Hi Dion, I'd say a bit of both.
The junction is topologically correct but looking at the aerial imagery and the node that you circled, https://www.openstreetmap.org/node/1691043684 , then it could be moved very slightly north and a bit more aggressively west to lessen the change of direction and better fit the actual physical situation. You could also map in the pedestrian crossing and close the gap further.
OSMAnd may also need some tweaking. I helped test a commercial routing product and know that these situations are difficult to get right. At the most simple, OSMAnd should measure the deviation from directly straight on (0 degrees) and assign anything up to, say 8 as "straight on", to around 45 degrees as "bear left" and anything more as "turn left". The may still not get it right. There are a couple more sophisticated things it could do: 1) Note that you are going from/to the same road name/classification and dynamically broaden the "straight on" angle test or even drop the navigation instruction entirely; 2) look ahead to the next node or two and create some kind of smoothed average angle, which will again help push the instruction to "straight on".
Mike On 2019-01-31 09:54, Dion Moult wrote:
G'day all! In the intersection of Liverpool road and Burwood road in Burwood, Sydney (see attached), if I am travelling in the direction shown by the red arrow, then my GPS device should tell me to continue and drive straight at the intersection. However, because at that junction, the map splits up Liverpool Road into two roads, OSMAnd tells me to turn left there, which is quite confusing. What is the appropriate way to fix this mapping? Or is it a problem with OSMAnd? Dion Moult _______________________________________________ Talk-au mailing list [email protected] https://lists.openstreetmap.org/listinfo/talk-au
_______________________________________________ Talk-au mailing list [email protected] https://lists.openstreetmap.org/listinfo/talk-au
