Galileo Project Observatory Class System Architecture

Brief

Fourteen authors from the Galileo Project and Harvard-Smithsonian CfA (arXiv 2506.00125, 2025, submitted to Sensors) detail the design and preliminary operational results of OCICP, which splits into an on-site Edge Computing Subsystem, responsible for real-time data acquisition, GPS-synchronized clock management, AI-based object detection and tracking, and data provenance, and an off-site Post-Processing Subsystem for commissioning validation, science operations, and system effectiveness monitoring. The sensor suite spans visible-light cameras, infrared cameras (including an all-sky array), radio spectrum analyzers, microphones, magnetometers, and a software-defined ADS-B radio. Object-level sensor fusion follows the Joint Directors of Laboratories (JDL) Data Fusion Model across five levels, translating per-modality detections into corroborated multi-sensor entity tracks. Preliminary operational results are contained in the paper's later sections (pages 13–33 were not available in the provided excerpt); the design framework is presented with sufficient detail for external replication by the open research community.

Metadata

Category

Hub & Overview

Venue

preprint (arXiv astro-ph.IM)

Type

Preprint

Year

2025

Authors

Phillip Bridgham, Alex Delacroix, Laura Domine, Andriy Fedorenko, Ezra Kelderman, Sarah Little, Abraham Loeb, Robert Lundstrom, Eric Masson, Andrew Mead, Michael W. Prior, Matthew Szenher, Foteini Vervelidou, Wesley Andres Watters

arXiv

2506.00125

Access

Open access

Length

6.0 M

Programs

Galileo Project, AARO, UAPx, Scientific Coalition for UAP Studies

Instruments

visible-light cameras, infrared cameras (all-sky array), radio spectrum analyzers, microphones, magnetometers, software-defined ADS-B radio, PTZ (Pan-Tilt-Zoom) camera

Data sources

ADS-B transponder data, GPS time synchronization

Tags

UAP-instrumentation, sensor-fusion, technosignature, edge-computing, computer-vision, data-governance, SETI

Key points

• OCICP divides into two subsystems with distinct goals: Edge Computing handles on-site real-time ingestion, object detection, sensor optimization, and data provenance; Post-Processing handles off-site commissioning validation, science workflows, and effectiveness monitoring.p.5
• The Edge Computing subsystem is built on Event-Driven Architecture (EDA), enabling asynchronous, loosely coupled processing of heterogeneous sensor streams and real-time pan-tilt-zoom targeting of priority objects.p.9
• The system implements the JDL Data Fusion Model across five levels, from raw signal/pixel fusion (Level 0) through object assessment, situation assessment, impact assessment, and human cognitive refinement (Level 5), enabling per-sensor 'objects' to be resolved into multi-sensor real-world 'entities'.p.11
• Three deployment tiers are defined: Observatory (long-term, comprehensive, full instrument suite), Portable (rapid deployment, smaller suite), and Mesh (inexpensive regional sensor network); OCICP covers only the Observatory tier.p.3
• All system clocks are GPS-synchronized and continuously monitored for drift to ensure temporal alignment across modalities for valid data fusion.p.10
• AARO receives 50–100 new UAP reports per month but cannot resolve the majority due to insufficient data quality, the central gap OCICP is explicitly designed to fill.p.3
• The primary science objective is to detect, identify, and characterize aerial phenomena 'distinctly different from both currently known natural phenomena (e.g., birds, weather, meteors, etc.) and artifacts of human technological culture (e.g. aircraft, balloons, satellites, etc.).'p.2
• All findings will be published in peer-reviewed journals with underlying data made openly available, with independent corroboration explicitly built into the governance model.p.5

Verbatim

• “Scientific investigation of Unidentified Anomalous Phenomena (UAP) is limited by poor data quality and incomplete data sets. Existing data are often fragmented, uncalibrated, and missing critical metadata.”

  p.1

• “The Galileo Project is a multifaceted scientific research program active in the search for evidence of extraterrestrial technological civilizations (ETCs). This includes the search for artifacts, remnants, or potentially active objects in space as well as for scientifically anomalous objects near Earth.”

  p.1

• “unidentified anomalous phenomena sightings themselves are not classified, it is often the sensor platform that is classified”

  p.3

• “EDA enables asynchronous processing, which allows better handling of data spikes and large volumes of requests.”

  p.10

Most interesting

• The paper uses an F-35 fighter jet photographing a bird as a concrete illustration of how UAP data becomes classified by default: the image would be classified to protect the aircraft's sensor capabilities, not because of anything anomalous about the subject.
• The JDL Data Fusion Model at the core of OCICP was originally developed for military battlefield target tracking; the Galileo Project applies it here to unclassified civilian aerial anomaly research.
• OCICP explicitly distinguishes its mission from traditional SETI: SETI targets distant electromagnetic signals from ETCs, while Galileo concentrates on scientifically anomalous objects near Earth's atmosphere, framing the two as complementary rather than competing programs.
• Decommissioning an observatory, even for a simple relocation, triggers a full restart of the Commissioning phase before Science Operations can resume, treating a moved system as an unvalidated new deployment.
• ADS-B aircraft transponder data is ingested via a local software-defined radio and classified as a local sensor rather than an external data feed, a distinction that has direct implications for data provenance and immutability guarantees.
• The commissioning phase must establish a sensor performance baseline and identify system limitations before any object can be declared a statistical outlier, a deliberate design choice to prevent false anomaly claims from uncalibrated detectors.