The paper presents the design of a controller for an EGR system to be applied on a turbocharged diesel engine, suitable for transient operation. The controller is developed and validated in Matlab-Simulink and GT-Power, while is finally tested on an engine. Though the subject is interesting, even if EGR control is a topic with a long history, the manuscript has to be enhanced in order to be accepted.

In particular: - Which is the final goal of the development of a new EGR controller, that is, which types of engine is the controller suitable for? This aspect is fundamental, because a number of papers are available for EGR systems control in the automotive field (see for example Shutty, Benali, Daeubler, Traver, Air system control for advanced diesel engines, SAE paper 2007-01-0970) and many problems have to be considered when controlling EGR system, first of all the interaction with the turbocharger, so it should be explained why a “simple” controller is required. - The model description is quite long, if referred to the overall manuscript length, usually it can be included in an Appendix section. - No information is given on the tested engine, while it should be specified if the controller is suitable for a high pressure EGR system only or it can be also applied to a low pressure EGR system (or even to both). - Equations 13 and 14 – line 130: four efficiencies are presented in equations, while only two are explained in the text. - Coefficients for matrixes A and B are obtained for one operating condition: what does it happen if this condition is changed in a significant way in engine speed, load and/or EGR rate? A similar remark: are the considered transient tests applied for the experimental validation sufficient to validate the designed controller? What if intake oxygen mass fraction is significantly lower than 21.3%, which is the minimum level considered in the experimental validation? - Lines 225-230: considerations on linear model are somewhat obvious. - There are a number of statements which are difficult to read and need to be rephrased, but “fuel is being completely combustion” (line 344 and 356) have to be absolutely avoided (for its English, not for the assumption). - Please check values in Table 2: error in the last line is wrong (1.23%, not 0.24%). - I don’t agree with the statement in lines 368-369: “… it is easy to implement for (probably authors mean four) other sensors in engineering practice”. Due to cost increase, each sensor to be added is subject to a strong discussion with engine manufacturers.