Enhanced Superresolution Path Delay Estimation for Channel Impulse Response Based Localization

As presented in recent studies, integrated sensing and communication (ISAC) is suggested as an essential component of the next-generation communications systems. ISAC can exploit channel estimations obtained from communication to perform sensing, particularly utilizing time delay information for applications like localization. However, conventional techniques, e.g., cross-correlation (CCR)-based methods, cannot provide accurate estimates, requiring the use of superresolution path delay estimation (SPDE) methods for precise sensing. In addition, SPDE can offer sub-sampling interval resolution, boosting the accuracy of these methods. Within the subspace-based SPDE methods, the path delay estimation (PDE) matrix, constructed using frequency-domain channel estimates, is crucial in obtaining signal- and noise-subspaces. Therefore, extensions to the PDE matrix, such as the forward-backward averaging (FBA), have been proposed to enhance estimation performance, along with flexible formulation for the PDE matrix. In this work, we propose a generalized expression that expands on this flexible formulation to integrate the FBA method. Inspired by this expression, we propose a virtual-band enhanced PDE matrix that improves the resolution of the path delay estimation. Simulation results indicate significant improvements in estimation performance, demonstrating the potential for more accurate sensing.