Distributed Environmentally-Adaptive Detection, Classification, and Localization Using a Cooperative Sensor Network

摘要

The specific objective of this effort is to develop distributed detection, classification, and localization (DCL) algorithms suitable for application to the nonlinear inversion problems encountered in ocean acoustics that can be nested within an over-reaching system concept of a cooperative sensor network. Joint parameter estimation processes were developed wherein both target parameters and environmental acoustic parameters (primarily bottom geoacoustic) are estimated. The latest tracking work incorporated a likelihood surface formulation with the JPDA algorithm. We've determined that work is still needed to improve the performance of the JPDA algorithm with the likelihood surface formulation. Results were encouraging for the baseline tracking scenario where the truth is known. An initial framework for creating target times series associated with a contact-based tracking data set was expanded and a physically-motivated feature set and classifier was improved with the addition of classification.