Miguel Nunes, Hawai‘i Space Flight Laboratory, University of Hawai‘i at Manoa; Robert Wright, Hawai‘i Institute of Geophysics and Planetology, University of Hawai‘i at Manoa; Paul Lucey, Hawai‘i Institute of Geophysics and Planetology, University of Hawai‘i at Manoa; Luke Flynn, Hawai‘i Space Flight Laboratory, University of Hawai‘i at Manoa; Mark Wood, Hawai‘i Space Flight Laboratory, University of Hawai‘i at Manoa; Eric Pilger, Hawai‘i Space Flight Laboratory, University of Hawai‘i at Manoa; Scott Ginoza, Hawai‘i Space Flight Laboratory, University of Hawai‘i at Manoa; Lance Yoneshige, Hawai‘i Space Flight Laboratory, University of Hawai‘i at Manoa; Gabriel Guerrero, Hawai‘i Space Flight Laboratory, University of Hawai‘i at Manoa
https://doi.org/10.64861/GMQY8641
Keywords: LWIR hyperspectral imaging, Active Thermal Architecture, cryogenic focal plane arrays, Space Edge Processing, edge AI in space, onboard data processing, CubeSat sensing architectures
Abstract:
The rapid expansion of resident space objects in LEO and beyond demands persistent, space-based sensing capable of extracting more than photometric signatures. Thermal hyperspectral imaging in the long-wave infrared (LWIR) offers a fundamentally new modality for Space Domain Awareness: material discrimination, thermal state estimation, and potential chemistry detection independent of solar illumination conditions. In 2022, we introduced the Hyperspectral Thermal Imager (HyTI), a 6U CubeSat demonstrating 25-channel LWIR spectroscopy (8–10.7 µm) using a no-moving-parts Fabry–Pérot interferometer and a cryogenically cooled Barrier Infrared Detector (BIRD) focal plane array. The project established TRL 6/7 maturity and demonstrated real-time L0–L1 onboard processing using a heterogeneous CPU/GPU/FPGA architecture, achieving 90%) necessary for space-based remote sensing. The new HyTI-2 payload, although larger than HyTI-1, preserves a CubeSat-class compatibility while enabling improved radiometric stability. HyTI-2 is built around the Active Thermal Architecture (ATA), developed at the Utah State University and Orion Space Solutions, in which thermal control and stabilization become the primary performance driver. The ATA enables high radiometric stability, reduced thermal drift, and high-duty-cycle operation (>90%), all of which are essential for operations and extended target-tracking capabilities. Also, the upgraded cryocooler system incorporates advanced control electronics, active ripple filtering, closed-loop thermal regulation, and improved structural-thermal isolation to minimize the coupling of exported vibration and electrical noise into the spacecraft bus. Coupled to ATA is a next-generation onboard compute layer leveraging Blue Marble Communications’ Space Edge Processing (SEP) platform. SEP provides radiation-tolerant, high-throughput data processing and real-time payload data management within SmallSat power and mass constraints. By integrating the SEP into the HyTI-2 architecture, interferogram reconstruction, spectral calibration, anomaly detection, and feature extraction are executed onboard the satellite. The SEP incorporates next-generation heterogeneous edge processing and containerized workflows, enabling near-real-time interferogram reconstruction and calibration, radiometric calibration and spectral cube generation, high-resolution spectral detection. This approach minimizes downlink burden by transmitting only L2-derived feature products or flagged detections, reducing latency from days/hours to potentially a few minutes. Unlike ground-based optical telescopes, LWIR hyperspectral sensing operates both in eclipse and sunlit conditions and enables material discrimination through spectral emissivity signatures. In particular, narrow-band LWIR measurements allow identification of composite materials, coatings, and surface degradation effects that are indistinguishable in broadband photometry. As one potential application, temporal hyperspectral observations will enable detection of active propulsion through plume thermal features and evolving emissivity patterns. With the increasing deployment of megaconstellations, the SSA community requires other sensing capabilities beyond reflected visible light. HyTI-2 demonstrates that cryogenically cooled LWIR hyperspectral imaging, combined with onboard AI acceleration, is becoming viable within SmallSat-class SWaP (Size, Weight, and Power) constraints and scalable to constellation architectures. Looking forward, HyTI-2 can potentially establish the technical foundation for hyperspectral LWIR constellations operating as distributed thermal-material sensing networks. By combining the Active Thermal Architecture for sustained temperature stability with commercial Space Edge Processing for autonomous onboard analytics, future constellations could deploy multiple LWIR hyperspectral nodes across orbital regimes to provide persistent, low-latency characterization of resident space objects. Cooperative survey and cross-cue architectures can enable distributed detection, triangulated plume analysis, and temporal monitoring of material degradation or anomalous thermal behavior. Feature-level data fusion across nodes can dramatically reduce downlink burden while improving classification confidence. Such constellations can extend beyond detection to enable continuous behavioral characterization of space assets, thereby adding a new layer of multimodal Space Domain Awareness in which hyperspectral emissive signatures become as operationally relevant alongside visible photometry and radar cross-section. HyTI-2 represents a step toward that distributed, intelligent sensing infrastructure for future space security and traffic management architectures.
Date of Conference: September 15-18, 2026
Track: Space-Based Assets