Research on Longitudinal Control Algorithm of Adaptive Cruise Control System for Pure Electric Vehicles

Round 1

Reviewer 1 Report

This article presents an interesting approach to address solving non-linear control in EV. CNN is adopted and integrated in the control loop along with PID control scheme. The results are verified by a couple of simple responses assumed on the road.

 

Xc in equation from 5 to 7 is not defined.

 

The main purpose of this article is to solve nonlinear control seen in the ACC system of an electric vehicles. A EV model figure has been shown. However, it is not clear how serious a problem that nonlinearity appears in this vehicle dynamics. Most literatures are also related to CI vehicle models, and hence more may be included in the introduction section to address EV dynamics and control problems.  

 

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

The title of the work refers to the study of a system for an electric vehicle. At the same time, the authors do not consider the processes in the electric drive system of an electric vehicle.

Would be appropriate to formulate the novelty of the article material and  obtained results in more detail

What are the advantages, compared to FFNN, received by application of RBNN to obtain the Jacobian for implementing the algorithm for finding the changes of increment PID controller coefficients?

Would be appropriate to indicate the change in the coefficients of the incremental PID controller and the settings of the traditional PID controller in driving and braking regimes. 

What explains the acceleration fluctuations after 5 sec (fig. 9)?

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

The paper focuses on an adaptive cruise control system for electric vehicle applications. In particular, the paper asses a comparison between a traditional PID controller and an adaptive PID exploiting an RBFNN for online parameter tuning. The topic is quite interesting and actual. However, before suggesting the manuscript for publication, the following point should be addressed.

-      Please, better highlight in the introduction and conclusion the novelty of the work and its incremental contribution to the scientific literature.

-      Please avoid bulk referencing as [1-3], [4-6] etc... . It is suggested to use only one, or if all are relevant, explain their point of view or differences. Moreover, check the references. As an example " The specific modeling process of the vehicle model will not be discussed here [24]." at rows 129-130 suggest that the used model has been previously discussed in other work, but the linked references are from different authors and don't use carsim.

-      Please clarify the model section. It is not clear if there is some hardware as suggested by " A pedal simulator equipped on the vehicle is used to simulate 122 pedal feeling". It is suggested to revise the whole section to avoid possible misunderstandings and improve clarity. Moreover, a table reporting the main assumed vehicle parameters can be added to improve the reproducibility of the results.

-      Please improve the discussion on the development process of the RBFNN. How has it been trained? What data have been used for it?

-      Is the controller's sampling rate the same as traditional and RBFNN-augmented PID controllers? It can be useful to report the adopted value and the constant tuning parameters of the traditional PID. 

-      There are some minor grammatical and stylistic issues with the text. For example, in some phrases, there are some repetitions which can be avoided. " The specific modelling process of the vehicle model will not be discussed here [24]." can be rewritten as " The specific vehicle model development process will not be discussed here [24]." Avoiding modeling/model repetition.

-      The number of figures is relatively high, particularly in relation to their data contents. It can probably be helpful to merge some graphs using a multi y-axis or stacked charts. As an example, 7 and 8, 9 and 10, 11 and 12, and so on, can be merged as described. 

-      It can be interesting, if it is possible, to add the variation of the PID tuning parameter due to the use of RBFNN, compared to the constant set of traditional PID.

 

-      It can be useful to introduce a summary table reporting the main performances, such as maximum error and settling times, for all the tested conditions. 

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 3 Report

Thanks tha authors for the great revision work. Now the quality of the work presentation is improved significantly. 

Author Response

We appreciate for the time and effort that you have put into reviewing the manuscript. Your suggestions have enabled us to improve our work. We have benefited a lot through your guidance. We have learned more professional knowledge and writing experience in our communication with you. We hope to have more opportunities to discuss academic issues with you in the future. Finally, thank you again for your guidance!