Michael Kirk
Michael Kirk
My understanding is that due to vastly undercounting offstreet parking, we'll get more realistic results with `--infinite_parking`, and that I should re-run the results / prebake with that option. Is...
> And risk exposures: Are those risk exposures comparing: A: "Finite parking w/o Broadmoor edits" vs "Inifinite parking w/ Broadmoor edits" or B: "Infinite parking w/o Broadmoor edits" vs "Inifinite...
I've been poking at this and wanted to check in and verify my understanding @dabreegster. I think what's happening is: A car's route is a list of lane IDs (and...
Thank you for all this excellent feedback! I will consider over night and respond in the AM - I'd like to keep working on this.
Thanks for the clarity around how doing lane-based pathfinding avoids certain unsupported scenarios (e.g. mid-segment lane changing). I see why it would be complicated to switch to a road-based route....
> It looks much better! But all but one of the cars in the video are going straight, so it's still far from realistic -- all 3 lanes should be...
> That's also pretty intuitive -- getting stuck behind a slow vehicle for very short distance doesn't matter so much. Where would it get prioritized in this tuple? If we...
As an update, I'm still working on this. My work in progress (https://github.com/dabreegster/abstreet/compare/master...michaelkirk:mkirk/lane-change-2?expand=1) attempts to update router.rs so that an entire uber turn can be optimized just before the agent...