Fix problems with checkRewindLinkLanes (trac #913)

[behrisch]

see failing tests in tests/sumo/basic/dynamics/avoid_spill_back

another problem: sometimes CRLL doesn't find enough space because of insufficient look-ahead (there aren't enough driveProcessItems). (improved in bbf6378 (aka r14934). hard to say whether it will never show up again but should be much less likely now)

Migrated from http://sumo.dlr.de/ticket/913

{
    "status": "new", 
    "changetime": "2013-10-25T14:27:10Z", 
    "description": "see failing tests in tests/sumo/basic/dynamics/avoid_spill_back\n\n~~another problem: sometimes CRLL doesn't find enough space because of insufficient look-ahead (there aren't enough driveProcessItems).~~ (improved in [14934]. hard to say whether it will never show up again but should be much less likely now)", 
    "reporter": "namdre", 
    "cc": "", 
    "resolution": "", 
    "_ts": "1382711230767628", 
    "component": "simulation general (micro)", 
    "summary": "Fix problems with checkRewindLinkLanes", 
    "priority": "major", 
    "keywords": "", 
    "time": "2013-05-30T13:35:01Z", 
    "milestone": "1.0.0", 
    "owner": "namdre", 
    "type": "defect"
}

[behrisch]

@namdre changed description from:

see failing tests in tests/sumo/basic/dynamics/avoid_spill_back

to:

see failing tests in tests/sumo/basic/dynamics/avoid_spill_back

another problem: sometimes CRLL doesn't find enough space because of insufficient look-ahead (there aren't enough driveProcessItems).

[behrisch]

@namdre commented:

see another failing test in bugs/ticket699/sigma

[behrisch]

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

[behrisch]

@namdre changed milestone from "" to "1.0.0"

[behrisch]

• @namdre changed component from "microsim - junctions" to "simulation general (micro)"
• @namdre changed description from:

see failing tests in tests/sumo/basic/dynamics/avoid_spill_back

another problem: sometimes CRLL doesn't find enough space because of insufficient look-ahead (there aren't enough driveProcessItems).

to:

see failing tests in tests/sumo/basic/dynamics/avoid_spill_back

another problem: sometimes CRLL doesn't find enough space because of insufficient look-ahead (there aren't enough driveProcessItems). (improved in [14934]. hard to say whether it will never show up again but should be much less likely now)

[namdre]

the problem of #1052 shows up in the Stuttgart network where a big intersection is followed by two short edges. It might be mitigated by #4437 but a more solid solution is still needed.

Either the lookahead-distance must be increased or CRLL ignored if the drive items provide insufficient lookahead.

[namdre]

There are two failing tests:

• 'no_space_behind_junction_blocker_braking' represents a scenario that probably cannot be solved (the leading vehicle brakes and stops suddently)
• '2junctions_2nd_blocked' is caused by invalid approach information from an accelerating vehicle (approach time is underestimated) This causes the ego vehicle to drive and then stop, blocking the first junction. Thus, it doesn't really show an issue with the CRLL algorithm.