2 April 2019
Opportunities in the autonomous drone and UAV industry
by Wojciech Gryc
While people focus aggressively on self-driving cars, the opportunity of self-flying planes, drones, and other aerial vehicles is cannot be understated. In some ways, the problem might be easier to solve, ushers in a more interesting future, and presents a range of opportunities for startups – from racing leagues to data collection to actual building of entire devices.
Why autonomous UAVs?
Autonomous drones and UAVs seemingly present a bigger challenge than self-driving cars: you’re now optimizing for faster speeds and three dimensions of navigation, rather than two. On the bright side, there are fewer physical hurdles: autonomous UAVs would need to manage themselves in three dimensions, but also don’t have the risk of running into as many people or interpreting street signs and other traffic signals.
Part of the romance of autonomous UAVs is also the enablement of more efficient travel. Yes, I’ve watched The Fifth Element and Bladerunner, and the idea of having flying cars that drive me between destinations (and avoid traffic or the need for roads!) sounds like a fantastic future to be building towards.
So why not autonomous UAVs, ones that can ferry people around and ship cargo in straight lines between major destinations? Today, the issues are a combination of technological capability gaps and regulatory burdens:
- Technology still needs to develop to enable more powerful motors, longer-lasting batteries, and safer operation. I won’t be breaking down all the technical challenges in this industry here, but suffice it to say, there are quite a few.
- Regulation needs to change to enable autonomous flight. The regulatory landscape in the autonomous drone and UAV space is nascent. We need to learn to distinguish between consumer/hobbyist drones versus commercial UAVs, and there is a major difference between line-of-sight operations (i.e., where an operator or company can see the UAV and interrupt its operations) versus beyond-line-of-sight operations, where the UAV is fully autonomous and its operations cannot be interrupted easily, or can only be interrupted via a radio or cell phone link.
The issues above need to be addressed to really enable fully autonomous, long-distance flight. It’s a big vision, but fortunately not everything above needs to be achieved to enable the launch of a successful company in this space.
To get our imaginations going, here are some examples of companies building UAVs that are autonomous and/or represent a precursor to the unmanned long-distance flight vision above. What’s magical about these companies is that they are working within the confines of the burdens mentioned earlier and are doing so with great early success.
First, let’s begin with technologies that could enable you to sidestep the technical issues mentioned earlier.
- Tethered drones. Given the challenge of short battery lifetimes, one option is to tether drones and pass electricity via cable to the motor itself. This enables the drone to fly more permanently, for a much longer period. Applications are clearly limited due to the cable being present, but there are still problems one needs to solve. Apparently, this market is growing at 70% per year, with notable companies like Elistari and Aria Insights. Some examples of solutions provided via tethered drones include:
- Short-term mobile surveillance for events lasting one or two days.
- Pop-up structures (lighting for events, or temporary advertisements).
- Continuous data collection of a changing environment. For example, using DJI’s Terra in a tethered environment would enable you to collect a continuous stream of data for a long period. This could be useful for environmental monitoring – for example, collecting data around a wildfire.
- Disposable drones. If you’re delivering military or cargo payloads, you might want to use an autonomous drone that steers the cargo to a landing site and is then discarded. The main use case here is precise delivery of cargo without actually landing a plane itself, with such systems already being used by the military.
- Loyal wingmen. Most drones today require oversight from a human, but what if you simply follow pilots and their (human-piloted) planes? That’s the idea behind “loyal wingmen”. This idea could be used across a number of industries. Following (surveilling, guiding, etc.) civilian aircraft or following other entities for photography, entertainment, etc. could be useful. The technology can also be used with other drones. Imagine autonomous planes following a plane that is operated by humans, for example to deliver additional cargo or transport other passengers.
Creative companies and service providers
Those are some example of interesting drone technologies, and I’m specifically avoiding examples where companies provide a flight service with human oversight. In all of the cases above, however, you’d be producing the entire drone technology yourself. You can also use a strategy to provide other solutions to the drone space without running an actual drone yourself. Specific examples are below…
- Computer vision for drones, like Iris Automation. Managing computer vision for drones and providing the optical and computer vision technology to do so. This is for collision avoidance but could be used for other applications too.
- LIDAR (or other data collection technologies). As with the above, this could be used for collision avoidance (i.e., measure if another drone is nearby) or for data collection (e.g., scanning the area and building a 3D map of it). This is a fast-changing space and LIDAR companies are very popular with investors now.
- Training data and simulation software. One of the biggest challenges with autonomous flight (indeed, autonomous agents of any sort), is the generation of actual training data to enable the drone or UAV to actually know about its environment and learn from it. While you ultimately want to test your agents in the physical world, early models and training could be done in simulated virtual environments. Microsoft’s AirSim, Baidu’s Apollo, and even Project Chrono are examples of software products that exist to fill this need.
- Drone racing. Another interesting application is the environment and community around drones. The Drone Racing League is a good example of this, and their recent launch of the Artificial Intelligence Racing League (AIRL) fits within this area quite nicely.
- Performing arts. Verge Aero runs performances with drones – by flying multiple drones in a synchronized fashion, you get beautiful light shows.
- Government work and support. Working with regulatory agencies to understand the implications of drones – via consulting, lobbying, etc. – could be an interesting opportunity especially since regulatory bodies have struggled with how to manage autonomous drones, planes, and other devices. You have examples like Singapore creating mandated spaces for drone testing, while other agencies and bodies have struggled with how to regulate or oversee drones. The struggles with unsanctioned drones around Gatwick Airport is an apt example.
This article’s goal was meant to encourage creativity in how to start companies within the autonomous drone and UAV space. While the most common vision is simply building drones that fly on their own, there is a great deal of nuance in how one can overcome technical hurdles or even provide products and services without building a fully autonomous robot yourself.
Top photo by Aaron Burden.