Page 1 of 1

Variable radar accuracy

Posted: 01 Feb 2007, 15:19
by KDR_11k
The idea is that instead of having a constant accuracy, each pixel on the radar map has a float value that determines how accurately the position of a unit on it will be displayed.
New tags for radar units:
RadarAccuracy (float, 0.0-1.0): How accurate this radar is at its best. 0.0 to 1.0 would vary the accuracy from "somewhere on this side of the map" to "exactly here".
RadarFalloff (int, 0-inf): Distance up to which the radar will keep its maximum accuracy, defaults to the radar's range. After that it will linearly fall off.
RadarMinAccuracy (float, 0.0-1.0): The radar's accuracy at the edge of its range.
RadarStackable (bool): If true the radar will add its accuracy to the coverage other radars provide of the area (otherwise the maximum accuracy that covers a pixel is used). Well, not add but make it 1-((1-oldaccuracy) * (1-newaccuracy)).

Jammers would have similar tags and the final accuracy for a pixel would be radaraccuracy-jammeraccuracy. That means a weak jammer wouldn't make your units disappear from the radar but would make them harder to hit.

Posted: 01 Feb 2007, 23:38
by manored
With variation of the radar precision you mean bigger/smaller area where the unit may be?

Posted: 01 Feb 2007, 23:45
by Radja
that sounds so resource intensive....

but toying with radars sounds good, maybe cheap tiny good for nothing units that appear like strong units on the radar, or a unit that in radar appears like a group of 5, or 20 units together would just do

Posted: 02 Feb 2007, 00:09
by manored
Maybe if we give the radar "diferent detection strenghs" depending of the range of the unit instead of gradual increase it would make it less workfull to do. What I mean is like " from 1 to 1000 the radars detects well, from more not so well, etc".

Posted: 02 Feb 2007, 10:37
by KDR_11k
Linear interpolation is trivially easy to do.

Posted: 02 Feb 2007, 15:13
by bamb
afaik real radars either see or don't see an object. If you get a return, you know the bearing pretty exactly and the distance is known from the lightspeed delay. But I'm no expert on this of course.

Posted: 02 Feb 2007, 16:44
by KDR_11k
Yes but Spring doesn't have real radars.

Posted: 02 Feb 2007, 18:58
by MadRat
Don't forget to add steerable conical-shaped radar beams to this too, KDR_11k. :)

Posted: 02 Feb 2007, 21:09
by manored
bamb wrote:afaik real radars either see or don't see an object. If you get a return, you know the bearing pretty exactly and the distance is known from the lightspeed delay. But I'm no expert on this of course.
This is true.
But real radars cant detect ground units since terrain obstacles would block it, what means that TA radars are probaly diferent from real radars... :P

Posted: 02 Feb 2007, 21:59
by Argh
No, Spring radars obey LOS. They're really pretty physically accurate, tbh. They just act like sight-radii, basically. Seismic, on the other hand, doesn't even care about LOS, which is great for simulating things that aren't radar.

Posted: 03 Feb 2007, 08:22
by KDR_11k
I wish it was mod definee whether a radar is blocked by terrain.