search

OpenSimulationInterface / open-simulation-interface

A generic interface for the environmental perception of automated driving functions in virtual scenarios.
Other
267 stars 125 forks source link

Add sensor-specific data to detected stationary objects #702

Closed pmai closed 1 year ago

pmai commented 1 year ago

This addresses #462 partially, probably data should also be added to detected traffic signs and lights. Maybe add to DetectedItemHeader instead? Needs to be harmonized with ISO 23150 as well.

jdsika commented 1 year ago

Where is the definition of the field? :)

pmai commented 1 year ago

Where is the definition of the field? :)

Already exists in sensorspecifics...

jdsika commented 1 year ago

Considering the fact that the RCS is a measured unit somehow with dependencies on frequency, viewing angle etc... maybe it is suited better in an detectedItemHeader

pmai commented 1 year ago

Considering the fact that the RCS is a measured unit somehow with dependencies on frequency, viewing angle etc... maybe it is suited better in an detectedItemHeader

Or something similar; let's see where ISO 23150 nowadays puts that information, maybe this gives a clearer picture, at least for 4.0.0; for 3.6.0 we likely should keep as similar to the moving objects as possible...

jdsika commented 1 year ago

Considering the fact that the RCS is a measured unit somehow with dependencies on frequency, viewing angle etc... maybe it is suited better in an detectedItemHeader

Or something similar; let's see where ISO 23150 nowadays puts that information, maybe this gives a clearer picture, at least for 4.0.0; for 3.6.0 we likely should keep as similar to the moving objects as possible...

Then I would opt for a creating an issue for a potential breaking change in v4.0.0 as the fields of movingObjekts would need to be moved as well and at the same time for consistency merge this proposal for v3.6.0

thomassedlmayer commented 1 year ago

As far as I can tell, sensor specific data are only referenced in moving objects in the ISO as of 2022-11-22 (even though they generically refer to a 'recognized object' in the description of RCS/lidar reflectivity).

I'd also say that it generally makes sense to put it in the generic header in the longterm as they are not specific to some object type but rather a commonality among any detected objects. If this is what we want, I can create another PR for the 4.0 milestone.

thomassedlmayer commented 1 year ago

Sensor Model WG (08.02.): No opinions on that. Passed on to harmonization WG to check with ISO23150.

PhRosenberger commented 1 year ago

Sensor Model WG (08.02.): No opinions on that. Passed on to harmonization WG to check with ISO23150.

I would say it does not matter what is written in the ISO23150. We use these information for a different purpose and you want to have this information on all objects for your sensor model, anyways. The actual sensor will calculate e.g. an RCS for any detected object, as all are potentially moving at first sight and could be actually static objects by error. So in simulation you need it on all objects to be able to replicate such erroneous detected (potentially moving) objects.

pmai commented 1 year ago

CCB 2023-03-13: Merge as-is.