DRCKNN

Detecting and Locating Cable Faults with hybrid AI

This project introduces an innovative diagnostic method for cable networks, combining Time-Domain Reflectometry (TDR), the K-Nearest Neighbors (KNN) classifier, and Dynamic Time Warping (DTW). The approach is divided into two phases: an offline phase, where numerical simulations (FDTD) generate training data for classification and regression models; and an online phase, where these models are used to detect, locate, and characterize faults in tested networks. Both numerical and experimental results will demonstrate the effectiveness and feasibility of this method.