The aim is to develop the theory of rank estimation for the regression parameters in measurement error models. R-estimators are studied as robust alternatives to least squares estimators.