at Ground Truth are most interested in learning more about. And though designing military systems and satellites would seem in some ways the exact opposite of developing self-driving cars for roads on Earth, according to Seminatore, there are far more connections between the work, and the organizations behind them, than you might expect.
- Celebrating and nurturing STEM skills
Long before he joined the military or Argo, Seminatore was interested in pursuing a career in the fields of science, math, technology and engineering (STEM).
“There are old home videos of me playing with a LEGO Space Shuttle,” as a child, he recalls.
But Seminatore did not immediately pursue an opportunity with the military. He did not see himself fitting into the military stereotype that some outsiders associate with the Armed Forces.
He formalized his interest in STEM through his education, receiving a Bachelor of Science in Aerospace Engineering from the Massachusetts Institute of Technology (MIT) in 2003.
After he graduated from MIT, he found that job opportunities for aerospace engineers were not plentiful in the post-9/11 time period. “People forget it, but back then, there were lots of layoffs in aviation,” he says. “It was really hard for someone fresh out of college to get a job.”
Instead, he worked as a freelance stagehand, handling audio-visual equipment and stage setup for clients including the Boston Ballet.
At the same time, Seminatore looked around and realized that he had several friends and MIT professors who had applied their skills to the Air Force. And as he did more research and had more conversations with veterans, he came to learn that he could pursue a role as an officer through the Air Force’s Officer Training School located in Montgomery, Alabama.
After he passed training and was stationed at Los Angeles Air Force Base, Seminatore found that the skills and interests he’d developed from his childhood were not only sought after, but celebrated and nurtured by fellow STEM experts.
“I think of the Air Force as the most high-tech of the [Armed] Services, and Space Force is the most high-tech section of the Air Force,” Seminatore says. “There’s a much higher concentration of advanced degrees, engineers and scientists, than you have in other parts of the military.”
At Space Force, Seminatore and his fellow servicemembers worked on developing portfolios of advanced new technology to aid wider military and strategic goals of the United States government, including classified technology that has only come to fruition more recently, such as hypersonic flight and the Air Force X-37B Orbital Test Vehicle, a kind of military space shuttle that can orbit for hundreds of days, nearly two full years, at at a time. Additionally, he worked on satellites that provide information and communications to ground personnel.
In the Air Force and now while helping develop autonomous vehicle systems at Argo, Seminatore’s strong STEM skills and familiarity with STEM concepts are crucial.
But STEM is celebrated in more than just the day-to-day work at Argo and the Air Force. Both organizations frequently partner with universities and academic institutions, as well as other tech-focused special interest groups.
Argo, for example, established the Center for Autonomous Vehicle Research at Carnegie Mellon University (CMU) where advanced autonomous vehicle technology research is conducted, and the company has principal scientists on staff from CMU, Georgia Tech, and the Technical University of Munich, to help work on fundamental research and technology issues that can advance the entire field of autonomous transportation. These universities look to Argo as an excellent resource for graduates seeking employment, and Argo, in turn, relies upon them and many others as pipelines for new talent.
The same is true of the Air Force, which routinely looks to the academic community for new recruits as well as to work on new military-funded research that it can use.
“A lot of the developments that I worked on at the Air Force had numerous ties to research universities,” Seminatore says.
Meanwhile, there’s also the extracurricular grassroots support for STEM. At Argo, numerous employees lend their time to support and mentor students pursuing STEM through avenues like the FIRST Robotics Challenge.
Altogether, at the Air Force and now at Argo, Seminatore sees his STEM skills being used in ways he never could have imagined when he first started out as an audiovisual technician.
- Teamwork, logistics, and leadership are critically important
The Space Force and Argo both fundamentally rely on teamwork to accomplish their organizational goals.
One of the most important aspects of establishing good teamwork is mastering something found in organizations of all types and sizes around the world: logistics.
For a military satellite, Seminatore says, “a lot of people need to talk to the satellite at the same time, so you need to sort out the scheduling of who gets time on the [ground] dish to send their communications to it, and who gets priority.”
Seminatore sees a direct parallel between juggling satellite communications requests, and how Argo must juggle the various needs and demands placed on its autonomous cars, such as “mapping missions, data collection missions, and taking customers,” in ride hail or in goods delivery. “All that workflow is very very similar,” to the military, Seminatore says.
The logistics skills Seminatore learned from his time in U.S. Space Force have translated directly into his work for Argo.
“I build [computer] models that say, ‘if you give me this many cars, operating this many hours a day, I can take this many images, it will take this long to label them, here’s how long it will take to build a model on a new generation of car,’” he explains.
While both the Air Force Space Force and Argo have many experienced, well trained and talented technical experts, at the end of the day, someone needs to be responsible for leading the organization and harnessing all the talents of the other individuals toward a common goal. In the military, leadership is extremely hierarchical and comes from commanding officers.
At Argo, it’s team leaders and Lead Technical Program Managers such as Seminatore. In his day-to-day at Argo, he draws upon the leadership lessons he learned at the Air Force.
“People don’t think of leadership as something that can be taught, but the military has been doing it for 300 years,” Seminatore says. “I was an Officer, but anyone enlisted gets taught how to lead. That is the biggest strength that military vets have, they’ve been taught how to lead,” in part because the job is so high-stakes and involves situations that can risk individuals’ health and lives, to say nothing of the expensive equipment and technology involved.
And while the leaders at both organizations must have enough technical expertise to know what their personnel are working on and the obstacles they are facing, a big part of their job in both cases is keeping projects on track.
“You need to make sure teams don’t get too hyperfocused on problems and are working with other teams,” Seminatore says. “Technical Program Managers (TPM) look much broader than individual problems. The ‘T’ in TPM means you have the technical know-how to engage and solve the problem, even if you can’t write software.”
- Transformative technology
Both the Air Force and Argo work on technology that provides clear benefits to society.
For example, Seminatore points to the fact that much of the U.S. population now relies on the Global Positioning System (GPS) satellites operated first by Space Force, then Space Force, for navigation on their smartphones and other location services and map apps. Among those who utilize GPS technology is Argo and its autonomous cars, as one of the data inputs that help determine a vehicle’s location.
It’s not something people outside of the military often think about. Even for those inside the Armed Forces, it can be hard to remember who is behind the world-changing technology we take for granted.
A lot of “people in the military don’t even know what they’re getting is coming from space,” Seminatore says.
After suffering a snowboarding injury while serving in the Air Force Reserves, Seminatore ended up getting treated at Walter Reed Medical Center. There, he saw many other veterans with serious injuries, and was impressed by how technology helped them heal, recover, and regain some of the mobility they lost.
He remembers going to a local Best Buy near the hospital and “there was a young guy, my age, a double amputee, but he was able to walk around,” thanks to robotic prosthetics, Seminatore recalls.
Seminatore is hopeful that Argo’s autonomous vehicles can benefit people in need. He mentions his 96-year-old father, who suffered an injury in recent years that prevented him from driving, and had to wait for long time periods to get ride-hail drivers to pick him up.
Seminatore believes autonomous vehicles could provide such persons more transportation options by adding to the available number of cars in high-demand areas.
- Designing durable hardware that works without people
Autonomous cars and U.S. military satellites need to be able to operate for long periods without any human intervention or assistance, and in tough environments where they are subject to lots of recurring wear and tear.
“I describe satellites as: imagine you are going to build a computer, you seal it in a room, you brick the door shut and just have a tiny little cable that comes out the end, and then it has to run for eight or 10 years,” Seminatore says. “They have to be able to operate without you being in direct communication at all times.”
At Argo and the Air Force, the technology is subject to extensive testing before and even after deployment.
“Testing of the environment is actually kind of similar,” between satellites and autonomous vehicles, Seminatore explains. “You’ve got to vibrate your tech, put it through temperature cycles, do all that stuff.”
Self-driving car hardware doesn’t go into outer space, but it must be able to withstand heat, sunlight and UV radiation, rain, humidity, dust, dirt, debris, mud splatters, bug splatters and animal droppings, cold, frost, ice and snow. Its computers need to be able to withstand bumps, vibrations of the road, acceleration and rapid braking, and sharp turns.
In terms of the software, Argo engineers put updates to the company’s self-driving software through a detailed computer simulation of a real-world driving environment — say, a map of Hamburg, Germany. Using such a simulation, engineers can introduce specific or randomized variables, including other virtual vehicles, pedestrians, bicyclists, and various obstacles, and evaluate how Argo’s self-driving system updates perform when driving around them.
Once Argo engineers determine that the updated self-driving software performs well in simulation, they will test it with select cars driving autonomously, under supervision of Test Specialists riding inside the vehicles, on closed courses at Argo’s facilities in Munich, Germany, and near Pittsburgh, Pennsylvania. If the vehicles perform safely on closed courses, only then is the software deployed on public roads under the supervision of Test Specialists riding inside.
Once a satellite launches into space, if all goes well, it begins orbiting the Earth at thousands of miles an hour. Having astronauts service it by hand in space would be prohibitively expensive and dangerous. “If a part breaks, you can’t just send someone up to fix it,” Seminatore puts it. And commands sent over wireless signals may not always work, either, due to issues with weather on Earth and solar activity.
Similarly, computers inside autonomous vehicles need to be able to operate on roads without any human intervention or support, even when traveling through wireless “dead zones” – areas of poor reception and slow data transfer speeds – or during times when the network goes down.
Each one of Argo’s self-driving cars is equipped with computer systems that contain the entire self-driving software suite needed to operate safely on the roads, including an updated, high-resolution 3D map of the area in which it is designed to drive (known as its operational design domain, or ODD) complete with richly detailed data like landmarks, traffic signals and lane markings.
Inside each vehicle are two separate computers: an Autonomous Vehicle System (AVS) and a Complimentary Autonomous Vehicle System (CAVS), each with its own set of software, circuits and wires, so that in the event of equipment failure or a collision, the car will be able to stop safely.
“In an autonomous car, if something breaks while someone is in it, first off, it has to stop safely,” Seminatore notes. “And secondly, it has to pull over. You can’t just call AAA.”
- Handling vast amounts of data
“There’s a tremendous amount of data that comes off these things,” says Seminatore, and that’s true of both Air Force satellites and autonomous vehicles.
In the case of military satellites, the data may be anything from classified photographic imagery of the ground to infrared imagery showing the heat signatures of people or bases, measurements like topography, weather monitoring, and forecasting, climate research, wireless surveillance, or something else entirely.
For autonomous cars, the data that is used to help the car navigate comes from its sensors, including video and still imagery of public road signs, road conditions, other vehicles, pedestrians, cyclists, scooter riders, and any other road users
But Argo’s autonomous cars also use lidar to create 3D imagery of the car’s surroundings and measure the distance between the car and other objects, as well as their movements, which helps the car forecast where they are going to go.
When it comes to extracting all this data from either craft, “It’s the same problem,” Seminatore continues. “You’ve got to do it efficiently, in a timely way, you can’t lose it.”
For satellites, it used to be done manually: Seminatore recalls seeing an old 1960s-era satellite the size of a city bus at Space Force that used to drop film canisters of photography from orbit which would parachute down to military planes for collection.
Now, on both cars and autonomous vehicles, data transfer happens over communications links, wireless or wired, depending on the requirements of the mission.
For Argo, as long as there is a stable wireless connection, the cars are constantly sending some data back to the company’s servers while out on the roads, including their position calculated by GPS and by their own onboard camera imagery. There’s also Argo’s Remote Guidance service, which can be activated on-demand and exchanges data between the car and the remote operators in the depot to help an Argo autonomous vehicle navigate on its own around unique obstacles, such as temporary construction or street closures.
But at the end of every self-driving car’s shift on public roads, it goes back to an Argo depot where it is hooked up to Argo servers and its ride data is downloaded.
While Argo and the Air Force both collect and move vast amounts of data, the real difficulty lies in sorting through it and interpreting it, finding the most useful and important information in a timely fashion – deriving signal from the noise, so to speak.
“It comes down to storing it, making it accessible and secure, and making it easy to do,” Seminatore says. “When you have huge volumes of data, the intelligence is in deciding what to ignore.”
Both organizations use data to monitor the performance of their technology and develop ideas for how to improve it going forward.
So even though they differ in their goals and the kinds of technology projects they ultimately work on, Seminatore’s career trajectory from the Air Force to Argo shows how there are some surprising similarities between the two organizations and their work, and how the same skills that someone develops and sharpens in the military can be a great fit for helping to build and deploy Argo’s self-driving technology.