Difference between revisions of "Movement Challenge 2022/Prior Probability Challenge"

From Robowiki
Jump to navigation Jump to search
m
m
 
Line 9: Line 9:
 
: Random targeting.  
 
: Random targeting.  
 
; Ablation experiments of danger models
 
; Ablation experiments of danger models
: Danger models should be replaced by uniformly distributed angles between PPMEA as well.  
+
: Danger models should be replaced by uniformly distributed angles between PPMEA as well (don't use monte-carlo here).  
 
: Original danger models can also be tested, which tests their ability to accurately fit random targeting as well.  
 
: Original danger models can also be tested, which tests their ability to accurately fit random targeting as well.  
  
 
[[Category:Movement Challenges]]
 
[[Category:Movement Challenges]]

Latest revision as of 07:02, 20 March 2022

This article is a stub. You can help RoboWiki by expanding it.

Test your bot against random targeting, which is prior probability of getting hit, in order to compare strategies including distancing, max escape angle and precise wave intersection.

Random targeting
Defined as a robot firing uniformly distributed angles between Precise Positional Maximum Escape Angle (PPMEA).
Firer movement
A variety of movements should be selected, including stationary, random movement, simple wave surfing & sophisticated wave surfing. (precise list is TO BE DISCUSSED).
Participant targeting
Random targeting.
Ablation experiments of danger models
Danger models should be replaced by uniformly distributed angles between PPMEA as well (don't use monte-carlo here).
Original danger models can also be tested, which tests their ability to accurately fit random targeting as well.