The FAA recently released a rule for commercial use of small unmanned systems: Part 107. But noticeably missing from it was the use of beyond line of sight. While the FAA is still trying to figure out how beyond line of sight will become a factor in the drone industry, other companies are already working on the technology. Iris Automation for instance is working on an autonomous and situational awareness system for drones so that when beyond line of sight does become reality, drones will be ready.
“We see this huge niche in the market where we need sense and avoid technology, and we need collision avoidance especially for bigger commercial drones,” said Alexander Harmsen, CEO and cofounder of Iris Automation. “This is a completely different ways of using drones and using robotics for industrial and commercial purposes. That is the real power of autonomy and automation.”
We talked to Harmsen to learn more about the autonomous drones and where the industry is going.
InterDrone: What is the importance of having drones equipped with autonomy and situational awareness?
Harmsen: Up until now we had a lot of manual flights, especially on the consumer side of things. It is a lot of drones being used for toys. Even for close-up use for inspection and work with real estate or insurance claims, these drones are definitely useful. But they require an operator to be there, and we don’t get the full use of these drones. It is still very expensive to operate the drones because people need to be there as an observer or as an operator, and so building more autonomy into drone systems allows us to become a lot more operationally efficient. It reduces the cost tremendously, but also allows for scalability in these systems.
Being able to do mining exploration or a pipeline inspection with one operator and a hundred drones is huge—that is going to change industries. Allowing a farmer to have his own drone that just takes off 5 a.m. every Monday morning and does a scan for a field, that is so much more powerful than what is currently happening where we have operators coming to the field, flying the drone for a couple of hours, charging a couple thousand bucks.
Another major piece of autonomy is just the reliability of drone systems. We need a certain level of safety there that just isn’t there right now. We still don’t trust these drones to be completely independent, and that is a huge piece that we want to help fix.
How does providing these solutions help foster a sense of safety?
On one level what we are building with the situational awareness is a safety system. We are building reliability into the drones because for the first time drones will be able to see the world. They will be able to recognize that there are trees or power lines or buildings in front of them and they will be able to determine if there are any dynamic objects in the scene as well like aircraft or helicopters. Anything that they could possibly hit, they can now finally see, whereas before they were just kind of blindly relying on the operator's control or blindly relying on GPS waypoints they are following.
This is very crucial. This is why we trust the drones to fly with the operator on the ground, because the eyes and the situational awareness is there. This is why we allow pilots to fly an aircraft, because they can see the world around them, and they are able to take that last minute avoidance.
How will the industry utilize autonomous drones?
The major use cases that we see for these drones are in these big natural resource sectors. There is so much land that needs to be covered for farming and for mining exploration, for example. There is about 2.5 million km of pipeline across North America, and a lot of it just isn’t being inspected because it is way too expensive. We could have drones doing that. It is expensive because we need helicopter pilots, fixed-wing pilots in aircraft in these remote locations. This is perfect stuff for drones to do.
When we talk about mining exploration, we have big issues with helicopter pilots dying because they are flying really low to the ground for 10 hours at a time. They get fatigued. A lot of industrial work for search and rescue or for forestry inspection is very routine, and it is in the middle of nowhere. It is very expensive to get people to those locations. If we can do that with drones, that changes this whole entire industry.
If drones can operate themselves, does that potentially mean they will replace workers?
I don’t think people necessarily lose their jobs because of this. Those current workers aren’t put into those dangerous situations anymore. Essentially the work that they do gets better and more intelligent, especially for the drone operators. There are a lot of drone operators that are bored flying one system at a time. If suddenly they can control a fleet of a hundred systems, this makes their work so much more effective; their own time goes so much further. It is a lot more about using their brains and really operating the entire fleet rather than doing the manual flights that autopilot could do that this point.
How far along are you in creating a 100% reliable system, and are you working with other industry leaders?
We started flying our system onboard with a number of different pilot customers. Our systems are starting to get into the air. It is a similar challenge with, say, the Google self-driving car or any other self-driving car technology. We will never get to 100% perfect rates. Even human pilots have a ton of issues. It is more about getting those reliability numbers up to a certain point.
What kind of software and technology is necessary for this type of system?
What we are building here is computer vision technology, deep learning and artificial intelligence that goes into being able to see the environment in the same way that you or I would see an environment.
Why do you think this is a problem you need to tackle now?
Because it is in the gray area. If we can fix [the uncertainty and autonomy], then suddenly we are able to unlock this beyond visual line of sight market, and that is where the real money is. That is where we could really use these drone systems. It is this full autonomy that the entire drone market is going to truly explode, and we are able to explode this technology. If you and I know that there is a gray area here because we are missing situational awareness and a real trustworthy system, and if we can build that trust, it makes so much sense for us to do that. That is the perfect opportunity for us to make a huge impact on this sector.
How far out do you think autonomous drones are in the future?
I think we are very close. We are starting to use drones in a lot of these use cases already; they are just relative limited because we need operators right now to fly with those drones.
Was the FAA’s release of Part 107 a step in the right direction for the drone industry and moving toward autonomous systems?
Definitely. I am very excited about these news regulations coming out. A lot of it is formalizing the 333 exemption. There is a new certificate of waiver process that really helps push forward on a number of new technologies.
Where do you hope to see the drone industry going?
I really hope this allows us to get a lot more out of the current systems that we have, the current expense that we put into these natural resources sectors, and a lot of the current time that is spent. I hope drones will help automate and really help make systems more efficient. This is the goal of these autonomous systems: to make those human hours go so much further. I don’t think we will ever get to a point where we will replace humans. A lot of these autonomous vehicles and systems are there to help make [humans] more efficient.