Least-Squares Estimates of Automotive-Drag Coefficients from Coast-Down Tests

Determining vehicle drag coefficients from experimental data is an important problem for engineers working to improve automotive fuel economy. The least-squares method offers a convenient way to compute the coefficients, however, the method uses a badly conditioned matrix, condition number = 3 × 10⁵, so that the computed coefficients are sensitive to measurement errors. Examination of sensitivity to velocity-measurement errors indicates that the coefficients can be computed with sufficient accuracy for certain applications. Emphasis of several unresolved issues suggests directions for further investigation.