Multivariable control based on incomplete models via feedback linearization and continuous‐time derivative estimation