Signal-Integrity Optimization for Complicated Multiple-Input Multiple-Output Networks Based on Data Mining of S-Parameters

In this paper, an efficient signal-integrity analysis and optimization method for complicated multiple-input multiple-output (MIMO) networks is proposed, in which data mining is applied to discover the concealed information in black-box S-parameter models. Instead of performing a number of circuit simulations, data mining employs mathematical search algorithm directly into the model, which can save significant analyzed time and improve the efficiency. An optimized mining flow is presented for large scale data set, where the data processing and data mining are performed simultaneously, and thus it is unnecessary to save large extracted data. The proposed data mining method consider both the interconnect structure and the stimulated pattern of a MIMO system, which can perform thoughtful analysis and optimization with high efficiency. Two examples, signal-integrity analysis of two coupled microstrip lines and noise coupling among multiple signal vias through a power-ground plane pair, are presented to demonstrate the efficiency of the proposed data mining method in signal-integrity optimization design.