Skip to main content

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [sumo-user] [EXTERNAL] Re: Netconvert GeoConvHelper

I think the projection might be a side issue. I'd recommend that you check whether the original geometry (OSM?) is good enough:
- create a list of <poi> elements with 'lon', 'lat' attributes corresponding to your vehicle trajectory
- load those pois as an additional file together with your .net.xml (the one you wish to convert to OpenDRIVE) into sumo-gui.
(see https://sumo.dlr.de/docs/Simulation/Shapes.html#poi_point_of_interest_definitions)

Only if those pois lie on the road lanes, will you be able to find an x,y mapping that also fits.

Am Mi., 2. Feb. 2022 um 15:27 Uhr schrieb Ovsenjak, Gregor AVL/SI via sumo-user <sumo-user@xxxxxxxxxxx>:

Hello again,


On the screenshots that I provided, we have an offset between the road and the vehicle trajectory. (just the far right trajectory is relevant). The vehicle collects GNSS data (latitude,longitude) with a frequency of 1Hz. Do You have any suggestions on how to align the car positions with the road, so that the car is driving on the road, using SUMO.


Our current Pipeline for road generation is --> Extract the min,max values of latitude and longitude from the collected data  --> download a OSM map with a rectangle around min,max values  --> Transform to SUMO map (.net.xml)


Our Pipeline for vehicle trajectory creation is --> Extract the min,max values of latitude and longitude from the collected data --> Create a local ENU coordinate system with a reference point (origin) at the extracted min, max values --> Calculate x,y positions of the vehicle based on the ENU coordinate system. 


Next we use the offset netconvert -(offset.x,offset.y) options to set the origin of the SUMO map to the ENU coordinate system reference point. At the end we create the OpenDrive map from the SUMO map.


If we could align the road with the vehicle trajectory in a better, simpler way this would bring our research much further.


Thank You again,

Gregor Ovsenjak




Od: sumo-user <sumo-user-bounces@xxxxxxxxxxx> v imenu Jakob Erdmann <namdre.sumo@xxxxxxxxx>
Poslano: sreda, 02. februar 2022 13:41:21
Za: Sumo project User discussions
Zadeva: Re: [sumo-user] [EXTERNAL] Re: Netconvert GeoConvHelper
 
Actually, there are two reference positions we could be talking about: a) the point where the distortion of the projection is minimal (for UTM this would be the center of the current UTM zone, I think). b) the point that is called (0, 0) ‍ ZjQcmQRYFpfptBannerStart
This message is from an external sender
This email originated from outside of the organization. Do not click links or open attachments unless you recognize the sender and know the content is safe. If you are not sure, you can use the ReportPhish Button to submit the message for analysis.
ZjQcmQRYFpfptBannerEnd
Actually, there are two reference positions we could be talking about:

a) the point where the distortion of the projection is minimal (for UTM this would be the center of the current UTM zone, I think).
b) the point that is called (0, 0)

Point b) can be shifted arbitrarily by setting the appropriate x,y offset (netconvert options --offset.x --offset.y)
If this is what you need to change, it should suffice to compute the x,y coordinate of your desired lon,lat reference point and set the offset the negative of this value

If you need to change a) the you must use a projection other than UTM and you can set this with netconvert option --proj and the appropriate proj-string (https://proj.org/usage/quickstart.html) but I cannot tell you which string that would be.

regards,
Jakob


Am Mi., 2. Feb. 2022 um 13:28 Uhr schrieb Ovsenjak, Gregor AVL/SI via sumo-user <sumo-user@xxxxxxxxxxx>:

Hello Jakob,

Sorry maybe I did not explain my question very well. Here is a shorter summary.

 A GEO to ENU system needs a reference position (Lon, Lat,Alt). This reference position gives a road representation in ENU system with x and y positions (flat earth). Due to the ENU transformation a map is good only for areas <2km without position drift. SUMO generates road network in UTM shifted to ENU system with the reference point always on the lower left corner (min lon, min lat )


Question: is there a chance to set this reference position (lon, lat) by our own defined values ?


With kind regards,

Gregor Ovsenjak



Od: sumo-user <sumo-user-bounces@xxxxxxxxxxx> v imenu Jakob Erdmann <namdre.sumo@xxxxxxxxx>
Poslano: sreda, 02. februar 2022 12:46:30
Za: Sumo project User discussions
Zadeva: [EXTERNAL] Re: [sumo-user] Netconvert GeoConvHelper
 
Are there any other differences to the projections besides the shift?  Are the axis directions different (i.e. increasing from north to south)? Am Mi., 2. Feb. 2022 um 09:03 Uhr schrieb Ovsenjak, Gregor AVL/SI via sumo-user <sumo-user@xxxxxxxxxxx ZjQcmQRYFpfptBannerStart
This message is from an external sender
This email originated from outside of the organization. Do not click links or open attachments unless you recognize the sender and know the content is safe. If you are not sure, you can use the ReportPhish Button to submit the message for analysis.
ZjQcmQRYFpfptBannerEnd
Are there any other differences to the projections besides the shift?  Are the axis directions different (i.e. increasing from north to south)?

Am Mi., 2. Feb. 2022 um 09:03 Uhr schrieb Ovsenjak, Gregor AVL/SI via sumo-user <sumo-user@xxxxxxxxxxx>:

Hello Jakob,


I am working on a project where a vehicle is collecting data. We are trying to recreate the map in an OpenDrive format. For that we need the SUMO netconvert  software (which is awesome, by the way !!). The vehicle collects data, and creates a local coordinate system based on the ENU transformation, which means that they create a local tangent plane coordinate system. But they take the first measured GNSS point as the ENU reference point, which is (in this case) always in the top-right corner of the map. 

This presents a problem for us because each SUMO network imported from a geo-referenced data sources comes with it's own ENU projection: By default, these are UTM coordinates which are shifted so that the lower left network corner is the coordinate origin. Now we have one ENU coordinate system in the top right corner and one in the bottom left corner, which introduces an offset when overlaping x,y data from vehicle with the SUMO converted OpenDrive file.

I am wondering if there is a way to align these two ENU systems, or if we can move the ENU tranformation to the top right corner. We have tried shifting the coordinate origin but this does not solve the problem.

Sorry for the long explanation,


With kind regards,

Gregor Ovsenjak

_______________________________________________
sumo-user mailing list
sumo-user@xxxxxxxxxxx
To unsubscribe from this list, visit https://www.eclipse.org/mailman/listinfo/sumo-user
_______________________________________________
sumo-user mailing list
sumo-user@xxxxxxxxxxx
To unsubscribe from this list, visit https://www.eclipse.org/mailman/listinfo/sumo-user
_______________________________________________
sumo-user mailing list
sumo-user@xxxxxxxxxxx
To unsubscribe from this list, visit https://www.eclipse.org/mailman/listinfo/sumo-user

Back to the top