Scalable Conjunction Processing using Spatiotemporally Indexed Ephemeris Data

Abstract

The collision warnings produced by the Joint Space Operations Center JSpOC) are of critical importance in protecting U.S. and allied spacecraft against destructive collisions and protecting the lives of astronauts during space flight. As the Space Surveillance Network (SSN) improves its sensor capabilities for tracking small and dim space objects, the number of tracked objects increases from thousands to hundreds of thousands of objects, while the number of potential conjunctions increases with the square of the number of tracked objects. Classical filtering techniques such as apogee and perigee filters have proven insufficient. Novel and orders of magnitude faster conjunction analysis algorithms are required to find conjunctions in a timely manner. Stellar Science has developed innovative filtering techniques for satellite conjunction processing using spatiotemporally indexed ephemeris data that efficiently and accurately reduces the number of objects requiring high-fidelity and computationally-intensive conjunction analysis. Two such algorithms, one based on the k-d Tree pioneered in robotics applications and the other based on Spatial Hashing used in computer gaming and animation, use, at worst, an initial O(N log N) preprocessing pass (where N is the number of tracked objects) to build large O(N) spatial data structures that substantially reduce the required number of O(N2) computations, substituting linear memory usage for quadratic processing time. The filters have been implemented as Open Services Gateway initiative (OSGi) plug-ins for the Continuous Anomalous Orbital Situation Discriminator (CAOS-D) conjunction analysis architecture. We have demonstrated the effectiveness, efficiency, and scalability of the techniques using a catalog of 100,000 objects, an analysis window of one day, on a 64-core computer with 1TB shared memory.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Sep 01, 2014
Accession Number
ADA619600

Entities

People

  • Clay Alberty
  • Irene A. Budianto-ho
  • Randall Scarberry
  • Robert M. Sivilli
  • Stephen L. Johnson

Organizations

  • Air Force Research Laboratory

Tags

Communities of Interest

  • Air Platforms
  • Space
  • Weapons Technologies

DTIC Thesaurus Topics

  • Air Force Research Laboratories
  • Algorithms
  • Artificial Satellites
  • Collisions
  • Computations
  • Earth Orbits
  • Efficiency
  • Ephemerides
  • Filters
  • Filtration
  • Hash Tables
  • Low Earth Orbits
  • Orbits
  • Reliability
  • Space Objects
  • Space Surveillance
  • Spacecraft

Readers

  • Distributed Systems and Data Platform Development
  • Image Processing and Computer Vision.
  • Space Exploration and Orbital Mechanics.

Technology Areas

  • AI & ML
  • AI & ML - Machine Learning Algorithms
  • Autonomy
  • Space
  • Space - Orbital Debris
  • Space - Satellites
  • Space - Space Objects