Detectability of Potentially Colliding Space Objects via Star Trackers on at-Risk Satellites

The proliferation of objects in Low-Earth Orbit (LEO) is forcing satellites to increasingly maneuver out of potential collisions, consuming fuel and interrupting operations. Most of these burns are unnecessary, driven by conservative decisions based on limited trajectory information. Space-based observation capabilities could improve tracking accuracy, but typical onboard sensors have limited performance, making it challenging to detect faint, fast-moving objects. This paper investigates the capabilities of typical onboard electro-optical sensors to detect objects in LEO that pose a collision risk to the host platform, providing a feasibility assessment for an autonomous avoidance system. The concept is simulated in a Matlab–System Tool Kit environment over hundreds of scenarios, modeling imaging systems of varying quality, with a focus on star trackers. An observation strategy to maximize target SNR is presented. Realistic synthetic images of the scenes are generated and post-processed to derive target SNR and detection probability for each sensor. The study also assesses potential accuracy improvements from processing these observations through orbit determination (OD). Low-quality or small-aperture (2–5 cm) star trackers can only detect large-class secondary objects (>1–2 m), especially with small relative angular rates. More advanced systems achieve consistent detection and can occasionally detect objects as small as 40 cm reliably. Readout noise stands as a primary limitation for the SNR. Concerning observation timing, two favorable windows per orbit are available in about half the cases, located at opposite orbital points before closest approach. This feature efficiently constrains the orbit solution space during OD processing.