I am a PhD Student in Earth and Space Sciences at York University. I completed my Bachelors in Space Engineering, graduating first class with distinctions. Space Engineering is a branch of systems engineering focused on the Space Sector.
My Research focuses on Space Situational Awareness (SSA), which is a field of study which includes detection, identification, and tracking of various space objects. My current work focuses on Very-Short Arc Initial Orbit Determination, a major problem for tracking space objects from space-based sensors.
Outside of research I love to travel, be outdoors, and eat good food. If you are interested in my work or just chatting, feel free to send me an email or a message on LinkedIn.
Warning
Problem: The current name of your GitHub Pages repository ("Solution: Please consider renaming the repository to "
http://".
However, if the current repository name is intended, you can ignore this message by removing "{% include widgets/debug_repo_name.html %}" in index.html.
Action required
Problem: The current root path of this site is "baseurl ("_config.yml.
Solution: Please set the
baseurl in _config.yml to "Education
-
Lassonde School of Engineering / York UniversityDepartment of Earth and Space Sciences
Ph.D. StudentSep. 2025 - present -
Lassonde School of Engineering / York UniversityB.Eng Space EngineeringSep. 2021 - Jun. 2025
Honors & Awards
-
Enbridge Graduate Student Award2026
-
Queen Elizabeth II Graduate Scholarship in Science and Technology2026
-
Mitacs Globalink Research Award Abroad2026
-
York Graduate Scholarship2025
-
Lassonde Undergraduate Research Award (LURA), Fall / Winter2024
-
Lassonde Undergraduate Research Award (LURA), Summer2024
-
Lassonde Undergraduate Research Award (LURA), Fall / Winter2023
-
Lassonde Undergraduate Research Award (LURA), Summer2023
Selected Publications (view all )
Hardware-in-the-loop validation of physics-informed neural network for angles-only initial orbit determination in very short arcs
Ian Porto, Marissa Myhre, Sofia Meson-Perez, Angel Porras-Hermoso, Gunho Sohn, Regina S.K. Lee
Accepted to 46th Committee on Space Research (COSPAR) Scientific Assembly 2026 Spotlight
The exponential growth of the Resident Space Object (RSO) population in Low Earth Orbit (LEO) necessitates robust Space Situational Awareness (SSA) capabilities. However, the high relative velocities in LEO frequently result in very short arc observations, which cause traditional Initial Orbit Determination (IOD) methods to diverge or yield physically impossible solutions. Developing algorithms capable of handling these regimes requires data that accurately reflects the complex, non-Gaussian noise characteristics of real optical sensors. This work presents a robust IOD pipeline developed and validated using a hardware-in-the-Loop optical starfield simulator incorporating the engineering model of the UPMSAT-4 multi-use star tracker. The payload comprises a 4.2-megapixel sCMOS sensor paired with a 29.7$^\circ$ $\times$ 29.7$^\circ$ field-of-view lens. By physically stimulating this sensor with high-fidelity generated starfields and RSOs, we generate a training dataset that preserves flight-representative radiometric properties. Crucially, this hardware-in-the-loop setup captures degradation modes that cannot be reliably simulated. This hardware-validated imagery is used to train and evaluate a Physics-Informed Neural Network (PINN). The architecture employs a two-stage hybrid approach to navigate the difficult very-short arc solution space. First, a metaheuristic global optimizer identifies the approximate orbital basin by searching the admissible region. Second, the PINN performs local refinement by embedding the governing equations of orbital motion as a regularization constraint. This physics-guided loss function allows the network to distinguish between true orbital dynamics and the specific sensor artifacts present in the Hardware-in-the-Loop data. Preliminary analysis highlights the algorithm's resilience against these realistic sensor degradations compared to industry-standard iterative solvers. Results indicate that the physics-informed regularization significantly mitigates the singularity issues common in angles-only inversion, maintaining solution stability even when subjected to the non-linear noise profiles. These findings provide critical validation for the use of physics-guided machine learning in processing real-world optical data, demonstrating a pathway toward more robust, noise-resilient SSA architectures for future distributed sensor networks.
Hardware-in-the-loop validation of physics-informed neural network for angles-only initial orbit determination in very short arcs
Ian Porto, Marissa Myhre, Sofia Meson-Perez, Angel Porras-Hermoso, Gunho Sohn, Regina S.K. Lee
Accepted to 46th Committee on Space Research (COSPAR) Scientific Assembly 2026 Spotlight
The exponential growth of the Resident Space Object (RSO) population in Low Earth Orbit (LEO) necessitates robust Space Situational Awareness (SSA) capabilities. However, the high relative velocities in LEO frequently result in very short arc observations, which cause traditional Initial Orbit Determination (IOD) methods to diverge or yield physically impossible solutions. Developing algorithms capable of handling these regimes requires data that accurately reflects the complex, non-Gaussian noise characteristics of real optical sensors. This work presents a robust IOD pipeline developed and validated using a hardware-in-the-Loop optical starfield simulator incorporating the engineering model of the UPMSAT-4 multi-use star tracker. The payload comprises a 4.2-megapixel sCMOS sensor paired with a 29.7$^\circ$ $\times$ 29.7$^\circ$ field-of-view lens. By physically stimulating this sensor with high-fidelity generated starfields and RSOs, we generate a training dataset that preserves flight-representative radiometric properties. Crucially, this hardware-in-the-loop setup captures degradation modes that cannot be reliably simulated. This hardware-validated imagery is used to train and evaluate a Physics-Informed Neural Network (PINN). The architecture employs a two-stage hybrid approach to navigate the difficult very-short arc solution space. First, a metaheuristic global optimizer identifies the approximate orbital basin by searching the admissible region. Second, the PINN performs local refinement by embedding the governing equations of orbital motion as a regularization constraint. This physics-guided loss function allows the network to distinguish between true orbital dynamics and the specific sensor artifacts present in the Hardware-in-the-Loop data. Preliminary analysis highlights the algorithm's resilience against these realistic sensor degradations compared to industry-standard iterative solvers. Results indicate that the physics-informed regularization significantly mitigates the singularity issues common in angles-only inversion, maintaining solution stability even when subjected to the non-linear noise profiles. These findings provide critical validation for the use of physics-guided machine learning in processing real-world optical data, demonstrating a pathway toward more robust, noise-resilient SSA architectures for future distributed sensor networks.
Thermo-Optical Characterization and Qualification of a Space Situational Awareness Payload
Ian Porto, Marissa Myhre, Vithurshan Suthakar, Regina S.K. Lee
5th IAA Conference on Space Situational Awareness 2026 Spotlight
In this study, we present a Technology Readiness Level (TRL) 7 payload comprising a 4.2-megapixel sCMOS sensor and a lens with a 29.7 by 29.7 degree field of view, which has flown three times on stratospheric balloon missions with the Canadian Space Agency and the Centre National d'Études Spatiales.
Thermo-Optical Characterization and Qualification of a Space Situational Awareness Payload
Ian Porto, Marissa Myhre, Vithurshan Suthakar, Regina S.K. Lee
5th IAA Conference on Space Situational Awareness 2026 Spotlight
In this study, we present a Technology Readiness Level (TRL) 7 payload comprising a 4.2-megapixel sCMOS sensor and a lens with a 29.7 by 29.7 degree field of view, which has flown three times on stratospheric balloon missions with the Canadian Space Agency and the Centre National d'Études Spatiales.