Optimal design of sequential experiments for error-in-variables models