WP3: JCAS analytics

Objectives: Developing adaptive self-optimising algorithms for allocation of on-board resources; and leveraging knowledge distillation techniques for large ML model compression with minimal accuracy loss (Obj5); Devising signal processing techniques for multiband channel estimation and channel-event matching and leveraging federated learning to preserve privacy (Obj6); Optimising JCAS for accurate object detection/tracking, and low-latency reinforcement learning for real-time operation (Obj7); designing context-aware solution to fuse multi-modal sensor information to enhance robustness and minimise the energy and computational resource requirements (Obj8).

WP Leader: University of Trento