From: Björn H. <hen...@ib...> - 2011-09-27 08:56:33
|
Hi Daniel, > 2011/9/22 Björn Hendriks <hen...@ib...> > > Depending on the flows inserted into the highway and the on-ramp I get > > some traffic jams. But, they are always limited to the right lane of the > > highway while cars on the left lane pass the traffic jam with, for > > example, 150 km/h > > (passenger cars may go with up to 50 m/s in my configuration). On Monday 26 September 2011 10:53:22 wrote Daniel Krajzewicz: > what I think to have observed so far is that vehicles on the jammed lane > get stuck because the others leave no room for a safe lane-change. The I was afraid that this is the reason. > quick workaround is to place a variable speed sign on the neighbor lane > and force vehicles to slow down - you probably can do this via TraCI, too. > This works in most cases, but is not really pretty as the speed at the > whole lane is reduced and it's quite clumsy to get a good position of the > jam and good speeds, too. > Probably better would be to let vehicles slow down automatically at > neighbor lanes, either always if their neighbor speed is low - but this > would probably yield in some strange behaviour, as for example, an > off-ramp may have a jam which does not affect neighbor vehicles. Or slow > down if vehicles try to get to the lane - this is happening to a degree, > but obviously not strict enough. Thanks for these tips but, unfortunately, I need jams which develop and dissolve as realistic as possible. So, I'm not sure if above methods would introduce some hard to find and explain artifacts. > I do not have any evaluations of real-world behaviour during accidents. Is > something similar known to you? Nope. But I will search for some results about how the synchronization in a traffic jam exactly works. I know it's related to lane changes. Maybe I can extend the lane change model to reflect that. But, this will probably take some time. Regards Björn |