Accession Number : ADA305669
Title : Cluster Recognition Algorithms for Battlefield Simulation.
Corporate Author : OKLAHOMA UNIV NORMAN
Personal Author(s) : Trytten, Deborah A.
PDF Url : ADA305669
Report Date : JAN 1996
Pagination or Media Count : 225
Abstract : The target acquisition fire support model (TAFSM) is a large scale, automated, artillery combat simulation model. This model has been developed over a period of years to include increasingly large and realistic battlefield situations and still retain its efficiency. To process more complicated battlefield scenarios within time constraints, units on the battlefield must be grouped into clusters, which will then be treated as a single entity in further processing. These clusters fall into three categories: circular clusters, linear clusters, and on-line clusters. To perform clustering within the time constraints, two classic clustering algorithms were modified to meet the stringent efficiency constraints. To detect circular clusters, a template-matching technique was designed. Linear and on-line clusters were found using an algorithm that has roots in the single-link clustering method. These algorithms were implemented and tested on 59 data sets from the TAFSM simulation. The new algorithms ran within a few seconds and created reasonable clusterings. (AN)
Descriptors : *COMPUTERIZED SIMULATION, *PATTERN RECOGNITION, DATA BASES, MATHEMATICAL MODELS, ALGORITHMS, IMAGE PROCESSING, SCENARIOS, POSITION(LOCATION), STOCHASTIC PROCESSES, DATA MANAGEMENT, REAL TIME, BATTLEFIELDS, PARAMETERS, TARGET ACQUISITION, MILITARY VEHICLES, TARGET RECOGNITION, RESOLUTION, EIGENVECTORS, EFFICIENCY, INPUT OUTPUT PROCESSING, EIGENVALUES, CLUSTERING, GROUND VEHICLES, ONLINE SYSTEMS, HIERARCHIES, DATA LINKS, FIRE SUPPORT, OBJECT ORIENTED PROGRAMMING.
Subject Categories : Computer Programming and Software
Distribution Statement : APPROVED FOR PUBLIC RELEASE