Due to the great importance of defining performance, reliability and safety requirements for advanced composite products and services a considerable effort has been devoted to the study of their mechanical material properties and many inverse methods based on vibration tests have been developed in the last three decades. In each inverse technique based on vibration tests, an identification of material properties is performed minimising the error functional between the experimental and numerical parameters of structural responses using direct or indirect, based on the planning of experiments, optimisation methodologies. Unfortunately the dynamic parameters of test specimens are very sensitive to experimental errors and numerical models used in the identification procedure due to an approximated nature of the numerical solution or measurement inaccuracies. So, the experimental modal parameters of a structure and correspondingly the identified material properties are mostly disturbed by the errors due to poorly simulated boundary conditions, additional mass and damping effects from exciting or sensing devices and measurement noise influence. These are systematic errors arising from effects that interrupt the measurement results in a systematic way. They originate from the known defects of instrumentation, the wrong measurement procedure or imperfections of the test samples.