Frequency-domain experiments (FDEs) have been considered for experiments when errors are correlated and time-lagged responses may exist. Although to date the application of FDEs has been largely restricted to simulation meta-model identification, this approach is appropriate in many manufacturing settings. In much industrial experimentation, limited equipment or operator availability may mandate the use of sequential data collection. Analysis techniques which deal with correlated errors translate into shorter required data collection times, as do those which allow for the study of time-lagged responses. This research effort will combine the philosophy of robust product design with the methodology of frequency-domain experiments. Selection approaches will then be used to identify desirable product design configurations. The resulting experimental design and analysis methodology should be highly suited to the automated manufacturing environment, where the assumptions required for traditional experimental design and analysis may not be applicable. Additionally, in order to ensure the applicability of the work and facilitate the technological transfer, visits will be conducted to manufacturing firms currently employing off-line quality design or those with automated data collection capabilities. These plant visits will allow the investigators to ascertain the issues most critical to the implementation of this research.
