EMSI proposes to demonstrate the feasibility of extending our machine-learning-based SAR classifiers to provide real-time high confidence maritime target identification from ISAR data collected and processed onboard Tomahawk and other high-speed weapons. To this end, we will modify our classifier architectures to accommodate target motion and missile SWAP constraints, simulate realistic ISAR data sets, and to determine: which ISAR features are salient to ship identification and to which physical features they correspond; and what ATR performance is achievable from a missile platform, as a function of ship type, environmental and radar parameters, and missile trajectory.
Benefit: The increasing proliferation of A2AD environments is driving ISR platforms to operate at standoff ranges. Consequently, they may not be able to provide sufficiently accurate tracking information to guide an anti-ship missile to its target, which may move considerably during missile flight. The proposed effort will enable missile radars to classify vessels, so as to select the correct target.
Keywords: maritime ATR, maritime ATR, Machine Learning, Radar, ISAR
