Zian Choy <[email protected]> writes: > OsmAnd routes drivers going to buildings near freeways to use a freeway > offramp rather than the nearby street. It should prefer streets because it > is illegal to park a car on a freeway offramp.
I think the real issue is that "compute car route to point" is really "compute route to point by car to close, assuming you will walk the rest of the way, and that the rest of the way walking is possible". This is even harder because there are many things about the world that are not on the map, like parking on some streets, and entire parking lots. So, I think a routing algorithm needs two things: ignoring preferences on the last segments having some notion of weight for ending place, so that a non-limited-access highway way is better, even if a bit farther wawy. Perhaps choosing the road point to go to should be a pre-step before route computation, or perhaps it should be integrated. Overall, I think routing needs to move away from fastest time to some blend of time, distance, and other factors. For example, ending on a motorway_link could be worth a penalty of 500m, compared to some other road. Also, one could expect routes to be a joint car/walking route, and use footpaths, or not to cross fences. -- You received this message because you are subscribed to the Google Groups "Osmand" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/d/optout.
signature.asc
Description: PGP signature
