Driver model behaviour from ECE regulation

[asam-oss]

Driver model behaviour from ECE regulation

[arauschert]

Hi,

an interesting suggestion. We haven't considered that yet. The driver model would be a specific controller for the ego vehicle, which could be used instead of the ALKS function. So this is a bit of a different focus, but I can check, if maybe esmini already provides such a controller. This would improve the demonstration, if the ego vehicle isn't driven by the default controller and does not crash into the targets all the time.

[arauschert]

@eknabevcc
Do you have a recommendation, which of the esmini controllers would be suitable for a human driver model that brakes as well (doesn't need to have exactly the parameters of the ALKS driver model)?

[tmIdiada]

Hi, yes for sure there is the possibility to implement it as a controller. But there is also the possibility to create such scenarios with appropriate actions for the ego vehicle, then for each scenario a second xosc would be needed with this driver model (quick and dirty solution to create such a baseline).

But for sure the proper solution would be a controller.

[eknabevcc]

Good discussion. First, esmini is unfortunately lacking a longitudinal driver model/controller. However it is planned to add a simple one, probably during September (no commitment, just prognosis :)).

I also like @tmIdiada's idea of implementing driver behavior by means of OSC maneuver (including relevant events with conditions and actions) put in a catalog which could be referred to in any scenario. Haven't tested it, but it seems conceptually feasible.
But I agree, a controller is a cleaner solution.

To conclude, unfortunately no such controller in esmini yet.