Rietveld Code Review Tool
Help | Bug tracker | Discussion group | Source code | Sign in
(670)

Issue 103470050: new WAVE example VANET routing compare

Can't Edit
Can't Publish+Mail
Start Review
Created:
6 years, 4 months ago by scarpen
Modified:
5 years, 11 months ago
CC:
ns-3-reviews_googlegroups.com
Visibility:
Public.

Description

new WAVE example VANET routing compare

Patch Set 1 #

Patch Set 2 : previous upload failed to include new files #

Patch Set 3 : added new files via mercurial queues commands #

Total comments: 24

Patch Set 4 : Changes per code review #

Total comments: 4

Patch Set 5 : tweaks for patchset4 #

Patch Set 6 : tweaks to patchset5 #

Total comments: 13

Patch Set 7 : additional code review changes #

Unified diffs Side-by-side diffs Delta from patch set Stats (+4724 lines, -0 lines) Patch
A src/wave/examples/low99-ct-unterstrass-1day.filt.7.adj.mov View 1 2 1 chunk +1515 lines, -0 lines 0 comments Download
A src/wave/examples/vanet-routing-compare.cc View 1 2 3 4 5 6 1 chunk +2380 lines, -0 lines 0 comments Download
M src/wave/examples/wscript View 1 chunk +4 lines, -0 lines 0 comments Download
A src/wave/helper/wave-bsm-helper.h View 1 2 3 4 5 6 1 chunk +159 lines, -0 lines 0 comments Download
A src/wave/helper/wave-bsm-helper.cc View 1 2 3 4 5 6 1 chunk +321 lines, -0 lines 0 comments Download
A src/wave/helper/wave-bsm-stats.h View 1 2 3 4 5 6 1 chunk +202 lines, -0 lines 0 comments Download
A src/wave/helper/wave-bsm-stats.cc View 1 2 3 4 5 6 1 chunk +139 lines, -0 lines 0 comments Download
M src/wave/wscript View 1 2 3 2 chunks +4 lines, -0 lines 0 comments Download

Messages

Total messages: 16
Konstantinos Katsaros
Hi Scott, This patch contains only the changes to the wscript file for WAVE examples ...
6 years, 4 months ago (2014-06-23 09:20:06 UTC) #1
scarpen
On 2014/06/23 09:20:06, Konstantinos Katsaros wrote: > Hi Scott, > > This patch contains only ...
6 years, 4 months ago (2014-06-23 20:35:46 UTC) #2
Tom Henderson
thanks for contributing this, it looks useful. it looks like it shouldn't take much more ...
6 years, 4 months ago (2014-06-25 17:25:08 UTC) #3
scarpen
Thank you for these excellent review comments. I have addressed most of them, but before ...
6 years, 3 months ago (2014-07-05 19:53:04 UTC) #4
scarpen
I have refactored and addressed the code review comments. For the DSR issues I previously ...
6 years, 3 months ago (2014-07-18 18:15:22 UTC) #5
Konstantinos Katsaros
Hi Scott, I have two comments on the BSM application. These messages are generally sent ...
6 years, 3 months ago (2014-07-22 12:10:21 UTC) #6
Konstantinos Katsaros
I can't get any BSM messages using the trace file. The goodput is on the ...
6 years, 3 months ago (2014-07-22 14:26:05 UTC) #7
scarpen
Thank you for the comments. I have been out of town for several days. I ...
6 years, 3 months ago (2014-07-28 12:51:43 UTC) #8
scarpen
Thank you for the observations. Again, I have been out of town. I will review ...
6 years, 3 months ago (2014-07-28 12:54:39 UTC) #9
scarpen
On 2014/07/28 12:51:43, scarpen wrote: > Thank you for the comments. I have been out ...
6 years, 3 months ago (2014-07-29 20:25:51 UTC) #10
Tom Henderson
In general, most of the ns-3 helper Install() methods are responsible for plumbing things together ...
6 years, 1 month ago (2014-09-06 02:02:12 UTC) #11
scarpen
Tom (et. al.) Thank you for the comments. They are truly useful. I have addressed ...
6 years, 1 month ago (2014-09-06 16:13:12 UTC) #12
Tom Henderson
On 2014/09/06 16:13:12, scarpen wrote: > Tom (et. al.) > > Thank you for the ...
6 years, 1 month ago (2014-09-09 19:25:54 UTC) #13
scarpen
Thanks again for the comments, and the code updates. Re: logging vs. verbose reporting. I ...
6 years, 1 month ago (2014-09-09 20:07:17 UTC) #14
Konstantinos Katsaros
Hi Scott, I am working on a BSM/CAM application and I plan to use your ...
5 years, 11 months ago (2014-11-27 14:34:03 UTC) #15
scarpen
5 years, 11 months ago (2014-11-27 20:13:18 UTC) #16
Thank you for the questions.

1.  The use of r-squared instead of just r is merely a performance
optimization.  Consider Euclidean distance between two points, R, s.t. R =
sqrt ((x2-x1)^2 + (y2-y1)^2).  Then R^2 = (x2-x1)^2 + (y2-y1)^2.  Sqrt is
generally a complex operation.  Thus, since I need to do so many
comparisons to see if the distances between two points are within a certain
"range", I simply pre-compute the distance-square = R^2, and then compare
that to the sum of the squares for the two points.  Just saves a
square-root operation with every compare.  This is just a performance
optimization, and should be "hidden" from the user.

2. I believe the use of the index-1 is simply a 0-based vs. 1-based issue.
That is, we want to have a series of distances.  E.g., the first distance,
second distance, etc.  Give them an index = [1, .. n].  Now the actual
values are stored in a 0-based array.  So, to find the I-th indexed value
in the array, use (I-1) as the array index.  It just seemed more intuitive
to me to use index=1 to mean the "first" distance that I am interested in.
Storing it in a 0-based array is an internal implementation decision.

Please let me know if there are further questions, or if I have failed to
address your concerns.

-Scott

On Thu, Nov 27, 2014 at 9:34 AM, <dinos.katsaros@gmail.com> wrote:

> Hi Scott,
>
> I am working on a BSM/CAM application and I plan to use your stats
> helper.
> I have a question regarding the txSafetyRangeSq element. Why do you use
> the distance squared and not as it is?
>
> with the default
>   m_txSafetyRangesSq[0] = 50.0;
>   m_txSafetyRangesSq[1] = 100.0;
>   m_txSafetyRangesSq[2] = 200.0;...
>
> The used would assume that this is the different radius, but in fact it
> is the radius squared.
>
> Also, this part looks a bit strange. (when calculating the expected Rx)
>
> ...
> int rangeCount = m_txSafetyRangesSq.size ();
> for (int index = 1; index <= rangeCount; index++)
> {
>    if (distSq <= m_txSafetyRangesSq[index - 1])
>    {
>    // we should expect dest node to receive broadcast pkt
>    m_waveBsmStats.IncExpectedRxPktCount (index);
>    }
> }
>
> I understand you want to identify the expected receivers in each area.
> If the distance (distSq) is less than or equal to the (index-1) then you
> increase the index counter.
> e.g. if distSq = 40 (the actual distance of the two nodes is sqrt(40) =
> 6.3) and index = 1, the expected Rx to be increased should be the
> index=0 (for 50). Which is basically what you do in the BSM stats. You
> actually increase the index-1.
>
> void
> WaveBsmStats::IncExpectedRxPktCount (int index)
> {
>   m_wavePktExpectedReceiveCounts[index - 1]++;
> }
>
> My point here is that this (index-1) that you use while you require the
> index, looks confusing.
> Is there any particular reason?
>
> https://codereview.appspot.com/103470050/
>



-- 

Best,
Scott Carpenter
Graduate student (PhD track), NCSU Department of Computer Science
scarpen@ncsu.edu
919-413-5083
Sign in to reply to this message.

Powered by Google App Engine
RSS Feeds Recent Issues | This issue
This is Rietveld f62528b