Large scale simulation of a distributed target tracking system

Abstract
For a large class of defense problems involving distributed, multisensor, multitarget tracking based on data fusion techniques, there are many factors that can affect the cost-performance tradeoff for the overall system. Although the complexity of such problems depends on the overall dimensionality, there is usually sufficient complexity involved that the system design cannot be approached with formal methods. Given this perspective, we have developed an initial version of an analysis and design testbed for large-scale distributed tracking problems that allows the effects of various influencing features to be empirically explored. The initial version of the testbed has been synthesized for a decentralized multiple target tracking system under a distributed multiple sensor environment, with the purpose of studying the effects of tracking algorithms, track fusion algorithms, and topology of internodal connectivity. With the present testbed, we examine the tradeoff between the overall computational load of different large scale distributed systems and the composite tracking accuracy. Simulation-based results using the testbed will be presented along with the implications for system design.

This publication has 4 references indexed in Scilit: