Hunting Dark Ships in the Baltic Sea

Welcome to this month’s Lightbulb Moment, where our Head of Systems Engineering, Leonardo Ghizoni, shares the engineering insights and lessons learned as we push the boundaries of satellite system design. This month, Leonardo draws from his recent feasibility study on monitoring the Baltic Sea – a region of intense commercial traffic and immense strategic importance.
 

The Challenge of Maritime Domain Awareness

Securing maritime borders and protecting critical subsea infrastructure is increasingly complex. One of the most significant hurdles is the presence of "dark ships"— vessels involved in illicit activities that intentionally disable or manipulate their Automatic Identification System (AIS) transponders to avoid detection.

When designing a space-based surveillance model to track these vessels, we faced a fundamental sensor paradox. Radio frequency (RF) sensing can monitor vast areas and detect a ship's electromagnetic emissions even when AIS is disabled, but it provides limited visual identification capabilities. Conversely, optical payloads offer the high-resolution imagery needed for positive classification, but they are constrained by narrow observation windows.

 

The Lightbulb Moment: Cooperative Two-Layer Architecture

Our "lightbulb moment" was realizing that the most effective maritime surveillance architecture cannot rely on a single sensor type. Instead, we developed a cooperative two-layer architecture using our Apex Predator satellites: "The Tip" (BC Lynx) and "The Cue" (BC Leopard). 

In this architecture, Lynx operates as the wide-area survey satellite, constantly scanning the region to detect suspicious RF signals. Once Lynx detects an anomaly, it generates a cueing request to Leopard, which is then tasked to observe the location with its optical payload to confirm the scenario. 

However, identifying this workflow highlighted a critical physics problem: ships move. A standard container ship can travel at speeds of 16 to 22 knots (up to 40.74 km/h). If there is a delay between the initial RF detection and the optical confirmation, the dark vessel could travel several kilometers, potentially exiting the optical satellite's Field-Of-Regard. 

To solve this, we engineered a concept called "Target Information Maintenance". The true power of the cooperative two-layer architecture is that while Leopard is waiting for its next imaging opportunity, the Lynx constellation continues to revisit the target, constantly updating the estimated position and computing a trajectory based on successive observations. This progressively refined data and target update – what we call “Target Information Maintenance – drastically reduces the temporal uncertainty before the optical image is acquired, enabling Leopard to point its camera toward the expected target location at the time of imaging. 

To minimize mission cost while enabling early operational capability, the architecture foresees the deployment of Lynx–Leopard pairs into the same orbital plane. This provides each deployed plane with both RF detection and optical imaging capabilities from the outset, allowing the cooperative surveillance concept to become operational without requiring the full constellation to be deployed. Subsequent Lynx–Leopard pairs can then be added incrementally, progressively improving mission performance through reduced revisit and reaction times and increased overall coverage.

Proving the Concept: Simulation Results

To validate this approach, we ran extensive orbital coverage simulations over the Baltic Sea using 550 km Sun-Synchronous Orbits (SSO). Our findings proved the immense capability of the Apex Predator series:

• A constellation of 16 Lynx satellites distributed across 8 equally spaced planes can refresh the full Baltic Sea region in approximately 39 minutes, with revisits every 20 minutes. By scaling the Lynx constellation to 32 satellites in the same equally spaced 8 planes, the regional coverage completion time drops to just 18 minutes, with revisits every 10 minutes. 

• A constellation of 16 Leopard satellites provides reaction times, or the time between RF detection and optical confirmation, of approximately 15 minutes. A constellation of 32 Leopard satellites reduces the reaction times down to 7 minutes. 

• Launching Lynx and Leopard pairs together into the same orbital plane avoids separate insertion campaigns, minimizing costs while providing immediate operational readiness.

The Takeaway: 
By seamlessly combining persistent RF surveillance with rapid optical confirmation, Space Inventor’s "Tip and Cue" architecture offers defense and commercial operators a highly scalable, dedicated path toward sovereign maritime domain awareness.