![]() ![]() Note: Lower scores are better, see Setup section for how this score is calculated Because of these wide roads, this is the only intersection where switching lanes like this can lead to an advantage, which is why this intersection performs better than the rest, although, big caveat, this intersection took a lot of attempts to balance the lane switching to get it to perform to this level. ![]() This intersection has multiple 3-lane roads to carry vehicles in multiple directions, this allows vehicles to speed up and pass others in the middle of the intersection. Usually this does not matter as switching lanes in an intersection almost always slows traffic down, I say almost, because there is one big and distinct exception, intersection no.30. One of the disadvantages of controlling lane changing so tightly is that slower traffic holds up faster traffic because the traffic is usually not allowed to switch lanes. Most of the other intersections took just 2 or 3 attempts at most. No24 and no30 gave particular problems as the optimal configuration is not obvious, and these ones took many attempts to get a good score. I set up what I think might be the best rules and let the test run, then I re-run the test if I think I can make improvements. ![]() For every intersection I set up TMPE lane-crossing rules for every single road segment in and around the intersection. TM:PE Lane changing rules:This is actually the biggest variable in the tests, handling lane merging and diverging. The intersections have no output vehicles in the first minute, it's only in the third minute that we can see the output fully catching up the initial input vehicles. To allow me to more directly compare between intersections of different sizes, the road segments that were used for counting were placed quite far from the centre of the intersections, having them in the same place for each intersection makes direct comparisons more correct, but it does introduce a lag, this can be seen in the graphs. To get the final score I get the per minute average for each intersection for the duration (14 minutes) of the test. Then it tallys up the entering and leaving counts every minute. To do this I designate two different sets of roads, one for leaving the intersection and one for entering. Vehicle CountingTo work out the score for each intersection I developed a small Traffic Counting Mod which counts the number of vehicles entering the intersection and vehicles leaving the intersection every minute. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |