Hi Harri, yes, the Garmin algo seems to work like this: try to get on a high class road as soon as possible and try to reach a point near the target without using a lower class road again. So, once a class 4 road is used it will accept rather long detours to get to a target. The img data even contains rather redundant information to help with this calculation. This algo works well for countries with motorways and car routing but for cycling maps you might get better results when you don't use high road classes unless you want a map for long and nice cycling tours where a long detour on a nicer road is welcomed. Besides that you should not rely on results that you got from much older mkgmap versions if they wrote different NOD information. We still learn how the bits and bytes in the IMG format are interpreted by Garmin and thus older releases often simply encoded wrong data. Sometimes it was a miracle to me how the Garmin algo still found reasonable routes. Gerd ________________________________________ Von: mkgmap-dev <mkgmap-dev-bounces@lists.mkgmap.org.uk> im Auftrag von Harri Suomalainen <hsuomal@welho.com> Gesendet: Donnerstag, 16. Juli 2020 10:35 An: mkgmap-dev@lists.mkgmap.org.uk Betreff: Re: [mkgmap-dev] Road speed through urban areas That approach based on POIs does not work too well. You have no idea if you lower the speed for a short segment or a long one, it just depends on eg. if there are traffic lights on long or short segment. Same problem with other poi types like crossings. BTW, For another message suggesting lower road classes. I'd not lower road classes too much. According to my tests garmin seems to ignore lower road classes on long trips (except near ends) and if there are huge number of equal road classes, route calculation time can get really large. I tried a few years back lifting cycleways to category similar to motorways (to get better cycling routing). That resulted in route calculation times like 10 or 15 minutes for a 10km trip in the city. (At least on my older Oregon). On 7/16/20 11:14 AM, Joris Bo wrote:
Hi,
In urban areas the 'practical maxspeed' is often lower then the 'official' maxspeed because of existence of traffic lights and traffic calmers. I would rely on the intended routing basics of road_class and corresponding maxspeeds and the common usage of highway types and tagging in osm.
Maybe it helps to make these roads less attractive for the routing engines to check for traffic lights. Because these are node-tags and not line-tags I don't know the exact impact.
highway = residential & highway=crossing & crossing=traffic_signals {set mkgmap:road-speed=-2} highway = residential & traffic_calming=table { set mkgmap:road-speed=-1} highway = residential & access = destination { set mkgmap:road-speed=-2}
Joris
mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev