Abstract
Optical correlation-domain reflectometry (OCDR) is a technique for locating reflective events and faults in optical fiber networks. However, conventional OCDR assumes that only one correlation peak exists within the measurement range; otherwise, signals originating from different positions are mixed in the detected signal, limiting the measurable length to the correlation-peak spacing. In this paper, we propose a compressed-sensing-assisted OCDR framework for measurement-range extension under multiple-correlation-peak conditions. The measurement process is formulated as a linear inverse problem using a correlation-peak response matrix, whose rows describe the spatial response of the correlation peaks generated at each modulation frequency. By exploiting the sparsity of reflection events, the proposed method reconstructs the reflectivity profile even when multiple correlation peaks coexist in the fiber. Experimental results demonstrate reflection-point localization over an extended measurement length, reaching approximately five times the conventional correlation-peak-spacing limit.