Battle Engagement Area Simulator/Tracker
Abstract
The Battle Engagement Area Simulator/Tracker (BEAST) Project was first proposed by Boris, et al, to demonstrate algorithms and a problem solving framework suited to processing the sensor contact reports and tracks from tens of thousands to hundreds of thousands of Satellites and missiles on ballistic trajectories and in low earth orbit. The goal was to rethink such a system from the ground up with mathematically optimal or near optimal parallel algorithms scaling at least as lnN on an N processor system. The Surveillance, Correlation, and Tracking problem is the computation-limited kernel of future battle management systems being developed for the military. In this report, we show how high-performance tracking can be performed on the Connection Machine (CM) using a Monotonic Lagrangian Grid (MLG). This high-performance algorithm scales better than traditional algorithms which scale as N2 or N3, where N is the number of objects. The CM, a fine grain Single Instruction Instruction Multiple Data (SIMD) computer, is described; and its affects on algorithm design are discussed. Fully parallel algorithms developed at the NRL Laboratory for Computational Physics and Fluid Dynamics for the CM are described for scenario generation, near-neighbor search within the MLG, sensor report generation, fusion of sensor reports and a tracking. Finally, a description of a Strategic Defense Initiative (SDI) operational system using a heterogeneous assembly of computers is given.
Document Details
- Document Type
- Technical Report
- Publication Date
- Oct 08, 1990
- Accession Number
- ADA227525
Entities
People
- J. Michael Picone
- Jay Paul Boris
- R. L. Kolbe
Organizations
- United States Naval Research Laboratory