Hello,
the lack of jam wave formation with default parameters is a known issue of the Krauss model. The waves do occur when increasing the sigma parameter from its default of 0.5 to at ~0.7. However, the value of 0.5 was chosen based on other calibration scenarios.
This is by no means the only issue of the Krauss model. By design it uses a very flat acceleration/deceleration profile which has prompted several users to use the IDM model instead.
However, until quite recently (1.0.1), the IDM model (or rather our implementation thereof) had known issues from the interaction with our lane changing models which discouraged us from using it as the default car following model.
There are probably other reasons why we're still holding on to Krauss:
- literally thousands of tests for the Krauss model that would have to be re-evaluated with another default model
- lots of experience with this model and all it's attendant issues
- the interested user is free to select the model anyway
In short, I'm not happy with the lack of jam wave formation but alternative parameter settings would cause other problems.
I honestly cannot tell you yet how the default model or it's default parameters will change in the next versions.
Among other things, we have some ideas on modelling imperfect driving which would lead to a redesign of the default model.
regards,
Jakob
