By Jacob Davisson, Head of Engineering
You may have noticed a bit of a delay in publishing this installment of First Things First — things have been particularly busy around the office, as we demo and ship our new Aertos™ 130IR and Pilot Controller System! Part of the recent delay in shipping this new product has been completing a new set of ducts that provide more thrust and flight time than our previous products. In the last blog post, I discussed my favorite piece of our patented technology — the Folded Geometry Framework™ (FGF™). In the next few of these blog posts, I’ll be discussing some of the incredible things that this framework allows us to do.
The Folded Geometry Framework™ has enabled the technology that allows us to fly stably in the most demanding, GPS-denied environments, be “instant on,” to be stable where others can’t, and to enable a first-principles design approach that enables unimaginable autonomy. One element of our framework is a patented, advanced AI technique that we call “Empathy Filtering™.”
Our first-principles design approach to flight control systems (and our other undertakings) led us to a number of initial pre-conditions. One in particular was that GPS would not be reliably available anywhere the initial generation of the Aertos™ family of drones would fly. GPS, and other external references like magnetic bearing (which would be unavailable or unreliable under a bridge, for example), are ‘cheats’ that allow other drones to correct the effects of uncertainty in their internal references — which give them stable flight where those sensors can gather data on those external references. With no data sources from external references, there will be no stability for drones that depend on them.
Empathy Filtering™ is one of the ways that allows the Aertos™ product family to manage sensor noise and uncertainty in ways that are otherwise difficult or impossible. By applying particular features of FGF™ in this AI technique, accumulated sensor noise and uncertainty do not overwhelm computation of important quantities — notably the gravity vector. Having a solid handle on which direction the gravity vector is pointing allows for a stable drone attitude everywhere you could think to fly an Aertos™ drone. To this GPS-denied (and other external reference denied) base of stability, we add various AI techniques (Empathy Filtering™) to further refine the ability of the Aertos™ to remain stable wherever you choose to fly.
This addition of Empathy Filtering™ AI technology allows the Aertos™ to learn about your flight and flight conditions. These include “feeling” the difference between the wind, a wall, a ceiling, or even a person. Our applications require that we fly into tight spaces where we have to “sense” and respond accordingly. Our products are primarily ducted and designed for stability even during such contacts. By paying attention to its response to pilot commands, among many other things, the Aertos™ is effectively participating in the flight with you.
Digital Aerolus is famous for flying stably, indoors, where GPS and other external references are unavailable. We continue to develop and refine the various mathematical and AI techniques that allow the Aertos™ to remain the number one choice for flying in the most demanding and challenging environments. We are excited to see what new environments you take us into.