Recent progress in EMC and reliability for automotive applications

Cable diagnosis is a crucial research topic for the reliability of embedded systems in automotive applications. This paper presents a methodology dedicated to the reflectometry analysis of branched networks in order to localize and characterize the faults which may affect it. The direct model (propagation along the cables) is modelled by a RLCG circuit model and the Finite Difference Time Domain (FDTD) method. This model provides a simple and accurate method to simulate Time Domain Reflectometry (TDR) response. Genetic algorithms are combined with this wire propagation model to solve the inverse problem and to deduce physical information's about defects from the reflectometry response.