Abstract
This paper proposes a multi-stage position and velocity estimation algorithm in the context of near-field (NF) integrated sensing and communications (ISAC). We consider a colocated multiple-input multiple-output (MIMO) orthogonal frequencydivision multiplexing (OFDM) system. Both the wavefront curvature in NF and the wideband feature of OFDM are utilized to jointly estimate the range and the angle of the targets. The proposed algorithm relies on spectral estimation methods and maximum likelihood (ML) refinement steps for target sensing. Also, the estimation performance is compared with the CramérRao lower bound (CRLB). Given the initial estimations of the targets' position and mobility parameters through our proposed algorithm, we further develop a precoding (i.e., beamfocusing) design. The proposed design exploits the finite beam depth and width in NF with minimal beam cross-correlation of different targets across both range and angular domains, allowing better overall resolution. The precoding design also accounts for a tradeoff behavior between communication and sensing performance.