Re: [sumo-user] Collisions during simulation

[Jakob E. <nam...@go...>]

Dear Pietro,
yes, using the latest version of sumo and your network in the new format i
still see collisions.
Using the default seed i get 3 and using different seeds i get between 2-5.
These collisions are indicative of some kind of bug and I'm looking into
that right now.

regards,
Jakob

2011/11/16 Pietro Niccoli <pi...@ti...>

> Dear Jacob,****
>
> I downloaded the last svn version of sumo and the wrong collisions have
> now disappeared, but some ****
>
> collisions (4) still happen. I checked and this time it’s because two
> vehicles actually crash. So probably ****
>
> there still is a little bug lying around.****
>
> I’m using the same map I gave you, converted to the new format. Did you
> experience those collisions****
>
> during your tests?****
>
> Regards,****
>
> ** **
>
> Pietro****
>
> ** **
>
> *Da:* Jakob Erdmann [mailto:nam...@go...]
> *Inviato:* giovedì 10 novembre 2011 14:56
> *A:* Pietro Niccoli
> *Cc:* sum...@li...
> *Oggetto:* Re: [sumo-user] R: arrivalPos option****
>
> ** **
>
> Dear Pietro,
> thanks for your example.
> I traced the invalid "collision" down and it should be fixed now.
> This may well have been a major source of generated "collision" warnings.
>
> I also noticed that your network was using the old format. You can easily
> convert it with the following call:
> netconvert -s old.net.xml -o new.net.xml
> Not only does get rid of deprecation warnings it also reduces network size
> by ~30%.
>
> regards,
> Jakob****
>
> 2011/11/9 Pietro Niccoli <pi...@ti...>****
>
> Sure Jakob, here you go. ****
>
> This is a map of Torino, Italy. I know it’s a little bit messed up, but it
> has been tweaked and modified for a long time.****
>
> I used the last version of sumo gui. If you need anything else, just send
> me an email.****
>
> Best regards, ****
>
> Pietro****
>
> *Da:* Jakob Erdmann [mailto:nam...@go...]
> *Inviato:* mercoledì 9 novembre 2011 13:08
> *A:* Pietro Niccoli
> *Cc:* sum...@li...
> *Oggetto:* Re: [sumo-user] R: arrivalPos option****
>
> Hello Pietro,
> I'm having a hard time reproducing this behaviour. Could you please send
> an example network and route (preferebly small).
> regards,
> Jakob****
>
> 2011/11/9 Pietro Niccoli <pi...@ti...>****
>
> Hello,
> I updated to version 0.13 and I rewrote the parameter, and now it works.
> But I found a little bug (supposedly). I run a simulation with a map of
> mine, and if a vehicle has a short road as a destination (arrivalPos is set
> to -0.1), when it exits sumo reports a wrong teleport and collision:
>
> Warning: Teleporting vehicle '1188'; collision, lane='-543232277_0',
> time=291.00.
>
> In this case, the road -543232277 is 6 meters long, the car was travelling
> at 7,88 m/s and at step 291 it was positioned at 1,08 m. in the destination
> road. In the following step, the car arrives and disappears, but sumo
> prints
> the above error message. There were absolutely no other cars around.
> My understanding is that since the car was travelling faster than the
> remaining length, this is recognized as a wrong collision. Does it make any
> sense?
> Thanks for your help!
>
> Pietro****
>
> ** **
>