Optimal Scheduling of Time-Sensitive Networks for Automotive Ethernet Based on Genetic Algorithm
Round 1
Reviewer 1 Report
First of all, I would like to say the paper fits journal topics. It is devoted to solving an important problem that concerns traffic scheduling with the optimal policy for a time-synchronized autonomous driving network. The mathematical background of the problem solving is based on a genetic algorithm.
Strong sides are:
- the paper is well structured,
- the problem has novelty and the proposed solution is well suited to the task,
- the proposed solution seems to be practically realized,
However, I would like to give some comments to improve the paper's readability. Comments have minor influence, but I think the authors can include them in the final version of their manuscript:
- to clarify a definition LCM in Eq 4, line 195,
- check the numbering, because 3.1.1 is after 3.2,
- to clarify abbreviations in Fig. 8,
- to explain what is "randomly created chromosomes" in line 414,
- in Conclusion, is better to explain the requirements for practical implementation of the proposed approach.
Author Response
Point 1: To clarify a definition LCM in Eq 4, line 195
Response 1: LCM means the least common multiple, and this is used to calculated the hyper period (hyper.T) which is the shortest period of all traffics which are repeatably transmitted.
To reflect your comments on the point 1, I have added “LCM” in line 188.
Point 2: check the numbering, because 3.1.1 is after 3.2
Response 2: To reflect your comments on the point 2, I have corrected “3.1.1” to “3.2.1” in line 218.
Point 3: to clarify abbreviations in Fig. 8
Response 3: To reflect your comments on the point 3, I have added the original words of abbreviations in Fig. 8.
Point 4: to explain what is “randomly created chromosomes” in line 414
Response 4: We generated 1000 chromosome samples to find the normal distribution of each indicator, and genes are randomly placed in generated chromosomes to represent different schedules. We have confirmed that chromosome samples are represented as a normal distribution, and the indicators are converted to z-score which is used as parameters for fitness function.
I have extended the 4th paragraph of the Experiments section to incorporate your comment.
Point 5: in Conclusion, is better to explain the requirements for practical implementation of the proposed approach
Response 5: To practical implementation of the proposed scheduling method, various components (sensors, ECUs, and switch etc.) which are supported by hardware for functions of IEEE 802.1Qbv standard related schedule. Unfortunately, using supported components is quite limited because they are rarely found in the currnet market. We expect the componets that provide the technology specified in IEEE802.1Qbv to be released soon and available. And we plan to evaluate the impact of improved network performance on the vehicle stability through the proposed scheduling method in the future.
To reflect your comments on the point 5, I have extended the last paragraph of Conclusion section.
Thank you very much for your valuable comments.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
The proposed schedule optimization method could be possible applied for an autonomous commercial vehicle. Therefore the standard parameters and related data should be used for comparing the results. Also, some new schedule optimization methods not for all (E2E,..) but some objects should be shown in the paper.
Author Response
Point 1: The proposed schedule optimization method could be possible applied for an autonomous commercial vehicle. Therefore the standard parameters and related data should be used for comparing the results. Also, some new schedule optimization methods not for all (E2E,..) but some objects should be shown in the paper.
Response 1: We have proposed a traffic scheduling method to improve network performance in applications such as autonomous vehicle where an amount of significant of data is exchanged. In this paper, the proposed schedule method is based on IEEE 802.1Qbv, one of Ethernet-related standards of IEEE, and requires hardware-based support for related standards in sensors, ECUs, and switches. Unfortunately, comparison with related data and standard parameters is difficult because the vehicle supported IEEE 802.1Qbv is rarely available at the moment. We expect the components that support schedule functions to be distributed on the market soon, and we plan to evaluate other objectives, such as the effect of the proposed method on vehicle stability in the future.
I have extended the last paragraph of Conclusion section to incorporate your comment.
Author Response File: Author Response.pdf
