Sat nav routing logic
Discussion
Been trying to figure this out.
As we all know even the best satnavs cannot beat good accurate local knowledge. Fair enough, can live with that.
Again as we all know when we get a new tool there is a fair bit of playing around in familiar circumstances to figure out what it will/won't do and why we sometimes get unexpected results.
So anytime the system gives me a strange result I've been putting it down to the settings, e.g fastest route chosen when actually the shortest route up a notionally slower road is actually faster, ie local knowledge prevails.
However I can't figure out one journey.
Same start and finish points. Details irrelevant.
Route 1, setting = fastest route.
Calculates a route which only a brain dead numpty would take, even though it is on A roads throughout.
Route 2, setting = shortest route.
Uses a combination of A and B roads but also calculates out at a shorter time than for route 1.
I'm confused
Supplementary question.
I'm used with PC based routing software to have a sliding scale of choices as to fastest <--------> shortest setting, not just the extremes.
Likewise preference to use/avoid certain types of road / features etc etc.
The in car satnav has very limited choices as in fastest or shortest nothing in between.
Again motorways avoid or not, forget exact wording.
Do any of the satnavs have the sort of graduated preferences we see on PC based software or is it like my car based system?
>> Edited by Flat in Fifth on Thursday 8th December 12:27
As we all know even the best satnavs cannot beat good accurate local knowledge. Fair enough, can live with that.
Again as we all know when we get a new tool there is a fair bit of playing around in familiar circumstances to figure out what it will/won't do and why we sometimes get unexpected results.
So anytime the system gives me a strange result I've been putting it down to the settings, e.g fastest route chosen when actually the shortest route up a notionally slower road is actually faster, ie local knowledge prevails.
However I can't figure out one journey.
Same start and finish points. Details irrelevant.
Route 1, setting = fastest route.
Calculates a route which only a brain dead numpty would take, even though it is on A roads throughout.
Route 2, setting = shortest route.
Uses a combination of A and B roads but also calculates out at a shorter time than for route 1.
I'm confused
Supplementary question.
I'm used with PC based routing software to have a sliding scale of choices as to fastest <--------> shortest setting, not just the extremes.
Likewise preference to use/avoid certain types of road / features etc etc.
The in car satnav has very limited choices as in fastest or shortest nothing in between.
Again motorways avoid or not, forget exact wording.
Do any of the satnavs have the sort of graduated preferences we see on PC based software or is it like my car based system?
>> Edited by Flat in Fifth on Thursday 8th December 12:27
they are all basedaround dykstra's graph algorithm with some heuristics the different 'costs' whos relevance get alterered by your preferences.
Fot those interested enough it's a very but of code :-
http://smccd.net/accounts/hasson/C++2Notes/graphs.ht
yes i did write a lot of mapping software in 1990/91
Fot those interested enough it's a very but of code :-
http://smccd.net/accounts/hasson/C++2Notes/graphs.ht
yes i did write a lot of mapping software in 1990/91
Gassing Station | In-Car Electronics | Top of Page | What's New | My Stuff