Re: [sumo-dev] Regrading understanding Car following Model

Sir,
As I mentioned earlier, I am currently studying the implementation of KraussOrig1 model. Sir, I have a few questions about it.

1. I don't understand the macros ACCEL2SPEED, SPEED2ACCEL, DIST2SPEED.

2. I don't completely understand the following function, their implementation, and use.
2.1 double MSCFModel::maximumSafeFollowSpeed(double gap, double egoSpeed, double predSpeed, double predMaxDecel, bool onInsertion)
2.2 double MSCFModel::maximumSafeStopSpeed(double g /*gap*/, double v /*currentSpeed*/, bool onInsertion, double headway)
2.3 double MSCFModel::brakeGap(const double speed, const double decel, const double headwayTime)

Can you please guide me through this?

Thanks and regards
Shantanu Phanse

On Fri, Jan 8, 2021 at 6:39 PM Jakob Erdmann <namdre.sumo@xxxxxxxxx> wrote:
Hello,
even the kraussOrig Model has quite some deviations from the work of Krauss. It is based directly on equation 4.8 in the dissertation and actually is more like the approach of Gipps discussed at the start of Section 5.1. (vSafe is derived directly from braking distance which are quadratic in speed).
The exact commit where the original Krauss equation was changed to the current version dates to 2007 (https://github.com/eclipse/sumo/commit/6f7b1d5ffba1ac575988efc5f59b6d6d376a1345#diff-41a035c9bfbce931e5a134ddb3f95190657aa6b183d8db8dfc4a48797d3dfc8c)
I would assume this was done to fix collisions but the exact reason would have to be given by D. Krajewicz.

Some points of note when relating Krauss to KraussOrig1:
- in the SUMO-Krauss model (not KraussOrig1), braking distances are computed depending on the position update rule since  an euler-like update (move with the speed of the current step) leads to other braking distances than a ballistic update (move with the average speed of previous and current step).
- follow speed is a special case of stopping speed (where stop distance = gap + leaderBrakeGap)
- special care is taken when vehicles have different braking capabilities because the Krauss equation 4.8 does not guarantee save driving in this case (https://arxiv.org/abs/1902.04927)

regards,
Jakob

Am Fr., 8. Jan. 2021 um 12:02 Uhr schrieb SHANTANU PHANSE <201911048@xxxxxxxxxxxx>:

Hello Sir/Mam,

I am new to SUMO. I am learning about the microscopic car following model. My problem is that I am unable to relate the kraussOrig model in the simulator and the krauss model's theory. I am referring to the attached papers for the theory.