Autonomous Technology: Q&A with Scott Erwin, Acting Chief Scientist for the Air Force’s Space Vehicle Directorate

Since the development of the first artificial satellite in 1957 (Sputnik), the evolution of what can be done in space by unmanned technology has gripped the imagination of engineers, scientists, and space-enthusiasts alike. Technology has made great strides since those early days and the possibilities for autonomous machines in space continue to grow.

In 2020, the Hyperspace Challenge looked to expand the conversations around space-faring autonomous technology and invited startup companies and university teams to put forth solutions that could help drive innovation in several key areas.

The Hyperspace Challenge team worked closely with several government scientists to develop a series of questions for a new cohort specifically addressing improved integration of platforms and algorithms designed for space missions. The participating teams would need to propose solutions for questions including:

  • How do we achieve next-level autonomy in remote environments for servicing modules/vehicles?
  • How do we improve automatic hazard detection/avoidance to enable safer operations?
  • How do we leverage machine learning for autonomous vehicle control?
  • How do autonomous systems detect/react to system/sensor failures?
  • How do we develop the future of hardware solutions for autonomous software in harsh environments?
  • How do we improve automatic hazard detection related specifically to space weather?

Critical to the development of these guiding questions was the voice of Dr. Richard “Scott” Erwin, acting chief scientist for the Air Force’s Space Vehicle Directorate. Erwin’s career spans over three decades dedicated to furthering space technology, and during his 25 year tenure at the Air Force Research Lab (AFRL) he has been the principal or co-principal investigator for six on-orbit flight experiments. He held simultaneous appointments as the Lead for space autonomy technology for the Space Vehicle, as well as the Directorate’s Special Projects group. He is also the author or co-author of over 120 technical publications in the areas of autonomy, spacecraft communications, dynamics and controls.
Hyperspace Challenge editors spoke recently with Erwin about his experience working with the Hyperspace program, and found out about his current views on the importance of autonomous technology in the future of space.

We often hear from startups that they struggle to understand what the government actually needs when it comes to space technology solutions. That’s why programs like Hyperspace Challenge are so critical in fostering better relationships between commercial entities and government scientists. As part of our process, we invite government subject experts to the table to transparently say, “This is what we’re exploring now.” When Hyperspace Challenge approached you several years ago to help develop the 2020 areas of focus, what were your initial thoughts to the team?

Because of my career with the Air Force Research Lab, I suggested a focus on space-trusted autonomy or some combination of those three words. I have seen over the years how the field has developed, and noticed that this area was gaining renewed and broad interest in the space community.

What I said at the time was that there are really three relevant mission areas of interest, and if we could fashion problem statements around these, then we’d be likely to get quite a bit of interest. The areas were:
Autonomous space exploration vehicles like the Mars Rover or an unmanned spacecraft going to another planet;
Geospatial intelligence including satellite imagery; and
Autonomous technology to address resiliency for contested space.

That third area – autonomous systems technology in preparation for contested space – is one of the least addressed areas of space technology development today. There is significant information available on the possibility of space becoming a contested environment in the future, and we felt like the questions that could be developed would be ones that NASA, as a civil space department, does not really deal with, and could be helpful across several departments and agencies beyond the DoD and Space Force.

How should companies and researchers be thinking about autonomous technology and how it applies to space today?

When I say “autonomy” and “space,” most people immediately think of NASA rovers and autonomous robotic exploration missions like robotic unmanned spacecraft. But, there are so many more autonomous technology opportunities. The Space Force for example is interested in space mobility and logistics as well as space security and space domain awareness in low Earth orbit and beyond, it will be interesting to see how companies address each of these areas.

We’re seeing progress in areas that feel familiar to air domain questions. There are commonalities in terms of the technologies that both space and aviation employ – not at the detailed systems-level but at an abstract level including reasoning in systems that don’t have very powerful processors and on vehicles that may not have direct communications to the ground at all times and may need to make decisions about whether to stay on mission or to alter from the mission in order to avoid harm’s way.

Consider a situation where a satellite is able to make some decisions about a debris collision avoidance autonomously or other mission-altering moves under certain conditions. If you develop this capability, you’re removing some vulnerability. Concepts like that start to sound very much like the same thing an aviation person might talk about in regards to why they want to put autonomy on an unmanned aerial vehicle (UAV) or on a remote piloted vehicle.

Thinking like this just points to the fact that there are so many opportunities for growth in this field.

Are there any developments in the field that are particularly interesting or exciting to you?

All of the areas that were important in 2019 and 2020, when we were putting the Hyperspace Challenge questions together, are still important now. I think we are seeing a shift in the prioritization amongst them, but it is most exciting to see the noticeable growth of the automation and autonomy community, which isn’t necessarily the same group of people as the broader “space community.”

The changes I’m seeing are actually happening in large part because the Space Force, which was just formed in 2019, has the ability to look at how they want to do things and doesn’t have to be as tied to the ways the Air Force was perhaps going to do them before. And, there is also the fact that people who are at the intersection of commercial space and autonomy at large, are taking interesting risks that the government just wasn’t able to take for historical or political reasons, and they’re making big progress because of the freedom they have from those constraints.

The most exciting thing to me is to finally see that people who have the resources to work in space and who are also willing to take risks are allowed to do so by the system, and are in some cases even incentivized to take risks. It has been gratifying to see a lot of the people I know, who used to work in government or in academia who have started space companies, and they’re doing really well.

And, there’s a diversity of business models that are being tried right now. One example is basically allowing for service-style products. In Orbit Fab’s case, one of the 2020 Hyperspace Challenge cohort participants, the company has in mind to provide fuel so other companies working in space don’t have to build out that area on their own. This may not seem totally revolutionary in abstract, but this is new for space because to date that just hasn’t ever been an option. You brought your own logistics chain with you, or more traditionally you simply didn’t bring a logistics chain at all. When a satellite runs out of fuel on orbit now, it dies. So now we are talking about servicing, or at least elements of servicing like fuel, as a commodity facilitated by a commercial entity.

You’ve supported other accelerators in the space arena, what makes the Hyperspace Challenge unique?

I participated in an accelerator out of Los Angeles and what’s interesting about the difference for me is that the California accelerator had already picked the participating companies. My role there was to sit down and talk with the companies and help them with networking and with understanding the government side of things. That was fun, but on the Hyperspace Challenge I really got into the setup process with the team.

Also, something that is different with Hyperspace Challenge is the fact that the teams applying have great ideas based on technology coming out of other arenas, not necessarily space, and they want to figure out how to apply it to space, and some already have an idea of how to apply it to space. But, it is often the case that they don’t have an existing network within the space community. So, to be a resource for a company that isn’t well hooked up in space but they’re hooked up in something like AI and machine learning, that seems to be useful for accelerators to find where a group needs a connection and to help make that connection for them.