REDUCING AND SPEEDING COMPUTATION IN DISTRIBUTED DATA FUSION

One of the primary Engagement Planning Processing Suite (EPPS) functions is to develop multisensor threat object tracks based on the local track estimates developed independently by the sensors and communicated to the EPPS nodes. The EPPS distributed system architecture affords a high level of system survivability and tracking/targeting accuracy but at the cost of potentially massive communications and data processing requirements. The focus of our proposed Phase I effort is to explore methods to significantly reduce both the volume of data communicated across the EPPS system links and the time required to perform the track fusion processing at the EPPS nodes. Our objectives are to investigate how information doman methods can be used to reduce the flow of local track data from the sensors to the EPPS nodes. And to explore methods for parallelizing the track fusion computations at the EPPS nodes.