This forum deals with any kind of routing computation whether it is simple A:B-routing, calculation of isochrones, simple matrix computation or nearest search.
a partner wants to build an ETA calculation based on a guided navigation process in the backend (not on the mobile device).
We discussed an idea, but before we implement this with efforts I'd like to get a quick feedback about the approach or maybe a correction / alternative approach. Here's how we would tackle this based on xServer2 (PTV Developer offers a dedicated method getEstimatedTimeOfArrival / concept)
Avaliable Input
Vehicle's current (lat/lon) plus Timestamp (let's call it CUR)
Precalculated BCR file including multiple "silent via points" (SVP)
Output
Reconstructed route with proper ETAs for each waypoint - geometry should follow the precalculated BCR polylines / SVP
and here's the meta description of the process:
use a 1:N matrix call to identify the SVP which can be reached fastest : let's call it SVP*. Gather the required traveltime from CUR to SVP* and determine the arrival time at SVP*
identify the subsequence of SVPs that begins with SVP* (in other words: cut all SVPs prior to SVP*)
perform a track matching with this subsequence from SVP* to SVP.Last
What we do NOT know:
Which profile was used for the BCR calculation. Therefore we have to let the user pick a profile.
The doubts I have:
Selection of SVP* : could the dima lead to the wrong SVP backwards?
Are SVPs positions valid input for the track matching?
Best regards,
Bernd
Attachments
Sample image: Waypoints: Karlsruhe, Bruchsal, Heidelsheim. Red Pins: SVPs between Karlsruhe and Bruchsal (1st leg) Cyan pins: SVPs between Bruchsal and Heidelsheim (2nd leg) Blue cross: current position Blue circles: candidates "close" tu CUR
Though the lower blue circle is faster to be reached from an individual perspective, the upper blue circle is where I would expect the guided navigation to return to the route. Or not?
Bernd Welter Technical Partner Manager Developer Components
PTV Logistics - Germany
Bernd at... The Forum,LinkedIn, Youtube, StackOverflow
I like the smell of PTV Developer in the morning...
Are SVPs positions valid input for the track matching?
For a usable output it depends how dense the SVP's are. If they are to spread out from each other you will get a lot of broken paths. In that case you are better off with a regular routing and use manipulate waypoints for the SVP's.
