Joint Robust Antenna Selection and Beamforming for Integrated Sensing and Communications

Multiple-input multiple-output (MIMO) integrated sensing and communications (ISAC) systems suffer from significant system overhead due to the large number of required radio frequency (RF) units. Moreover, prior knowledge of the channel state information (CSI) and target angles is inaccurate in practical scenarios. In this paper, we investigate joint robust antenna selection and beamforming for the transmitter of MIMO ISAC systems. We assume that the targets are constrained within a region of interest (ROI), while the communication CSI lies within an uncertainty ball centered around the estimated CSI. To reduce hardware costs, we incorporate antenna selection into the robust beamforming design. The reweighted $\ell_{1}$-norm technique is employed to approximate the $\ell_{0}$-norm constraint on the number of utilized antennas as a convex constraint. Consequently, the non-convex joint design problem is relaxed to a convex one and can be solved iteratively. Numerical results demonstrate that the proposed algorithm maintains radar sensing performance within the ROI without significant degradation and satisfies the quality of service (QoS) constraints for communication users, even with a reduced number of transmit antennas.