Shinsuke Takahashi, University of Southern California; David Barnhart, University of Southern California
https://doi.org/10.64861/NHHV1883
Keywords: Infrared sensor, Non‑cooperative objects, Cislunar space, Tracking, Space-based assets
Abstract:
This paper investigates how combining visible and thermal infrared (IR) sensors can extend observation arcs and improve maneuver detection for non‑cooperative objects in cislunar space. Objects in the region have long orbital periods from several days to weeks, which are difficult to monitor from Earth due to range and geometric constraints. Thus, space‑based sensors on cislunar orbits are powerful assets that are not bound by these constraints. In many current cislunar space situational awareness (SSA) concepts, passive visible sensors are treated as the primary option for wide‑area surveillance, and most prior work on cislunar orbit determination and architecture design assumes visible‑band imaging as the main measures. However, visible sensors inherently depend on solar illumination, so observability is strongly driven by the Sun–object–observer geometry, and objects can become difficult or impossible to track during eclipses or other poor‑illumination conditions. If a maneuver occurs in these gaps, the object may be lost, and recovering it with narrow‑field, high‑resolution visible sensors after a large trajectory deviation is operationally challenging.
To address this gap, we examine the use of thermal IR sensors as a complementary option and quantify, via simulation using Systems Tool Kit, how they contribute to longer observation arcs and more robust post‑maneuver tracking for non‑cooperative targets. In many scenarios, spacecraft and other space objects are expected to have effective temperatures on the order of 300 K, so their thermal emission is observable around a wavelength of approximately 10 micrometers.
In our simulation, we assess how visible (VIS) and thermal IR sensors support triangulation‑based position determination for non‑cooperative cislunar objects. Four space‑based observers, each carrying VIS and IR sensors featuring comparable optical systems, track 30 targets on an L1 halo orbit at three epochs (Day 1, 10, 19, representing three representative dates within a given month), assuming a detection threshold of SNR ≥ 5. VIS measurements yield high‑accuracy positions when available, but some targets become unobservable on later days, resulting in no position estimate. Specifically, VIS sensors observed 18 out of 30 targets on Day 10 and 23 out of 30 on Day 19. IR sensors, in contrast, can observe almost all targets at all epochs, although with position errors roughly five to ten times larger than VIS. This is due to the diffraction limit, which results in IR sensors having lower resolution compared to VIS sensors. Additionally, at Days 10 and 19, targets that are lost when using only VIS sensors can still be successfully localized by fusing measurements from IR sensors. It is also confirmed that the position errors are improved to about 70–80% of those obtained with IR-only solutions. Moreover, we find that, while VIS sometimes provides only a single observing platform in some cases, IR delivers three or more simultaneous observers for all targets, increasing geometric redundancy and reducing the risk of losing objects due to illumination or viewing constraints. Therefore, thermal IR sensors are powerful complements for VIS sensors to extend observation arcs.
Leveraging these results, we extend this work to track for non‑cooperative objects in cislunar space, adopting a conventional residual‑based approach in which maneuvers are inferred from statistically significant deviations in orbit‑determination residuals derived from multi‑sensor tracking data. In this context, thermal IR sensors play a complementary role by maintaining track custody through illumination gaps and reducing the likelihood that post‑maneuver objects leave the sensor field of view (FOV). They are well suited to “catch” objects after maneuvers and hand off updated track information to higher‑resolution visible sensors for refined orbit determination.
We outline ongoing and planned simulation studies that evaluate how much the probability of losing objects after maneuvers can be reduced when thermal IR sensors are incorporated into cislunar SSA architectures. These studies vary sensor geometry and observation cadence to assess trade‑offs between IR capability, constellation size, and track‑loss risk. By leveraging the low dependence of thermal IR sensing on solar illumination and its suitability for wider‑field tracking, our results suggest that cislunar architectures incorporating thermal IR sensors can achieve higher‑frequency post‑maneuver track updates and significantly reduce the probability of losing non‑cooperative targets compared with visible‑only approaches.
Date of Conference: September 15-18, 2026
Track: Cislunar SDA