Partially-Connected Hybrid Precoder Design for ISAC via Joint Grouping and Precoding

With the spectrum resources becoming overwhelmingly scarce, Integrated Sensing and Communications (ISAC) based on the multi-input-multi-output (MIMO) array is regarded as a pivotal approach for alleviating the spectrum congestion. Current researches on MIMO based ISAC predominantly focus on the design of the precoding matrix, while overlooking the potential performance improvement brought by the precoder structure. In light of this, we propose the partially-connected hybrid precoder design for ISAC via joint grouping and pre-coding optimization in this work. A fully digital approximation is applied to design the precoding matrices, incorporating the grouping structure optimization between radio frequency (RF) chains and transmit antennas via antenna selection. We consider two partially-connected hybrid precoder structures via overlapped and non-overlapped grouping, respectively. To address the coupling between antenna grouping and precoding, we propose to decompose the original problem into distinct subproblems using an alternating minimization (AM) method, and provide effective methods of tackling each of them. Experimental results demonstrate that grouping optimization significantly enhances the performance of the partially-connected hybrid precoder in both overlapped and non-overlapped grouping, validating the effectiveness of the proposed methods.