Re: [Speed-dreams-devel] Network test today

[Wolf-Dieter B. <wd...@wd...>]

Hello Simon, hello all,

The "only use position" approach can work if we have a fast network
connection.

> for example skid sounds and tire marks
Yes these are needed things and this is why I assume we need the forces as
well for synchronization.

For local interpolation the robot code could be used. Even a human driver is
a robot (for the race engine).
So we could implement some robot code to be used as interpolation engine for
human drivers.
This approach could be used as well for future driver assistance systems.

Cheers

Wolf-Dieter

-----Ursprüngliche Nachricht-----
Von: si...@mu... [mailto:si...@mu...] 
Gesendet: Sonntag, 30. September 2012 17:45
An: "Kristóf Kály-Kullai"
Cc: Wolf-Dieter Beelitz; Andrew Sumner;
spe...@li...
Betreff: Re: [Speed-dreams-devel] Network test today

> Hello Wolf-Dieter and all,
>
> But I do not want to deal with driver input, only with position and 
> velocity. Something like this:
> - interpolate to the last, say, 3 positions
> - use this to extrapolate until a new position arrives
> - then recalculate, and smoothly change to this new one.
>
> This way the car on the client will likely never be exactly where it 
> is on the master, but would never be too far away.

Personally I like the sound of this concept, but there are some other things
to consider - for example skid sounds and tire marks (which I believe are
created from the full simulation data).

This technique of a 'guidance' for the local simulation engine to follow
could also be used during recording/payback/flashback/etc to reduce the
amount of data stored.

Simon