ARETE ASSOCIATES — Department of Defense SBIR Phase I: N172-108

Description

Aret proposes to develop and implement an innovative deep learning process to fuse multiple sensor modalities for classification and identification of watercraft. Machine learning techniques and simulations capabilities will be used to construct a trained classification processor that will be implementable in SWaP-Computational limited operations. The Phase I demonstration will leverage a suite of existing EO, IR and radar simulation capabilities that will be transformed through a reification machine learning process and used to train a classification neural network. The trained classification neural network will then be ported and implemented into a classification processor on platforms for real-time classification of maritime targets. Applications range from targeting guidance to maritime surveillance.