Stuckness with long walk legs

[zneedell]

I'm finding that in rare cases when agents have a super long walk leg that can't complete before the end of the simulation they end up stuck in WaitingToDrive because a StartLegTrigger is never scheduled.

beam/src/main/scala/beam/agentsim/agents/PersonAgent.scala

Line 1213 in db2d22d

if (data.nextLeg.beamLeg.endTime > lastTickOfSimulation) Vector.empty

It's not a problem from a simulation perspective because they wouldn't complete the trip anyway, but perhaps we should either kill them or modify their trigger timing so they don't get stuck and we don't need stuckfinder on.