Artificial Intelligence / Machine Learning Enabled Battle Damage Assessment

SensorOps aims to develop a real-time, automated method for Battle Damage Assessment (BDA), using its TargetModeler (TM) simulation tool, incorporating state of the art gaming physics simulation driven by ML damage approximation. SensorOps will demonstrate the feasibility of: Training an ML model on real BDA data in order to generate an approximation of how the combination of • Ordinance, employment method and target produce various forces Building on top of TargetModeler’s user interface to allow a user to select a weapon system, delivery mechanism, environmental conditions and strike target • Use the ML approximation model to apply appropriate forces, deformation, marring and fracturing to the physics and destruction systems within Unreal Engine, the framework powering TargetModeler • Feed a large number of variants of these results into TargetModeler’s existing ability to stage and collect labelled training data • Pass this training data into TargetModeler’s existing no-code ML training pipeline to produce an ML model capable of detecting BDA results from the weapon/target pairing previously selected by the user