Multivariate dimensional accuracy data analysis in automobile assembly: Modeling, reduction and quality control issues

The problem of dimensional accuracy data analysis in automobile assembly operations is considered. The data are measured via Optical Coordinate Measuring Machines (OCMM´s), and issues of multivariate modeling, analysis, reduction (sensor number and location determination) and quality control are addressed. The study is based upon a novel multivariate time series analysis framework that accounts for both spatial cross correlation and serial autocorrelation. Its effectiveness is demonstrated via actual and simulated plant data.