Hi everyone,

This is a little on the lengthy side of things, I hope you don't mind.

I'm currently working on a traffic simulation for a commercial project. The
validity of the simulation is important, but in itself no metrics are
required; it serves only as a visual aid to challenge the product's users
when they complete various tasks.

Prior to this moment, I have been working with a multi-lane simulation
which did not include any off/on ramps or more complicated traffic systems.
At this point, I moved to a "node based" model, which analyses geometry of
a 3D mesh (using some helper information) to produce a graph of connected
nodes.

It has only been when writing the simulation for this graph that I realise
how much more complicated the simulation must become. The list below
indicates my new challenges:

   1. Ramp merges affect the vehicles in the lane as well as those which
   are about to merge. Unlike lane merging, which is optional and only
   performed if the impact is minimal, vehicles on ramps prefer to maintain
   velocity, and vehicles behind them in the new lane will slow down somewhat.
   This changes the purist IDM model.
   2. Routes require some level of planning in order to determine how far
   to the nearest vehicle. Whenever the vehicle ahead is on a separate lane
   (in the case of an on ramp, or a "decision" node where one lane can go
   either of two directions), we must investigate this and recall the path we
   took, so that our behavior corresponds with the vehicle we're following. We
   cannot consider artificially congesting routes unless we force a percentage
   of vehicles to follow those routes.
   3. Road representation. I'm using polygons, but this creates its own
   challenge - using portals as midpoints along polygon edges for smooth lane
   changes and ramp exit/entrance. Long story short, being able to work out
   which vehicle is ahead of which is difficult when you're working with such
   flexible implementations.

These are not the only problems I have or will continue to face, but I
notice that SUMO is a mature (13 years?) simulation framework and I hoped
to ask a few higher order questions.


   1. How does SUMO handle ramps? Both in terms of the linear acceleration
   models of the vehicles involved and in the planning stage.
   2. How does SUMO represent the vehicle in terms of its position on the
   road, and then convert to/from 2D cartesian coordinates?
   3. Would it be possible to ask SUMO (with the Python interface) to spawn
   vehicles in specific positions at a user-defined interval, perhaps
   irregularly? I suspect that given other requirements of the simulation, I'd
   need to write my own Python bindings for certain things.
   4. Is there any documentation on the overarching structure of the
   simulation design?


I'm sure that I will have more questions, but any help at this stage would
be great,
kind regards,
Angus Hollands.
------------------------------------------------------------------------------
Download BIRT iHub F-Type - The Free Enterprise-Grade BIRT Server
from Actuate! Instantly Supercharge Your Business Reports and Dashboards
with Interactivity, Sharing, Native Excel Exports, App Integration & more
Get technology previously reserved for billion-dollar corporations, FREE
http://pubads.g.doubleclick.net/gampad/clk?id=164703151&iu=/4140/ostg.clktrk
_______________________________________________
sumo-user mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/sumo-user

Reply via email to