Parallel Implementation of the Terrain Masking Algorithm
Abstract
The terrain masking algorithm is a key algorithm used in battle management and mission planning in choosing the optimal location site for mobile radar and jammer systems. This algorithm calculates the region of clear line-of- sight at a particular altitude for an emitting device. Currently, to calculate the site for one radar or jammer system it can take up to eight hours running on a uniprocessor system. This is unacceptable because it increases the time and cost involved in the planning of a successful mission strategy. A solution to this problem is to utilize a parallel architecture and modify the existing sequential version of the terrain masking algorithm so that it can effectively execute in parallel. The approach presented uses the transputer architecture for executing the algorithm. This architecture was selected because it allows the simulation of a number of MIMD message passing architecture topologies, without having to drastically rewrite the parallel implementation of the algorithm. The approach for parallelizing the terrain masking algorithm involves reverse engineering the sequential version. The algorithm decomposition involved defining the data dependencies inherent in the sequential version so that the data can be properly partitioned for parallel execution. Software development, Parallel processing, Software engineering, Transputers, Terrain masking.
Document Details
- Document Type
- Technical Report
- Publication Date
- Mar 01, 1994
- Accession Number
- ADA281017
Entities
People
- Milissa M. Benincasa
Organizations
- Rome Laboratory