Optimal Tasking of Ground-Based Sensors for Space Situational Awareness Using Deep Reinforcement Learning

Space situational awareness (SSA) is becoming increasingly challenging with the proliferation of resident space objects (RSOs), ranging from CubeSats to mega-constellations. Sensors within the United States Space Surveillance Network are tasked to repeatedly detect, characterize, and track these RSOs to retain custody and estimate their attitude. The majority of these sensors consist of ground-based sensors with a narrow field of view and must be slew at a finite rate from one RSO to another during observations. This results in a complex combinatorial problem that poses a major obstacle to the SSA sensor tasking problem. In this work, we successfully applied deep reinforcement learning (DRL) to overcome the curse of dimensionality and optimally task a ground-based sensor. We trained several DRL agents using proximal policy optimization and population-based training in a simulated SSA environment. The DRL agents outperformed myopic policies in both objective metrics of RSOs' state uncertainties and the number of unique RSOs observed over a 90-min observation window. The agents' robustness to changes in RSO orbital regimes, observation window length, observer's location, and sensor properties are also examined. The robustness of the DRL agents allows them to be applied to any arbitrary locations and scenarios.

Space Debris Tracking with the Poisson Labeled Multi-Bernoulli Filter

This paper presents a Bayesian filter based solution to the Space Object (SO) tracking problem using simulated optical telescopic observations. The presented solution utilizes the Probabilistic Admissible Region (PAR) approach, which is an orbital admissible region that adheres to the assumption of independence between newborn targets and surviving SOs. These SOs obey physical energy constraints in terms of orbital semi-major axis length and eccentricity within a range of orbits of interest. In this article, Low Earth Orbit (LEO) SOs are considered. The solution also adopts the Partially Uniform Birth (PUB) intensity, which generates uniformly distributed births in the sensor field of view. The measurement update then generates a particle SO distribution. In this work, a Poisson Labeled Multi-Bernoulli (PLMB) multi-target tracking filter is proposed, using the PUB intensity model for the multi-target birth density, and a PAR for the spatial density to determine the initial orbits of SOs. Experiments are demonstrated using simulated SO trajectories created from real Two-Line Element data, with simulated measurements from twelve telescopes located in observatories, which form part of the Falcon telescope network. Optimal Sub-Pattern Assignment (OSPA) and CLEAR MOT metrics demonstrate encouraging multi-SO tracking results even under very low numbers of observations per SO pass.

Multi-Site Simultaneous Time-Resolved Photometry with a Low Cost Electro-Optics System

Sunlight reflected off of resident space objects can be used as an optical signal for astrometric orbit determination and for deducing geometric information about the object. With the increasing population of small satellites and debris in low Earth orbit, photometry is a powerful tool in operational support of space missions, whether for anomaly resolution or object identification. To accurately determine size, shape, spin rate, status of deployables, or attitude information of an unresolved resident space object, multi-hertz sample rate photometry is required to capture the relatively rapid changes in brightness that these objects can exhibit. OSCOM, which stands for Optical tracking and Spectral characterization of CubeSats for Operational Missions, is a low cost and portable telescope system capable of time-resolved small satellite photometry, and is field deployable on short notice for simultaneous observation from multiple sites. We present the electro-optical design principles behind OSCOM and light curves of the 1.5 U DICE-2 CubeSat and simultaneous observations of the main body of the ASTRO-H satellite after its fragmentation event.