Sensing-Aware OTA-FEEL: Joint Scheduling and Beamforming Approach

In this paper, we propose a robust design for overt-the-air federated edge learning (OTA-FEEL) that leverages sensing capabilities at the parameter server (PS) to mitigate the impact of target echoes on the analog model aggregation. We derive novel expressions for the Cramér-Rao bound of the target response and mean squared error (MSE) of the estimated global model to measure sensing and aggregation quality. We then develop a joint scheduling and beamforming framework that optimizes the OTA-FEEL performance while maintaining desired sensing and communication quality. The resulting scheduling problem reduces to a combinatorial mixed-integer nonlinear programming problem (MINLP). We develop a low-complexity hierarchical method based on the matching pursuit algorithm that uses a step-wise strategy to omit the least effective devices in each iteration based on a metric that captures both the aggregation and sensing quality. Numerical results show that accurate sensing effectively suppresses target echoes on the uplink, preserving model aggregation quality despite interference.