Ubotica has demonstrated that its software, flying on the CogniSAT-6 CubeSat which was designed, built and operated by UK-based Open Cosmos, identified a vessel in an image and used that result to schedule a second observation without waiting for a ground station or human operator.

The first image was acquired over the Pacific in early October and according to Ubotica, algorithms onboard identified a cruise ship, calculated the next useful opportunity to revisit the location and the satellite acquired a follow-up image two days later. Whilst, arguably unsurprisingly, the vessel had gone it did represent an example of a satellite using an observation to decide what to do next.

From collecting images to choosing the next one

Earth observation has traditionally followed a longer process that starts when a satellite collects raw data, transfers that imagery via a downlink to ground systems that process it and then an operator or automated workflow decides whether to request another image. That is often the most appropriate method but it can be too slow or expensive when the question is changing rapidly.

Ubotica calls the new capability Autonomous Scheduling, which David Rijlaarsdam, Director of Space System Engineering at Ubotica explains in more detail:

“Autonomous Scheduling is about giving satellites the ability to interpret what they see and decide what to do next. Demonstrating this capability on a low-cost satellite in space, we are seeing how Orbital AI can enable satellites to respond dynamically to events on Earth, without having to wait for instructions from the ground. It is a form of tip and cue. The first observation provides a clue, then the system directs a follow-up observation towards the event, object or location that appears to matter. Potential applications include maritime surveillance, monitoring offshore infrastructure, wildfire response and volcanic activity.”

What edge computing means in orbit

Edge computing means performing computation close to where data is generated instead of sending everything to a distant central system first. In this case, the edge is the spacecraft where AI models run onboard to detect or classify features in imagery and turn a large data stream into a smaller, more immediate decision or alert.

That does not make ground processing obsolete. Users still need imagery, calibration, data fusion, quality assurance and a defensible chain of evidence but the value of edge AI is that it can decide which observations deserve urgent attention before the full workflow reaches Earth. For customers there is the prospect of better using limited satellite time, downlink capacity and analyst attention more selectively.

The important commercial test is therefore not whether a model can identify a vessel once, it is whether a service can deliver reliable confidence scores, manage false positives, work through weather and sensor constraints, respect customer tasking rules and fit into operations rooms where decisions carry real consequences.

A partnership route to operational services

CogniSAT-6 pairs Ubotica’s CogniSAT-XE2 onboard processing technology with Open Cosmos mission design and operations. The collaboration shows how specialist AI companies can reach orbit through satellite-platform partners rather than building an entire spacecraft business themselves. Ubotica and Open Cosmos are also collaborating with NASA’s Jet Propulsion Laboratory on the FAME autonomous intelligent satellite-network demonstration, which points towards a future in which several spacecraft can share the sensing task.

Ubotica is pursuing maritime security first through its Live Maritime Intelligence platform and in June, it announced an $11 million funding round led by Act Venture Capital and Greencode Ventures, with existing investor Atlantic Bridge participating. The company says the capital will support commercial rollout, but neither its latest announcement nor the funding release discloses customer contracts, pricing or independently verified performance outcomes.

Watch this space

For a coastguard, port authority, offshore-energy operator or maritime insurer, the eventual prize is an intelligence loop that can identify emerging risk, choose the right sensor and return evidence while it is still useful. For EO providers, it creates room for new products that combine spacecraft autonomy with SAR, optical, RF, AIS and existing customer data rather than selling a standalone image.

The demonstration should be treated as an encouraging proof point, not a promise of instant autonomous surveillance but it does illustrate a wider market shift. EO systems are beginning to compete on how well they convert data into timely decisions, not solely on resolution, revisit or the size of an archive.

Sources

Ubotica autonomous scheduling announcement
Ubotica funding announcement
Ubotica, Open Cosmos and NASA JPL FAME announcement