Precisely. I would suggest using the profile documentation to build a
simple profile from the ground up. The default profiles are very
complicated and take advantage of some features of the Lua language that
you may not understand when you first look at them. I didn't. I do
believe the profile documentation is pretty complete though; the default
profiles just have a lot of complex stuff built on top of that.
Nate Wessel
Jack of all trades, Master of Geography, PhD Candidate in Urban Planning
SAUSy Lab <http://sausy.ca>, Sid Smith Hall, University of Toronto
On 04/26/2018 04:00 PM, Patrick Agin wrote:
And by "handle subway lines", you mean "specify that only subway edges
are included"?
Can I ask you to point in .lua profile where you specify such a thing?
In restricted_highway_whitelist structure maybe?
The readme on profiles seems good but it's general (for example, there
is nothing about this restricted_highway_whitelist).
Do you know where can I find a complete tutorial/documentation on
writing profiles?
thanks again Nate
Patrick
2018-04-26 15:30 GMT-04:00 Nate Wessel <[email protected]
<mailto:[email protected]>>:
You can use a profile to determine which edges are included in the
graph, which are accessible, etc. If for example you only handled
subway lines in the profile and disregarded all other data from
OSM, you would only be able to match things to subways.
A GPS trace from someone e.g. walking aboveground but parallel to
a subway track would also likely match to the subway rails because
there would be no more-likely candidates to match to.
Nate Wessel
Jack of all trades, Master of Geography, PhD Candidate in Urban
Planning
SAUSy Lab <http://sausy.ca>, Sid Smith Hall, University of Toronto
On 04/26/2018 02:32 PM, Patrick Agin wrote:
Thanks again Nate, I will take a look for sure. But just out of
curiosity, what your ttc profile is useful for then? And how do
you (in general, big picture only) define in a profile that it's
subway only? And does it mean that when map matching against this
profile, only subway routes will be returned?
Patrick
2018-04-26 12:35 GMT-04:00 Nate Wessel
<[email protected] <mailto:[email protected]>>:
This one won't return a subway route because I haven't
configured it for that. I don't have subway trains in the
dataset I'm working with for Toronto. But it would be pretty
easy to make a subway only profile, or to add subways into
this one if you want to do all transit modes in one profile.
Take a look at this page for documentation on how profiles
work:
https://github.com/Project-OSRM/osrm-backend/blob/master/docs/profiles.md
<https://github.com/Project-OSRM/osrm-backend/blob/master/docs/profiles.md>
If you want to understand how map matching in OSRM works
generally, I think this paper is a good place to start:
https://www.microsoft.com/en-us/research/wp-content/uploads/2016/12/map-matching-ACM-GIS-camera-ready.pdf
<https://www.microsoft.com/en-us/research/wp-content/uploads/2016/12/map-matching-ACM-GIS-camera-ready.pdf>
When you change the profile, you're essentially configuring
the street/rail/whatever network that your GPS trace will
match against. Montreal definitely has subways in OSM; you
can see them in the main map at openstreetmap.org
<http://openstreetmap.org>.
Best,
Nate Wessel
Jack of all trades, Master of Geography, PhD Candidate in
Urban Planning
SAUSy Lab <http://sausy.ca>, Sid Smith Hall, University of
Toronto
On 04/26/2018 10:52 AM, Patrick Agin wrote:
Hi Nate and thank you very much for your help. I have a very
basic question about your ttc profile (sorry if it's too
silly). When you match a GPS dataset against this profile
instance, will it return a subway route for example? If yes,
why this subway route would be declared much plausible by
osrm than the street route? Only because of the ttc profile?
I realize that I don't understand very well the involved
mechanisms behind osrm and profiles, if you can shed some
light on this, I would greatly appreciate. Another question:
I suppose that the subway data has to be included in the osm
data of my city (Montreal) in order to get subway routes
from match . How can I know if it's there or not in my osm
data?
Thank you again Nate,
Patrick
2018-04-26 9:36 GMT-04:00 Nate Wessel
<[email protected]
<mailto:[email protected]>>:
I've had some success with transit matching by simply
modifying the car profile to allow routing on streetcar
tracks, allowing access for buses and public service
vehicles, etc. Here's a profile I'm using for transit now:
https://github.com/SAUSy-Lab/retro-gtfs/blob/master/etc/ttc.lua#L347
<https://github.com/SAUSy-Lab/retro-gtfs/blob/master/etc/ttc.lua#L347>
I had also tried to develop a more transit specific
profile from the ground up. My thought was that known
transit routes (provided in the OSM data) would be
preferred by modifying their edge weights. This however
ran into an issue with the matching service which hasn't
been resolved yet; edge weights simply aren't used in
match results.
https://github.com/Project-OSRM/osrm-backend/issues/4785
<https://github.com/Project-OSRM/osrm-backend/issues/4785>
If anyone has a suggestion for the question about
multimodal paths, I would be very keen to hear it,
however I suspect this is still an open topic of
research with no definite best practices.
Best,
Nate Wessel
Jack of all trades, Master of Geography, PhD Candidate
in Urban Planning
SAUSy Lab <http://sausy.ca>, Sid Smith Hall, University
of Toronto
On 04/26/2018 09:01 AM, Patrick Agin wrote:
Dear all,
I'm a newbie to routing engines and my first attempts
are with OSRM. Specifically I want to use the/match
/service. Questions I have is :
1) is it possible with match service to specify public
transport (e.g. subway) as the mode of transportation?
I see only profiles for car, bike and foot. If not, is
there a solution? I'm pretty sure it it possible but
how? I saw in some profiles the public_transport
parameter but I don't know how to use it.
2) is it possible for match to return a multi-modal
path given the timestamps of each coordinate? For
example, return a sub-path that is the most plausible
for pedestrian (chosen because of the low speed
inferred from timestamps) followed by another subpath
that is the most plausible for car (because of a higher
speed)? Or is the only solution to pre-process the data
to infer by myself the mode of transportation and make
subsequent calls to the corresponding profile instance?
Thanks a lot,
Patrick
_______________________________________________
OSRM-talk mailing list
[email protected]
<mailto:[email protected]>
https://lists.openstreetmap.org/listinfo/osrm-talk
<https://lists.openstreetmap.org/listinfo/osrm-talk>
_______________________________________________
OSRM-talk mailing list
[email protected]
<mailto:[email protected]>
https://lists.openstreetmap.org/listinfo/osrm-talk
<https://lists.openstreetmap.org/listinfo/osrm-talk>
_______________________________________________
OSRM-talk mailing list
[email protected] <mailto:[email protected]>
https://lists.openstreetmap.org/listinfo/osrm-talk
<https://lists.openstreetmap.org/listinfo/osrm-talk>
_______________________________________________
OSRM-talk mailing list
[email protected] <mailto:[email protected]>
https://lists.openstreetmap.org/listinfo/osrm-talk
<https://lists.openstreetmap.org/listinfo/osrm-talk>
_______________________________________________
OSRM-talk mailing list
[email protected] <mailto:[email protected]>
https://lists.openstreetmap.org/listinfo/osrm-talk
<https://lists.openstreetmap.org/listinfo/osrm-talk>
_______________________________________________
OSRM-talk mailing list
[email protected] <mailto:[email protected]>
https://lists.openstreetmap.org/listinfo/osrm-talk
<https://lists.openstreetmap.org/listinfo/osrm-talk>
_______________________________________________
OSRM-talk mailing list
[email protected]
https://lists.openstreetmap.org/listinfo/osrm-talk
_______________________________________________
OSRM-talk mailing list
[email protected]
https://lists.openstreetmap.org/listinfo/osrm-talk
