Air Force Space Command Satellite Orbit Predictor Using Parallel Virtual Machines
Abstract
Parallel computing is the wave of the future. As the need for computational power increases, one processor is no longer sufficient to achieve the speed necessary to solve today's complex problems. The Air Force Space Command (AFSPACEOM) tracks approximately 8000 satellites daily; the model used by the AFSPACECIM, SGP4, (Simplified General Perturbation Model Four), has been the operational model since 1976. This thesis contains a detailed discussion of the mathematical theory of the SGP4 model. The tracking of a satellite requires extensive calculations. The satellite can be track more efficiently with parallel processing techniques. The principles developed are applicable to a Naval ship tracking multiple incoming threats; the increase in the speed of processing incoming data would result in personnel being informed faster and thus allow more time for better decisions during combat. Three parallel algorithms applied to SGP4 for implementations on a Parallel Virtual Machine (PVM) are developed. PVM is a small software package that allows a network of computer workstations to appear as a single large distributed-memory parallel computer. This thesis contains a description of several algorithms for the implementation on PVM to track satellites, the optimal number of workstations, and methods of distributing data.
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 01, 1993
- Accession Number
- ADA276316
Entities
People
- Susan K. Brewer-matusiak
Organizations
- Naval Postgraduate School