Multiple Delay Identification in Long Interconnects via LS-SVM Regression

This work presents a novel approach for the accurate estimation of multiple time-delays from frequency response distributed system. The proposed is based on powerful and flexible machine learning technique, namely, least-square support vector (LS-SVM). LS-SVM regression used to construct metamodel transfer function describing generic linear time-invariant system in delayed-rational form. Specifically, after some manipulation model precisely identifies dominant propagation delays original essential steps critical criteria delay identification procedure are carefully discussed throughout paper. Once have been identified, rational part expansion then obtained by means progressive application conventional fitting algorithm. Numerical examples presented illustrate feasibility performance technique compare its performances with what provided state-of-the-art techniques. results clearly highlight capability identify systems, thus allowing compact delayed models.