Presented by:
Francis Chun, Professor, USAF Academy
Benjamin Roth, Director of Astronomical Research, USAF Academy
Timothy Giblin, Senior Scientist, i2 Strategic Services LLC
David Strong, Director, Strong EO Imaging, Inc.
Anil Chaudhary, Principal Scientist, Applied Optimization
Phillip Fishbein, Computer Scientist and Mathematician, Applied Optimization

This short course will present how traditional and advanced optical data is collected, accessed, and used by academia, industry, and government for basic research on Space Domain Awareness (SDA). There will be four sessions, which describe methods for data collection, image processing, and analysis of signatures collected using the USAFA Falcon Telescope Network (FTN) sites.

The course will cover the use of different types of filters, source extraction, and calibration of panchromatic and multi-spectral photometry, spectroscopy, and polarimetry data, and will include analysis of unique features of different types of data. The non-resolved source extraction will include point spread function (PSF) fitting, and aperture photometry. The image processing techniques will include the extraction of visual magnitude from different types of photometry data, and Stokes parameters from polarimetry. The calibration methods will include the basics of noise sources and their reduction using noise calibration frames. The course will use FTN data, and describe methods to support non-resolved object characterization and identification. Finally, the course will describe the use of a data standard to facilitate the reporting and exchange of optical data on the Unified Data Library (UDL).

Presented by:
Thomas Schildknecht, Professor Emeritus in Astronomy University of Bern, Switzerland

The proliferation of space debris and the increased probability of collisions and interference raise concerns about the long-term sustainability of space activities, particularly in the low-Earth orbit and geostationary orbit environments. During recent years governments, space agencies and civilian research organizations increased their efforts to build space object catalogues and to investigate the space debris population in different orbit regions. Understanding the nature and the sources of debris is a prerequisite to provide the scientific foundation for a sustainable use of near-Earth space.

This course will provide a general introduction to the space debris problem, give an overview on the current space debris research activities to detect and characterize space debris, followed by a presentation of the efforts to model the future space debris population and the international efforts to protect and remediate the space environment. Particular focus will be put on optical techniques to detect, track and characterize space objects including small-size debris. The techniques will be illustrated with examples from the long-standing optical observation programs.

Presented by:
Peter Zimmer, Astronomer, COMSPOC/J.T. McGraw and Associates, LLC

This course provides those new to the space situational awareness (SSA) and space domain awareness (SDA) community (as well as those seeking a refresher) with an introductory-level understanding of the tools and techniques used for detecting and tracking earth-orbiting satellites with optical instruments, focusing primarily on ground-based systems but with significant applicability to space optics. The course begins with an overview of optical telescopes and includes a discussion of many of the key terms and buzzwords one might encounter when reading about ground-based optical telescopes. From there, the course presents an overview of how these components are assembled into a sensor package for nighttime optical SSA and can be optimized to suit various mission goals. This includes a discussion of satellite visual magnitudes, terminator viewing, sensitivity, search rate and related topics. Finally, the course presents a brief look at the challenges and differences of optical systems for daytime optical and cislunar SSA.

This course brings value to those AMOS attendees who come from other technical fields or those who have a very limited knowledge of the tools and techniques of ground-based optical SSA, such as program managers or those whose SDA experience is with other techniques, such as radar. While attending the conference, lots of technical jargon and buzzwords are thrown around as if everyone understands the field, but for those new to the overall topic, understanding what people are talking about, or simply following a presentation can be a challenge. This course is intended to serve primarily as an introduction for the novice and the newcomer, but everyone can use a refresher now and then: if you already know all this stuff, think of how smart you’ll feel!

Presented by:
Harris Mohamed, Space Domain Awareness Software Development Engineer, Kratos
Kyle Overton, Technical Sales Engineer, COMSPOC
Jason Ruggieri, COMSPOC
Steven Williams, Sr. Systems Architect, Kratos

Modern space domain awareness (SDA) requires more than individual sensor performance, it requires the ability to integrate diverse observations and disparate data into coherent, time-relevant operational insight. As orbital regimes become increasingly congested and contested, analysts must combine data from electro-optical (EO), radar, and passive radio frequency (RF) sensor phenomenologies to accurately and quickly detect maneuvers, maintain custody, identify anomalous maneuver and signal behavior, and investigate interference events with defensible confidence.

This short course provides a hands-on exploration of multi-phenomenology SDA within an enterprise fusion environment designed for high-accuracy, low-latency, and high-capacity operations. Participants will work through a realistic geosynchronous (GEO) scenario involving normal and anomalous satellite maneuvers and electromagnetic interference (EMI) events. Using a combination of real-world and physics-consistent simulated data, attendees will explore how different sensor modalities complement one another and how fusion workflows translate raw observations into actionable assessments.

Students will also directly encounter the vital and unique SDA insights gleaned from automated deep-dive RF signal analysis and RF pattern of life (POL) deviation detection.  Passive RF sensing plays a critical role in modern SDA by providing persistent, all-weather satellite tracking and signal analysis. RF observations also contribute essential additional context for maneuver detection, satellite identification, interference investigation, emitter geolocation and intent assessment.

Presented by:
Yvette Rodriguez, Research Director, Warfighting Acquisition University;
Nicholas LeGrand, Director of Systems Engineering Modernization, Systems Engineering and Architecture, Office of the Assistant Secretary of War for Mission Capabilities;
Monique Ofori, Program Manager, SAIC Support to OUSW(R&E) Systems Engineering and Architecture (SE&A), Office of the Assistant Secretary of War for Mission Capabilities

• Strategic Compliance & Policy: Understanding the implications of the 2025 “Transforming the Warfighting Acquisition System” memo and mandatory MOSA compliance for MDAPs.
• Technical Frameworks: Evaluating emerging standards (e.g., SOSA, CMOSS, and CCSDS) and the role of the Space Domain Awareness (SDA) Tools, Applications,  & Processing (TAP) Lab in creating an “operational bridge” for modular software and AI.
• Digital Engineering & Twins: Leveraging digital twins to model black box interfaces, ensuring interoperability while safeguarding proprietary vendor code.
• Security & Assurance: Addressing the vulnerabilities of open interfaces and the challenges of certifying modular AI/ML components for autonomous SDA.
• Global Interoperability: Navigating the integration of international modular networks into the Unified Data Library (UDL) based on expected significant improvements as part of the Space Force’s 2025 Data & AI Strategic Action Plan.

Presented by:
Roberto Furfaro, Professor, University of Arizona
Richard Linares, Associate Professor, Massachusetts Institude of Technology
Weston Faber, Advanced Concepts Engineer, L3Harris

• Understand Foundational DL Architectures: Gain familiarity with core Deep Learning concepts, Dense & Convolutional Neural Networks, Variational Autoencoders, and the critical underpinnings of how they learn from data.
• Explore Transformer-Based Architectures & LLMs: Learn how the Transformer architecture has enabled breakthroughs in natural language processing and sequence modeling tasks, and explore emerging approaches for adapting LLMs to numerical, time-series, and multi-modal data relevant to SSA.
• Deep Reinforcement Learning for SSA: Get introduced to reinforcement learning (Q-learning, Policy Gradient methods) and see how these ideas can be extended using Transformer-based models in advanced control and decision tasks.
• Implement Hands-On Case Studies: Walk through Python-based Jupyter notebooks illustrating how to apply these models (including LLM-driven data processing or analysis pipelines) to TLE data, light curves, and other SDA-relevant problems.
• Discuss Emerging Trends & Challenges: Understand current limitations, ethical considerations, and the future research directions in applying advanced AI, particularly large-scale models, to space domain awareness

Presented by:
Brenton Smith, Chief Technology Officer, Zendir
Christopher Capon, Co-Founder & CEO, Zendir

• Identify the operational constraints that govern sensor tasking in proliferated architectures
• Recognize how dwell time, field-of-view (FOV), and revisit rate trade off against detection probability and coverage
• Understand how sensor geometry, illumination, and measurement precision influence observability
• Evaluate prioritization strategies when multiple objects compete for limited sensor time
• Compare human-driven tasking decisions against structured, probability-informed scheduling approaches

• Narrow-field high-SNR dwell vs wide-field lower-sensitivity search
• Rapid retasking vs fuel/attitude constraints and stabilization time
• Frequent revisits for covariance containment vs distributed monitoring across multiple objects
• Immediate reacquisition focus vs maintaining catalog-wide situational awareness
• High-latency space-based sensing vs lower-latency but geometry-limited alternatives

Presented by:
Brandon Jones, Associate Professor, The University of Texas at Austin

This short course will present the fundamentals of Uncertainty Propagation (UP) through the lens of SSA to help attendees understand the implications of how uncertainty may be represented and correctly propagated.  Accurate UP is a fundamental driver for timely and accurate SSA.  Broadly, UP is the science of quantifying, reducing, and understanding sensitivity to uncertainty and how they map through our computer simulations/models and algorithms.  Propagated uncertainty can determine how we interpret knowledge extracted from data and computer simulations, what potential data may help to reduce uncertainty, and ultimately influences operator decisions.  The primary goal of this course is to introduce attendees to the fundamental principles and methods of uncertainty propagation, which they may use in their work and research after this course.

Topics will include definitions of key components of UP, sources of uncertainty in simulations, an overview of existing tools for uncertainty prediction through computer models, and various methods of analyzing sensitivity to sources of uncertainty.  Concepts will be illustrated through SSA-centric applications such as orbit uncertainty propagation and sensor tasking.  Finally, attendees will be introduced to the latest areas of research in UP, including multi-fidelity methods and the intersections of UP and machine learning.

This course is intended for those looking to improve their understanding on how to propagate uncertainty beyond the traditional methods based on linearization.  While anyone working with uncertainty will benefit from this course, it will be especially useful to those using or implementing software tools for simulation and data analysis.  Attendees familiar with basic probability and statistics will have the foundation necessary for this short course.

Presented by:
Jonathan Smith, Systems Analyst, Principal, Parsons

Cislunar space is not merely a larger version of geosynchronous orbit, it is a distinct region with its own orbital mechanics, strategic importance, and operational challenges. The Foundational Cislunar Education course introduces participants to the unique dynamics of cislunar space with the intent to shift perspective from earth-centric thinking to the three-dimensional, three-body framework required for operations in the earth-moon system. Cislunar space represents the next frontier in space exploration, infrastructure development, commercial enterprise, national security, and geopolitical competition.

Understanding its environment is critical for ensuring the U.S. and its allies are prepared to lead in this rapidly evolving domain.

Designed for professionals, students, and enthusiasts with basic understanding of near-earth space, the course emphasizes clarity and accessibility while maintaining an operational focus through real-world case studies.

The Foundational Cislunar Education course covers key topics such as the orbital mechanics of the earth-moon system, the strategic significance of the Lagrange points, lighting and distance considerations, sensing phenomenologies, and the advantages and limitations of common lunar trajectories and orbits.

Whether a new space operator, analyst, or experienced professional seeking a refresher, this course provides the tools to navigate the challenges and opportunities of this critical and emergent domain.

Participants will leave with a solid grasp of foundational concepts and emerging strategic considerations necessary to avoid risks and surprise and capitalize on opportunities in cislunar space which is the necessary knowledge to acquire, operate, and instruct on systems designed for operations in, to, and from cislunar space. Enroll today to prepare for the next era of space exploration and operations in cislunar space.

Presented by:
Christopher Grant, Founder, Hyperion Vector Systems
Jon Terry, Machine Learning Tech Lead, Hyperion Vector Systems

Space mission analysis tools like GMAT, SatSim, and open-source image processing libraries are mature and capable, but onboarding new operators takes months of specialized training. This timeline constrains workforce development when personnel rotate faster than they achieve proficiency (a gap documented in RAND’s 2024 assessment of the SDA mission, RR-A2318-1). Even experienced practitioners spend more time on tool syntax than on analysis the tools exist to support. This short course teaches participants to build AI agent systems that translate natural language and voice commands into executable tool configurations, converting tool-syntax overhead into analytical throughput.

Participants build a working AI agent during the course using either a provided Raspberry Pi with audio HAT (upgraded registration option) or their own laptop, connected to a course-provisioned GPU-compute environment running an open-weight large language model. To ensure reliable connectivity, the course provides a dedicated high-speed internet link to the session room independent of the conference network. The agent architecture follows a modular pipeline: voice or text capture, LLM interpretation via a self-hosted model, structured command generation through standardized tool interfaces, execution against GMAT and SatSim backends, and result visualization returned to the operator.

The course builds on the SatSim Chat concept presented at AMOS 2024 (De Alba et al.) and extends it in three directions. Attendees will implement a multi-tool agent framework capable of orchestrating workflows across GMAT for orbital mechanics, SatSim for synthetic imagery generation, and standard image processing libraries for detection algorithm evaluation. Voice interaction is useful in hands-free operational contexts, for operators with mobility limitations, or when domain-appropriate natural language is faster than tool-specific syntax. Course instructor and attendees will demonstrate scripted activity plans: pre-defined multi-step analysis workflows that chain tool invocations into coherent operational sequences executable through a single high-level command, modeled on real SDA problem statements.
The technical implementation uses open-weight and open-source components: faster-whisper for speech recognition, a self-hosted model for command interpretation (hosted on course GPU infrastructure; participants access in time-shared rotation during exercises), LangGraph for agent orchestration with tool-calling, and JSON-schema interfaces to GMAT’s scripting engine and SatSim’s configuration system. Participants working on their own laptops will connect to a instructor-hosted environment via SSH and web browser with no local GPU required.
Complex mission analysis tools effectively gate who participates in space domain awareness work. An intelligence analyst who understands orbital dynamics but hasn’t memorized STK or GMAT script syntax can’t contribute without months of tool training. Voice and natural language interfaces let existing expertise reach the tool without a syntax tax. This matters for workforce throughput in a community where trained analysts are the bottleneck, not compute or sensors.

Participants leave with a working reference implementation, complete source code, and documentation for extending the framework to additional tools in their own environments.

Presented by:
Alinda Mashiku, NASA CARA Program Manager, NASA
Dolan Highsmith, NASA CARA Chief Engineer, The Aerospace Corporation
Hunter Morris, Member of Technical Staff – Flight Mechanics, The Aerospace Corporation
Eliot Toumey, Conjunction Assessment Research Analyst, Omitron, Inc.

• Better understand how uncertainty is quantified in satellite orbit determination and prediction
• Understand the strengths and weaknesses of the leading risk metric calculations
• Understand the different philosophies of risk pertaining to conjunctions 
• Appreciate the operational utility of the various metrics
• Understand the fundamental tradeoff between false-alarm and missed-detection
• Be able to make remediation decisions with a clearer understanding of the applicable risk metric

Presented by:
Patrick North, Chief Remote Sensing Engineer, Image and Computer Scientist, Ansys Government Initiatives (AGI)

This short course is meant to provide the theoretical background for modeling and simulating Space Situational Awareness (SSA) and Space Domain Awareness (SDA) relevant targets of interest, imaging systems, and environmental conditions in the optical wavelengths from the visible through the longwave thermal infrared.

Imaging both with traditional and novel methodologies is an essential element of SSA/SDA, and this course will provide both a general, detailed, and complete framework for the math, physics, and engineering of the imaging process. This will offer professionals and students directly in SSA/SDA and adjacent fields a deep-dive into the techniques, a guided tour of simulation options available, and most importantly the ability to ask any detailed questions.

In addition to the course a companion field-guide has been developed and will be made available to participants. This is the missing textbook for image simulation including the radiometric equations, constants, and system models along with specifically tailored SSA/SDA specific parameters such as exoatmospheric solar illumination ranges, visual magnitude calculations and variations, thermal emissions values for standard space objects, and more.

Presented by:
Jeff Cornelius, Principal Systems Engineer, Parsons

Today’s space environment is saturated with data from satellite telemetry to spacecraft observations to physical force model parameters such as space weather indices and drag profiles. Space Situational Awareness (SSA) engineers and analysts can take advantage of new data analytics techniques to help predict future events or establish patterns in current data sets.

For those new to the concepts of machine learning and predictive analytics, this course will provide an introductory survey of various data analytics techniques. The course will follow a data to insight to decision making construct, beginning with an introduction to machine learning and definitions of terminology within the field. It will continue with a review of the mathematical background required for machine learning and examine the process of storing and cleaning data to achieve an ideal data set to train models.

The course will examine unsupervised and supervised learning methodologies with a particular emphasis on supervised learning techniques. Some of these that will be explored include principal component analysis, K-means clustering, association rule mining, artificial neural networks, random forest ensemble learning, and extreme gradient boosting. Each technique will be evaluated with pros and cons. Additionally, the confusion matrix will be introduced as a method for evaluating the effectiveness of a predictive model. New hands-on direct SSA examples in python will be provided.

Finally, the course will examine practical applications in the SSA field of these analytics techniques. These will include hypothetical examples as well as current, on-going, real-world projects to predict patterns of life or to understand satellite maneuver intent.

Students in the course should come away with a basic level understanding of data driven analytics and application to SSA. They will understand the common terminology and be able to ascertain if a method or approach is effective and is applicable to specific projects.

Presented by:
Douglas Deok-Soo Kim, CEO & Founder, SPACEMAP Inc.
Peter (Joonghyun) Ryu, CTO, Research Professor, SPACEMAP Inc.
Shawn Seunghwan Choi, CPO, SPACEMAP Inc.