A powerful solar storm in May of 2024 brought with it a dazzling light show known as auroras which covered much of the northern hemisphere. While the auroras were beautiful to witness, these solar storms disrupted critical GPS signals that many automated systems use to navigate.
Across the United States, these disruptions in GPS signals caused farmers to delay planting with automated farm tractors that rely on GPS to plant extremely precise rows, losing days of productivity that threatened their planting schedules for the upcoming season.
What this event helped to highlight is that many so-called “autonomous” robotic systems today are not autonomous at all, they're highly automated. They rely on existing infrastructure systems and carefully pre-planned GPS routes. When one link in that chain fails, so does the perceived autonomy and usefulness of the technology.
When farmers across the midwest were left with un-drivable tractors because of some solar flares it was hard to point the finger only at the sun. The other culprit we need to call out is terminology.
See, for years companies selling robotic hardware that operated in a highly efficient and automated way convinced their customers that these products just work. They've R&D'd every conceivable externality and planned every if/then statement to make sure they're systems will always be up and running. But we all know that things break, or wear down. And then the impossible happens: an entire GPS network goes down. And that's when the automated system crashes.
A fully autonomous system doesn’t rely on existing infrastructure like GPS, wireless internet, or prior maps. Fully autonomous systems create their own internal 3D map of any environment on the fly and can then explore the environment with extreme precision.
But for this type of autonomy to be successful in the real world, it has to work flawlessly and be invisible to the end user. As we witnessed with solar storms, when bugs or technical issues arise with tethered automation systems, the usefulness immediately disappears and the illusion of autonomy is gone and the helpfulness evaporates along with it.
So the question is, how do we reach a level of autonomy where machines and robots can simply explore and function in the real world without any outside input or guidance?
For most people, this type of autonomy is considered something from science fiction. A perfect example is the beloved humanoid robot C-3PO from the Star Wars movie franchise.
In those movies, every other character around C-3PO has no reluctance or reservations when it comes to relying on it for a variety of tasks. In some cases, even trusting the droid in critical life-or-death situations.
In contrast, how many farmers who were stuck with non-functioning tractors during the solar storm now have complete faith in those systems?
Granted, the various droids in Star Wars are fictional, but they represent what true autonomous exploration should be striving toward. Systems that perform tasks while humans have 100% confidence in them to carry out their functions without any supervision or input.
The question for us, as engineers, is how do we set a path toward that goal and which technologies can bring us there the fastest.
Our ultimate dream in robotics is to one day be able to create machines like C-3PO that humans can rely on for even the most critical unsupervised tasks. To define that end goal, we need a system for determining the level of autonomy among different machines. With the levels defined, we can then chart a course and measure our progress toward our end goal of complete autonomy.
Currently, the only broadly accepted levels of autonomy are the SAE levels for driverless vehicles. For both engineers and customers who purchase autonomous systems, understanding these levels provides a yardstick to gauge the level of autonomy and what benefits that system can offer.
We sought to create some similar levels of autonomy for aerial robotic platforms when we unlocked volumetric exploration in our autonomy engine, ExynAI.
You can read more about how we defined these levels of aerial autonomy by downloading the whitepaper.
Currently, level 4 autonomy is the closest technology we have that simulates complete robotic autonomy like in the movies. To accomplish this feat, level 4 autonomy uses a technology known as SLAM. SLAM stands for Simultaneous Localization And Mapping and is a type of computer vision. (Lower level autonomous systems can use SLAM for navigation as well.)
SLAM combines vision and visual processing technology so that a robot can “see” its surroundings and adapt instantly. From a technical standpoint, SLAM may sound complex, but in practice, it simply mimics what a human would do when navigating a new area. A SLAM system just does it with extreme precision that’s accurate to within millimeters.
The most advanced SLAM systems use emitted laser light, which bounces back to onboard sensors. Every time the laser light reflects back, it creates a tiny data point. When these millions of points are combined, it creates a precise and feature rich 3D image of any environment.
The SLAM system continuously builds this 3D models as it moves through the environment. It’s the same way you see and understand more of an environment as you move through it.
As an example, let’s say you walk into a new house while shopping for a home. As you first walk in, you may only see the main room. But as you move forward, you see more rooms and start to build a map in your mind of how all the different rooms relate to one another and the distances between each one. You also make a mental note of obstacles, stairways, or other features.
GIF visualizing SLAM accuracy improving as it recognizes features and "closes the loop."
A SLAM-based robotic system operates the same way. This allows it to navigate any area without needing outside signals or input from a user. It operates exactly like a human who is exploring a new area. Even if the environment is constantly changing, the SLAM system simply updates its map in real-time.
SLAM technology is the bridge between simple automation like the tractors that failed during the solar storm and the amazing fully autonomous robots we know from science fiction. This is the technology driving us toward level 5 automation and it represents the true future of autonomous robotics that can keep people out of harm's way.
Explaining autonomous robotics is one thing, but witnessing the technology with your own eyes truly opens your mind to the superior reliability, cost savings, and improved safety that our Nexys level 4 autonomous system can bring to virtually any robotic platform.
Contact us today to book your free demo and learn what full autonomous exploration can do for your organization.
