The current Space Traffic Management (STM) community was not prepared for the exponential increase in the resident space object (RSO) population that has taken place over the last several years. The combination of poor communication infrastructure and long scheduling lead times of the Space Surveillance Network (SSN) prevent any type of reactive sensor tasking, which is required in event of anomaly detection. This dissertation was designed to survey extensions to the classical notions of covariance based sensor tasking strategies and develop a methodology for evaluating these techniques. A suboptimal partially observable Markov decision process (POMDP) was used as the simulation framework to test various reward functions and decision algorithms while enabling autonomous, reactive sensor tasking. The goal of this work was used the developed evaluation methodology to perform statistical analyses to determine which metrics were most reliable and efficient for Space Traffic Management (STM) of the geosynchronous Earth orbit (GEO) resident space object (RSO) catalog. Hypotheses were tested against simulations of 873 resident space object (RSO) in geosynchronous Earth orbit (GEO) being tracked by 18 heterogeneous, geographically disperse ground-based electro-optical (EO) sensors. This dissertation evaluates the ability of various sensor tasking metrics to produce rewards that maximize geosynchronous Earth orbit (GEO) catalog coverage capability of a sensor network under realistic communication restrictions. / Doctor of Philosophy / Space is getting crowded at an increasing rate. Communication issues and rigid scheduling of the Space Surveillance Network (SSN) prevent reactive sensor tasking, which is needed to alleviate this issue. This dissertation was designed to survey different sensor tasking strategies and develop a methodology for evaluating these techniques. A discrete time estimator called a suboptimal partially observable Markov decision process (POMDP) was used as the simulation framework to test various reward functions and decision algorithms while enabling autonomous, reactive sensor tasking. The goal of this work was used the developed evaluation methodology to perform statistical analyses to determine which metrics were most reliable and efficient for Space Traffic Management (STM) of the geosynchronous Earth orbit (GEO) resident space object (RSO) catalog. Multiple simulation scenarios were evaluated, with the first focused on determining the proper metrics in the ideal sensor network distribution case. From there, hypotheses were tested against simulations of a geographically disperse network of ground-based electro-optical (EO) sensors. This dissertation evaluates the ability of various sensor tasking metrics to produce rewards that maximize geosynchronous Earth orbit (GEO) catalog coverage capability of a sensor network under realistic communication restrictions

Traditionally, space situational awareness (SSA) sensor networks have relied on dynamic programming theory to generate tasking plans which govern how sensors are allocated to observe resident space objects. Deep reinforcement learning (DRL) techniques, with their ability to be trained on simulated environments, which are readily available for the SSA sensor allocation problem, and demonstrated performance in other fields, have potential to exceed performance of deterministic methods. The research presented in this dissertation develops techniques for encoding an SSA environment model to apply DRL to the sensor allocation problem. This dissertation is the compilation of two separate but related studies. The first study compares two alternative invalid action handling techniques, penalization and masking. The second study examines the performance of policies that have forecast state knowledge incorporated in the observation space. / Doctor of Philosophy / Resident space objects (RSOs) are typically tracked by ground-based sensors (telescopes and radar). Determining how to allocate sensors to RSOs is a complex problem traditionally performed by dynamic programming techniques. Deep reinforcement learning (DRL), a subset of machine learning, has demonstrated performance in other fields, and has the potential to exceed performance of traditional techniques. The research presented in this dissertation develops techniques for encoding a space situational awareness environment model to apply DRL to the sensor allocation problem. This dissertation is the compilation of two separate but related studies. The first study compares two alternative invalid action handling techniques, penalization and masking. The second study examines the performance of policies that have forecast state knowledge incorporated in the observation space.

Low Earth orbit represents a valuable limited natural resource. Of particular interest are sun-synchronous orbits; it is estimated that approximately 44% of low Earth satellites are sun-synchronous. A previously developed sun-synchronous orbit slot architecture is considered. An in-depth analysis of the relative motion between satellites and their corresponding slots is performed. The long-term evolution of Earth's orbital environment is modeled by a set of coupled ordinary differential equations. A metric for quantifying the benefit, if any, of implementing a sun-synchronous architecture is developed. The results indicate that the proposed slot architecture would reduce the frequency of collisions between satellites in sun-synchronous orbits.

Space debris growth and an influx in space traffic will create a need for increased space traffic management. Due to orbital population density and likely future growth, the implementation of a slot architecture to Sun-synchronous orbit is considered in order to mitigate conjunctions among active satellites. This paper furthers work done in Sun-synchronous orbit slot architecture design and focuses on two main aspects. First, an in-depth relative motion analysis of satellites with respect to their assigned slots is presented. Then, a method for developing a slot architecture from a specific set of user defined inputs is derived.