Abstract
The high mobility of unmanned aerial vehicles (UAVs) presents substantial challenges to conventional beam tracking schemes. Although integrated sensing and communication (ISAC) technology provides a promising solution, existing mono-static ISAC frameworks are constrained by limited observation angles and susceptibility to environmental dynamics. This paper proposes a sensing-enabled predictive beam tracking scheme for UAVs in cell-free massive multiple-input multiple-output (CF-mMIMO) ISAC networks. By exploiting multi-angle observations from distributed access points (APs), the proposed scheme employs an extended Kalman filter (EKF) to extract the coordinates of UAV user equipments (UEs) from sensing echoes, thereby enabling pilot-free beamforming updates. In addition, a neural network–based algorithm is developed to solve the AP–UE association optimization problem, aiming to maximize the achievable sum rate. Simulation results demonstrated that the proposed scheme can precisely track UEs' trajectories and effectively improve the achievable sum rate. Index Terms—unmanned aerial vehicle, integrated sensing and communication, cell-free massive multiple-input multiple-output, extended Kalman filter, beam tracking