Autonomous Collaborative ISR and Targeting

UAS are poised to play an increasingly prominent role in ISR and targeting applications. In situations where a human cannot be in the loop for every transfer, automatic methods for target handoff are required and must perform reliably. In the Phase I effort, Progeny developed capabilities to detect maritime targets from full-motion imagery at long range. Here, we propose to recognize specific targets and use known platform positions to estimate target position based on image-based measurements. We will explore ways to maximize the benefit from collaboration in targeting tasks, where data sharing across heterogeneous sensors and different perspectives can maintain and relay accurate target tracks, disambiguate closely spaced targets, etc. Further, we will extend recognition-based data association technology to support target handoff that is robust to changes in sensor, modality, and perspective. We will explore the use of synthetically rendered data to support these functions for targets that have limited available data for training. We will integrate these innovative technologies into a developmental prototype system, and then conduct in-flight evaluation experiments with increasing complexity to demonstrate capability.