Investigating and Modeling of Cooperative Vehicle-to-Vehicle Safety Stopping Distance

Round 1

Reviewer 1 Report

General comments
----------------
The paper presents a system to reduce the stopping distances based on DSRC, with a self-developed testbed.
I have a question about the research in order to maintain safe distances when driving. Is the size of the vehicles taken into account? As if only the distance between the communication devices is taken into account, the distance between vehicles will be very different depending on the type of vehicle: car, lorries, etc.

Some Typos or suggestions about the manuscript
----------------------------------------------
The manuscript is well written, but some errata have been found.

Missing blank space in many places along the text:
- Missing blank space before '(' (like in lines 2, 33, 34, 71, 95, 106, 153 twice, 183 twice, after gravity, 342, 345, 354 just before Eq.9, 358, 365 after 'Deceleration' twice followed by a misplaced comma), the most of them while defining acronyms.
- Missing blank space before '[' (like in lines 21, 31, 32, 34, 36, etc.), the most of them before a reference.
- Missing blank space after dot '.' (74, 93, 340, 414 "30mph.this is shown" -> "30mph. This is shown", 488)
- Missing blank space: "Fig.16as" -> "Fig.16 as" (line 451)

- Duplicated dot ".." -> "." (line 362)

- Missing commas ",": "Finally section VIII" -> "Finally, section VIII" (line 64); "The testbed although not the focus of the paper will be briefly explained." -> "The testbed, although not is the focus of the paper, will be briefly explained." (Line 149); "In this work 1.5 seconds" -> "In this work, 1.5 seconds" (line 187).

- Missing dot ".", ";", ":": "expensive[16], to counter" -> "expensive[16]. To counter" (line 81); "in Eq.1 The Round-Trip Time" -> "in Eq.1. The Round-Trip Time" (line 178); "the 3 variables are" -> "the 3 variables are:" (line 255); "measured We applied" -> "measured. We applied" (line 277)

- In the outline (lines 63-64), the word "combined" seems to be duplicated in the same statement.
- Uppercase: "many researchers" -> "Many researchers" (line 117); "30mph.this is shown" -> "30mph. This is shown" (line 448)
- "oprimary reason" -> "primary reasons" (line 123)
- "reaction time(Safety Distance(Ds)" -> "reaction time (Safety Distance or Ds)" (line 354)

- Table 4: there is a value with only 4 decimal places, whereas the other use 5 decimal places; similarly in Table 8 (1 decimal place is the most used but some does not use any).
- Table 5: "Lost(Consecutive" -> "Lost (Consecutive)" (space and right parenthesis)

- Acronyms not defined at their first use: WHO (line 32); DSRC (in line 33, although it is defined also in the abstract); "Software Defined Radios (SDR)" (line 40); V2I (105); ADAS (line 117, although it is defined in line 350);  

- About equations 1 and 2, labeled as Eq.1-3 and Eq.4-5, respectively, I suppose that it would be possible to several label both equations using only one number: Eq.(2) and Eq.(3). In this way it could be simplified "shown in Eq.2, Eq.3 and Eq.4," (line 183) as "shown in Eq.2"; the same to "shown in Eq.5 and Eq.6" -> "shown in Eq.3".
- Eq.8: seems that several parenthesis could be removed.

About figures
-------------
The figures show results very clearly.

About references/bibliography
-----------------------------
- Bibliography is very complete and exhaustive.
- A reference is needed: "6.44 m/s2[ref]." (line 373)
- [25] There blank spaces in the title "... 802 . 11a and 802 . 11p ..."

Author Response

General Comments.

I have additionally explained the use of vehicle distances and made a point to state the use of vehicle length is altered depending on the vehicle being driven.

Typo Fixes

All mentioned typos have been fixed following the comments given.

References

Added a missing reference needed. Line 363

Corrected incorrect spacing in reference 25.

Reviewer 2 Report

The paper analyses the use of DSRC communications to reduce the stopping distance in situations in which a vehicle follows another vehicle that brakes. The impact on stopping distance of reliability in the DSRC communications is also analysed. The analysis is based on results from field tests.

The topic is important. Knowing the impact of reliability on the performance of collaborative driving systems is critical to understand the requirements for communication solutions in those systems.

Unfortunately, the paper is at times hard to understand, and it is not easy to find out what really was done. Detailed comments below.

1. The paper states in the introduction: "we deem V2V to be the appropriate choice, hence DSRC is the technology being used [9]." But this is not necessarily the case, C-V2X also offers V2V communications using PC5 (sidelink) interface.
2. In section 3 the paper claims "Our testbed was designed... to be capable of operating to DSRC standards with the full protocol stack". It should be clarified if this means the access technology stack or the full DSRC protocol stack (access technology, networking & transport, facilities). For example, the tests use UDP with safety applications which is not the standard approach.
3. I do not understand equation 1. What is (T_R-T_S)? The paper states that the RTT was measured and that the propagation time was added, where is this reflected in equation 1?
4. The paper should not use MPH as the units of the speed of vehicles, in particular since the paper uses meters to measure distances. Is your distance unit is meters, use km/h or m/s as speed units.
5. Why two different formulas for calculating Braking Distance are needed? Why "thinking distance" contribution is not calculated?
6. In section 4.2. the paper, to justify the use of UDP, states: "This also allows the communication to occur faster as no relationship needs to be built between the nodes eliminating the time consuming process of the handshake procedure." Since CAMs/BSMs should be broadcast, you couldn't use TCP in any case.
7. In Equation 7 the term [1/(2gu)]  would result in units of s^2/m, while m should be expected, so something appears to be wrong.
8. Equation 8 requires a better explanation.
9. In section 5.1 the paper states: "Fig.5 and the figure used to show our distribution of losses shows that most consecutive losses are from 0 to 11 per 1000 packets" In figure 5 I can see the distribution of packet losses when sending 1000 packets but, what information is there about consecutive packet losses? There is also a problem with the X axis, the numbers do not match the bars.
10. In section 5.1 the paper states: "Most of our tests yielded a packet loss rate of 0 packets per 1000" However, Figure 5 does not show that, two losses seems to be the most common case.
11. I do not understand Table 6. I would expect a reference, for example the safety distance for vehicles at certain speed. Note that the paper uses safety distance in section 6. But I do not understand what is the reference distance in Table 6 and how you conclude that a negative value means that there is a collision.
12. Section 5.3 states: "communication, as at each instance of 10pps for every speed from 30-60 mph a negative value is produced indicating the crash would have already occurred and in the case of 70mph the distance reduction would be significantly reduced." Where do you see this result?
13. In section 5.3 your analysis is that more speed requires less packet sending rate and "this is due to the stopping distance being smaller", but one would think that more speed should require larger stopping distance and, therefore, more packets per second to allow a faster reaction time. 
14. In section 6.1 the paper mentions geocasting, but CAMs/BSM are broadcast.
15. In section 6.1 the paper states: "Braking time is not considered in safety distance as it is essentially a distance of space left between moving vehicles in order to give a driver enough time to react to emerging situations." This is obviously not safe, and is the reason of chain-reaction crashes in roads.
16. As already mentioned, I do not understand why the "safety distance algorithm" calculated in section 6.2 is not used also in section 5.
17. I do not really understand how the measurements are obtained in section 6 (for example, the blue points in Figure 16). You mention simulations, but what tools? which experiments?
18. Many typos all around the paper (missing spaces, "to solves", "oprimary", "al;so", "netqwork", "metghod"...

Author Response

Hello, Thanks for your review, I have corrected all of the issues you found and attempted to clear up the misunderstanding of items I had missed.

1. Corrected the assumptions about DSRC and C-V2X. Line 28-29
2. Clarified the type of protocol stack and the use of UDP. Line 127
3. Changed the equation to better represent the meaning and adjusted the explanation. Equation 1
4. Have added various indicators to both mph and km/h throughout the paper.
5. Calculation for thinking distance is now included and equations have been altered. (equation 2-3). We chose to show the braking distance differences between the standard method and a method used with more detail that gets more accurate results.
6. Corrected the explanation and removed unnecessary information. Line 196-197
7. Equation 7 has been corrected due to missing variable. Same for equation 8. These are now equation 4 and 5.
8. Better explained the use of equation 8. Between 204-205
9. Corrected the representation of 5 now fig. 3 and explained the representation clearly.
10. Explained the results shown in figure 3 Line 219-230
11. Explained the reason for table 5 and highlighted the implications of how the negative value could be interpreted. Line 236-243
12. Adequate explanation to show how the conclusion is drawn and where it is drawn from. 4-5
13. Altered the explanation and fixed the analysis. Line 289-295
14. Removed this sentence.
15. Explained the reason for the omittance of braking distance and what the section represents. Line 332-344
16. We only show the use of algorithms in section 6 and show that ADAS performs better, in section 5 the ADAS is the only method used.
17. Explained the experiments and measurements used in section 6, highlighting the tools used and experiments conducted. Section 6
18. Corrected the mentioned typos and many others.Paper has also been shortened.

Reviewer 3 Report

In this work, authors present a study about how stopping distance can be reducen usin DSRC communication in the scope of autonomous vehicles. The paper is an extension of a previously published work of the same authors.

Although the topic could be of interest, the structure and density of the paper makes difficult to measure the real contribution of this work. There are many grammar and typos mistakes in the text that must be fixed. Please, make use of proffesional proof reading services in order to improve the readability of the paper.

In section 4, communication assisted stopping distance, authors present 2 different formulas to measure the stopping distance. However, I do not really understand why authors use the first one in section 4.2 being the second formula more accurate. 

Regarding simmulation tests, in figures 9 and 10, authors must explain the unexpected behavior of curves 140pps, and 120 pps at distances of 30m and also at 60-90 meters (in figure 9). Also in figure 10, with 10pps at 80 meters something should explain this behavior. 

In subsection 5.3, in the second paragraph, authors claim that "As expected, our results show that at all speeds that consecutive packet loss leads to a loss in the stopping distance. We can state the assumption that after the first packet is received after the losses the vehicle would be able to adjust its position in motorway scenarios, however on road scenarios in some cases this would lead to unavoidable crashes." But, where this sentence is supported by the presented results???

In my opinion, the paper should be shortened and it would be more focused, because from section 6 onwards, the reader have lost the real intent of the paper and the experiments.

 

 

 

 

Author Response

Grammar and typos have been corrected after being proofread by 2 of the authors.

In section 4, a correction has been made to utilise the second and more accurate formula in the results. Line 205-211

Figures have been properly explained along with a reference to a further graph to emphasise our finding. Figures 7-10 and Figure 11, Line 304-311

Removed this assumption and edited content to represent results. Line 275-279

Manuscript has also been shortened by almost 5 pages and removed unnecessary content and figures.

Round 2

Reviewer 2 Report

I appreciate the effort that the authors have made to improve their manuscript. I still believe that some points require further clarification:

• Equation (1). Since we are discussing distances of, at most, hundreds of meters, propagation delay should be negligible (a few microseconds).
• Line 158, Fig. 3 is mentioned, but it should be Fig. 1.
• It is a little bit confusing that Thinking Distance and Reaction Time are defined as two separate things, and then in equation 2, D_R is called "Reaction or Thinking distance".
• The key remaining problem is that I still do not understand equations 4 and 5. The first point is that, shouldn't the round trip latency be one-way latency? But, more importantly, equation (4) tries to calculate "remaining stopping distance". The first part of the equation is the normal stopping distance without losses. The second part (after the minus sign) is the distance traveled while packets are being lost but, after that, the car still has to stop (if you want to calculate the difference between the stopping distance with and without losses). Equation (5) is even more difficult to understand. For example, in equation (5), larger L_PT results in larger remaining distance, which makes no sense. Related to this , Table 5 shows values for remaining stopping distance calculated with equation (5) and (4), and the values are lower for equation (5), which does not match the equations (equation 5 adds a positive quantity to equation (4). Also, with packet losses equal to 0, equations (4) and (5) give different results, which should not be the case. 
• Still some typos: line 207: "We"; line 252: "we, are measuring"; Fig. 11 is mentioned before Figures 8, 9, and 10...

Author Response

1. Equation 1, The propagation delay is negligible 33 microseconds, I have added this information.
2. Adjusted fig reference to fig 1.
3. Have adjusted this line and description to reduce confusion. Lines starting 160                                    Adjusted equation 2 to represent correct information.
4. I see the confusion and misunderstanding here. I have changed both formulas and added a missing parenthesis to equation 5 as there should not have been a positive quantity. I have also adjusted the values in the table as the first was indeed incorrect. Further to this I have added a paragraph to explain the equations in line 203-223.

“””” We have chosen to use the second model as this paper is focused on the impacts of how packet loss can alter the stopping distance when relying on communications and this equation highlights that impact. The model works by utilising Eq.3 to find the normalised stopping distance for a vehicle and then we subtract the distance lost via consecutive losses. The losses can be seen in the second part of Eq.5. The result of this equation will leave us with the distance remaining to a collision or stopping distance remaining. We know that after losing the consecutive packets the vehicle will still need to stop in order to show the full distance required to stop, the Eq.5 sign could switch. This would then represent the sopping distance plus the distance lost through consecutive loss, to leave us with total stopping distance. In our case, we are showing how lead and following autonomous vehicles communicate, with the following vehicle adhering to a communication assisted stopping distance. We then analyse how this stopping distance is impacted with the addition of consecutive packet loss and this is shown for different speeds and data rates. For our results we show the stopping distance when 0 packets are lost, and we then show how the stopping distance is reduced with each consecutive loss. This highlights how the reduction in reliability leads to reduced stopping distance and hence a higher chance of collisions with the lead vehicle. As the reduction value reaches 0 we deem this to be the point at which lead and follow vehicle will be occupying the same space or that a collision may have occurred. We also show a negative value which would mean the follow vehicle has gone past the lead vehicle and this would be classified as a collision and we show this to identify the impact of reliability.””””

5. Corrected mentioned typos and further.

Reviewer 3 Report

Authors have addressed the majority of reviewer's concerns and now the readability of the paper have been improved.

Author Response

Thank you for your review and comments.

Round 3

Reviewer 2 Report

The authors have addressed my previous concerns. I have no further comments, except that I do not see the relevance of the added references in this last version of the paper (references [59] and [60]).

Author Response

Thank you for your comments,

I have removed the references.