Sparse Array Design for ISAC in the Presence of Beam-Squint

This paper presents an efficient approach for joint antenna selection and hybrid beamforming for integrated sensing and communications (ISAC) in the presence of beamsquint. We proposed a manifold optimization-based technique that incorporates beam-squint compensation for hybrid beamformer design. To reduce computational and memory costs for antenna selection, we propose novel algorithms leveraging grouped subarrays, wherein the entire array is divided into non-overlapping groups. Numerical simulations show that the proposed approach exhibits up to 95% lower complexity for large antenna arrays while maintaining satisfactory communications with approximately 6% loss in the achievable rate.