Photon-IV Inc. of Cambridge, Ont., was one of several newcomers when the Canadian Space Agency (CSA) posted its latest Space Technology Development Program (STDP) awards last week, winning $350,000 to develop artificial intelligence that predicts when satellites will pass overhead.
The award is for a project called “AI Predictor pLEO Networks.” According to the CSA’s project summary, the goal is to “develop an AI-powered system that can predict when Low Earth Orbit (LEO) satellites will be available to communicate with devices on the ground,” and the work could “support environmental monitoring in remote regions, including the Arctic, while strengthening Canada’s sovereign capabilities in satellite communications.”
Satellite availability problems
Photon-IV is a Canadian space and defence technology company. It focuses on improving satellite connectivity through better connection security and more reliable handover between satellite-based and terrestrial communication. It is a qualified supplier under the Government of Canada’s Defence Drone Initiative and received a $250,000 contract from National Defence’s Innovation for Defence Excellence and Security (IDEaS) program for “defence-grade satellite and terrestrial network orchestration.”
In the case of this STDP award from the CSA, the relevant technology is a tool that Photon-IV’s calls “Predict.”
In essence, the issue Photon-IV is trying to solve is the low power and processing capabilities of remotely located Internet of Things (IoT) devices that rely on satellite-based non-terrestrial network (NTN) communications.
There’s an ever-growing number of satellites that could communicate with the devices. The pLEO in the project’s title stands for proliferated low Earth orbit, meaning networks built from large numbers of satellites in LEO.
Individual devices can’t reasonably model all of those satellites’ orbits to determine when and where satellites are within range, however, so they burn scarce battery power continually trying and failing to connect with whichever satellites might be available. That reduces the IoT devices’ usable lifespan and can introduce gaps in the data that could create problems.
Photon predicting pLEO passes
Predict is Photon-IV’s attempt to resolve this problem.
In an email to SpaceQ, Photon-IV said it could not go into detail on how the system works because of its involvement in the Controlled Goods Program. Run by Public Services and Procurement Canada under the Defence Production Act, the program governs who in Canada can examine, possess or transfer goods and technical data with military or national security significance. Companies that handle them must register, which includes a security assessment.
But what the company could share was that, like so many other technologies right now, it’s built around machine learning and AI. Predict looks at all the orbits of a pLEO network’s satellites and uses a machine learning algorithm to create the best possible power scheduling and duty cycling solution for the IoT devices across an operator’s fleet. Photon-IV’s AI models take the orbital information, combine it with network information and use all of it to predict the best times to attempt the uplink, sending that data to the client’s fleet of IoT devices.
The goal, according to Photon-IV, is to dramatically extend device battery life, resolving “the critical power constraint that limits IoT-NTN deployment at scale.” It pointed to potential radio frequency (RF) energy savings of 40 per cent, leading to battery life three to five times longer.
This could also make smaller constellations and pLEO networks far more effective. Instead of brute-forcing the issue with huge omnipresent constellations such as Starlink or Amazon Leo, networks of smaller constellations can still get needed data by using this pass prediction.
Efficient use of LEO capacity
Photon-IV says the work is “intended to support more efficient use of Canadian LEO capacity.” The company has already tested its patent-pending prediction technology at the O-RAN Alliance Global PlugFest Spring 2026, an industry testing event for open radio access network (O-RAN) equipment, working with Keysight Technologies at Virginia Tech’s Commonwealth Cyber Initiative xG Testbed in Arlington, Va.
Sanal Kamal, Photon-IV’s founder and CEO, said “a remote sensor shouldn’t spend its life listening for a satellite that isn’t there,” and that the growth of LEO constellations means terrestrial devices “need to get smarter” about connecting.
Kamal said the CSA’s support will “let us take our prediction technology into the field,” where the company can test it in remote regions such as the Canadian North. “We’re proud to be building this in Canada,” he added, “for the places where Canadians need it most.”