Resident Space Object Feature Identification using Hierarchical Mixtures of Experts

As the number of objects in orbit and spacecraft capability grows, the need for space situational awareness (SSA) also increases. The identification and cataloging of resident space objects (RSOs) is a large and complex problem, and new methods of performing these tasks are necessary to keep up with the rapid growth of objects in space. Emergent Space Technologies, Inc. (Emergent) and the University of Texas at Austin (UT) worked together to develop an algorithm to aid in detecting RSO characteristics using a Hierarchical Mixture of Experts (HME). This algorithm has been tested in simulation, and has been shown to be capable of detecting the size, shape, reflectivity, and maneuvers performed by an RSO. 1. BACKGROUND Space situational awareness (SSA) involves gaining and maintaining knowledge of objects in Earth orbit, particularly objects in the vicinity of valuable space assets. Hazards to space assets are increasing, largely due to the growing number of resident space objects (RSOs) as well as the growing capabilities of potential adversaries. How to effectively automate, schedule, and manage the decision-making associated with maintaining the U.S. Space Object Catalog is a challenging problem of increasing importance to our national security [1]. The U.S. Space Object Catalog currently lists approximately 15,000 trackable objects accounting for approximately 5,800 tons of on-orbit mass. The total population is thought to exceed 20,000 objects larger than 10 cm [2]. Before objects can b