Historical speed patterns - route calculation too optimistic

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.
Post Reply
michel_wsc
Posts: 8
Joined: Mon Mar 05, 2018 4:00 pm

Historical speed patterns - route calculation too optimistic

Post by michel_wsc »

Hello,
I've made several tests using historical speed patterns and PTV test tool. What I could observe is that the calculated time for the example routes are too optimistic. Compared with results from Google maps, the results from PTV are always on the lower end of the Google results.
Example: A route from Märkisch-Buchholz to Berlin-Kreuzberg.
On Sunday, Google calculated time between 0h:40m and 1h:00m. PTV calculated about 0h:46m.
On Monday, Google calculated 0h:50m to 1h:25m. PTV calculated about 0h:52m.

It seems that PTV is moving along the lower ends of the Google calculation. My question: why is the upper value not taken into account? It's a significant difference between 0h:52m and 1h:25m.

Best regards,
Uli
User avatar
Bernd Welter
Site Admin
Posts: 2564
Joined: Mon Apr 14, 2014 10:28 am
Contact:

Re: Historical speed patterns - route calculation too optimi

Post by Bernd Welter »

Hello Uli,

of course we depend on the data providers when it comes to those historical patterns. Most times I don't see such a big gap as mentioned by you.
What
  • coordinates
  • profile
  • feature layers
do you use? If you provide them to me I can have a look at them.

Best regards,
Bernd
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... :twisted:
michel_wsc
Posts: 8
Joined: Mon Mar 05, 2018 4:00 pm

Re: Historical speed patterns - route calculation too optimi

Post by michel_wsc »

Hello Bernd,
here your requested data:
Coordinates: Starting point 1530170,06486812134,4859
michel_wsc
Posts: 8
Joined: Mon Mar 05, 2018 4:00 pm

Re: Historical speed patterns - route calculation too optimi

Post by michel_wsc »

Hello Bernd,
the post was not complete, I accidently sent it.
Again, here the requested data:

Coordinates (format PTV_MERCATOR):
Starting point: 1530170,0648 6812134,4859
Destination point: 1491399,7297 6882788,3195

Profile: 'carfast' (refers to 'default' in the xroute profile configuration)

Feature layer: I'm not sure what you mean. Is it related to the 'Profile Snippet' to activate the histiorical traffic pattern?

Best regards,
Uli
michel_wsc
Posts: 8
Joined: Mon Mar 05, 2018 4:00 pm

Re: Historical speed patterns - route calculation too optimi

Post by michel_wsc »

Hello Bernd,
did you have time to check my results?

Best regards,
Uli
User avatar
Bernd Welter
Site Admin
Posts: 2564
Joined: Mon Apr 14, 2014 10:28 am
Contact:

Re: Historical speed patterns - route calculation too optimi

Post by Bernd Welter »

Hello Uli,

I have to excuse - I might have overseen your post. But I'm back in the loop now.
Overview of a testclient that calculates the same waypointlist at 00:00, 01:00, 02:00.... and returns the diagrams that show the difference of some KPI. Looks like your route has always the same geometry but takes between 48:58 and 52:35 minutes.
Overview of a testclient that calculates the same waypointlist at 00:00, 01:00, 02:00.... and returns the diagrams that show the difference of some KPI. Looks like your route has always the same geometry but takes between 48:58 and 52:35 minutes.
Looks like the gap isn't that big in the data.
If you wish to have a 1:1 websession get backto me so we can have a look at this scenario together (and at some more tools).

Regards
Bernd
Attachments
test_wsc.xml
Sample request. I change the RoutingParameter.START_TIME and RoutingParameter.START_TIME_ROADEDITOR via a loop.
(3.91 KiB) Downloaded 257 times
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... :twisted:
User avatar
Bernd Welter
Site Admin
Posts: 2564
Joined: Mon Apr 14, 2014 10:28 am
Contact:

Re: Historical speed patterns - route calculation too optimi

Post by Bernd Welter »

HERE map produces distances between 57.92 and 61.03km (5.37% difference) and 1:06:03 and 1:13:16 (10.93%)
HERE map produces distances between 57.92 and 61.03km (5.37% difference) and 1:06:03 and 1:13:16 (10.93%)
TOMTOM map returns between 56.76 and 60.98 km (7.45%) and between 0:50:28 and 1:03:56  (26,68%)
TOMTOM map returns between 56.76 and 60.98 km (7.45%) and between 0:50:28 and 1:03:56 (26,68%)
Providermin periodmax periodmin distancemax distance
TOMTOM0:50:251:03:56 (+26.68%)56,76 km60,98 km (+7.45%)
HERE / TLN1:06:031:13:16 (+10.93%)57.92 km61.03 km (+5.37%)
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... :twisted:
michel_wsc
Posts: 8
Joined: Mon Mar 05, 2018 4:00 pm

Re: Historical speed patterns - route calculation too optimi

Post by michel_wsc »

Hello Bernd,
in the meantime, the issue got a little bit out of sight from my side due to other projects. Nevertheless, I'm very interested in clarifying this issue and would like to take the offer of a 1:1 session. Can we schedule the session for the beginning of next week? Thanks a lot in advance.
Best regards,
Uli
User avatar
Bernd Welter
Site Admin
Posts: 2564
Joined: Mon Apr 14, 2014 10:28 am
Contact:

Re: Historical speed patterns - route calculation too optimi

Post by Bernd Welter »

OK, as discussed behind the scenes: here's the link for a GERMAN speaking session next Tuesday...
For those who check this post: feel free to join the session on July 10th, 2018 - 10am.
https://attendee.gotowebinar.com/regist ... 2942293249

Best regards,
Bernd
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... :twisted:
User avatar
Bernd Welter
Site Admin
Posts: 2564
Joined: Mon Apr 14, 2014 10:28 am
Contact:

Re: Historical speed patterns - route calculation too optimi

Post by Bernd Welter »

here's the video...

and the snippet I used

Code: Select all

<?xml version="1.0" encoding="UTF-8"?>
<Profile>
	<Routing majorVersion="2" minorVersion="0">
		<Course>
			<AdditionalDataRules enabled="true"/>
		</Course>
	</Routing>
	<FeatureLayer majorVersion="1" minorVersion="0">
		<GlobalSettings enableTimeDependency="true"/>
		<Themes>
			<Theme id="PTV_SpeedPatterns" enabled="true"/>
		</Themes>
	</FeatureLayer>
</Profile>
Combine this with the RoutingOptions START_TIME and IS_DESTTIME to consider your current time scope.

Best regards,
Bernd
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... :twisted:
Post Reply