search

eclipse-sumo / sumo

Eclipse SUMO is an open source, highly portable, microscopic and continuous traffic simulation package designed to handle large networks. It allows for intermodal simulation including pedestrians and comes with a large set of tools for scenario creation.
https://eclipse.dev/sumo
Eclipse Public License 2.0
2.55k stars 1.43k forks source link

vehicle acts strange when trying to help another with lane changing (trac #937) #937

Closed behrisch closed 7 years ago

behrisch commented 11 years ago

veh1 is already going very slowly and then stops to help veh2 which is more than 2 vehicle-lenghts behind it change to its lane. When veh2 is next to veh1, veh1 starts to accelerate again, preventing veh2 from actually changing its lane.

Instead, veh1 should wait until veh2 is next to it and then slow down to let veh2 change lanes.

Migrated from http://sumo.dlr.de/ticket/937

{
    "status": "closed", 
    "changetime": "2014-02-12T09:59:19Z", 
    "description": "veh1 is already going very slowly and then stops to help veh2 which is more than 2 vehicle-lenghts behind it change to its lane. When veh2 is next to veh1, veh1 starts to accelerate again, preventing veh2 from actually changing its lane.\n\nInstead, veh1 should wait until veh2 is next to it and then slow down to let veh2 change lanes.", 
    "reporter": "namdre", 
    "cc": "", 
    "resolution": "fixed", 
    "_ts": "1392199159058703", 
    "component": "microsim - lane change", 
    "summary": "vehicle acts strange when trying to help another with lane changing", 
    "priority": "major", 
    "keywords": "", 
    "time": "2013-06-26T10:54:16Z", 
    "milestone": "1.0.0", 
    "owner": "", 
    "type": "defect"
}
behrisch commented 11 years ago

@namdre commented:

another case I noticed: veh1 wants to drive from an on ramp onto the highway but is blocked by veh2. veh2 starts to decelerate to let veh1 in but since veh1 is nearing the end of the lane it also starts to break (by design since lane-change is not certain) and thus the slowing down of veh2 does not help but rather makes it worse, forcing veh1 to come to a full stop.

To get realistic dynamics in this behavior would require an artificially lengthed on-ramp to correspond to the way vehicles who fail their lane change can continue driving on the shoulder of the road instead of slamming into a wall. However, since lane-changes fail more often in the simulation than in reality this would also lead to an unrealistic number of vehicles on the shoulder of the road (especially in jammed conditions). To fix this would require special rules which allow longer on-ramps only under certain conditions of the main road (i.e. minimum average speed and maximum density)

behrisch commented 11 years ago

@namdre commented:

see tests/sumo/basic/dynamics/ramp

behrisch commented 11 years ago

@namdre changed component from "simulation general (micro)" to "microsim - lane change"

behrisch commented 11 years ago

@namdre changed milestone from "" to "1.0.0"

behrisch commented 11 years ago

these problems are mitigated in lane change model JE2013 (see #1045)