Connecting the world through nonparametric learning
Jeremy Henderson spent over two decades as a military pilot in the Canadian Air Force, with multiple overseas deployments and years coordinating global air force operations. During his time in the skies, he learned more than a few things about airspace and navigation systems and grew fascinated by ways to improve them. After leaving the military in 2019, Henderson consolidated his niche expertise through Stratotegic Inc., with the goal of assisting high-altitude platform station (HAPS) companies with flight and air traffic control coordination on a global scale.

Navigation is one of the biggest problems facing the HAPS aerospace industry. For instance, how do you navigate in an area you don’t know much about? How do you make decisions for flight direction around a factor as complex as the wind? And how do you account for a multitude of unpredictable circumstances in high-conflict areas or post-natural disaster? There had to be technologies that could mitigate risk, reduce the chance of human error and deploy communications to underserved areas.

Henderson visualized methods to improve on networks of autonomous high altitude balloons. He would build upon proven concepts of stratospheric communications that were limited by imprecise navigation of winds. If you can develop an AI system that comprehensively understands stratospheric winds, then you have essentially harnessed the wind. The balloons would arrange and keep themselves physically situated above an area where high-speed internet is required for a temporary amount of time, like after an earthquake that knocks out all power, or a war zone where the means of communication have been compromised. The AI algorithms could also be used to manage the position of drones or underwater devices.

Taming the wind

Temporary networks are not a new idea. Scientists began experimenting with “balloon communication” back in the early 20th century, using enormous, helium-filled balloons to carry radio transmitters into the atmosphere. Later, meteorologists used weather balloons to collect information on atmospheric pressure systems and thus predict weather patterns. More recently tech giants such as Alphabet Inc. and SpaceX have invested in balloon technology to provide internet access to remote areas across Africa, South America and Oceania.

A relatively small player like Stratotegic can win with these behemoths by solving the technology’s biggest problem: wind currents. Wind currents are part of a very complex flow system that change constantly and are very hard to predict. Climate change only increases the complexity. This summer, we observed right at home how ground telecom infrastructure can be vulnerable to forest fires, especially in the Western provinces and Northwest Territories. The latter had to resort to Canada Post for dispatching evacuation notices due to destroyed communication infrastructure.

Meanwhile, the Maritime provinces have been experiencing a marked uptick in hurricanes annually, resulting in the loss of significant ground-based communication. While satellite internet might seem like an alternative, its speed is often limited, and its capacity to connect to standard cellphones is restricted to SOS messages. Complementarily, HAPS can integrate seamlessly in the stratosphere with existing cellular infrastructure and can fill the communication void in disaster areas as well as remote regions. 

The key lies in the intelligence behind keeping the balloons and their positions synchronized despite the weather. And the tool for solving it is mathematics. 
A mathematical matrix

Stratotegic is approaching their research as an optimal control problem, essentially trying to identify functionals in real time in an efficient manner. The mathematics involved, then, are essentially the tools used for stability analysis, averaging analysis and all the complexities associated with time-bearing dynamics.
The main issue is having to rely on learning techniques that are data efficient. There are many learning techniques that require gigs and gigs and gigs of data. But for an application like this, when you have an environment that changes all the time, they have to rely on a nonparametric learning tool that learns from a functional perspective, instead of getting the data point and regressing against what we have and trying to optimize parameters all the time.
The sky is no longer the limit 

Henderson partnered with Martin Guay, a Professor of Mathematics and Engineering at Queen’s University. Guay connected him with Fields, where he learned about the emerging Multiplier program. Linking up with Fields, where Stratotegic could tap into our vast network of mathematicians and academic partners, made sense from a research perspective. For the industry side, Multiplier principals, Donna Shukaris and Martin Croteau, have helped advise on everything from securing intellectual property protection for the company’s control system to structuring relationships with international partners.

At present, Stratotegic is testing prototypes and forging international partnerships in the US and Europe. Later this fall, they are also starting a major multi-year government project that pertains to QKD quantum communications and at the same time currently undergoing an initial investment round aimed at gathering the necessary funds to take the company’s development to a commercial level. The funding effort has already garnered considerable attention from investors, making him confident that it will be a resounding success.

Having coordinated HAPS airspace and approvals with over half the countries in the world for their clients, Stratotegic is well underway to being known as one of the key experts in the industry. We expect great things from this smart start-up and believe that, in this particular case, the sky is no longer the limit.
Had it not been for the funding support from the Fields Multiplier program and Mitacs, we would have remained unaware of the vast potential inherent in this area of mathematics to bring transformative changes to the HAPS industry. Martin Guay’s engineering expertise in this field of study, combined with Donna Shukaris and Martin Croteau’s exceptional business insights and product commercialization skills, were the linchpins that propelled us to the prototype phase of our technology.

Jeremy Henderson