Difference between revisions of "@workgun"

From Robowiki
Jump to navigation Jump to search
Line 5: Line 5:
 
* create a Pattern based on data
 
* create a Pattern based on data
 
* find similar pattern(s)
 
* find similar pattern(s)
* shot as PM, create a virual bullet containig the Pattern and is weigth
+
* shot as PM, create a virual bullet containig the Pattern and his weigth
 
* if bullet hit modify actual gun weigth with virtual bullet weigth
 
* if bullet hit modify actual gun weigth with virtual bullet weigth
  

Revision as of 16:32, 23 March 2009

@work gun

I'm thinking about this new ibrid gun, it's work is quite simple: get and save data as PM gun, but can choose if use a PM or GF targeting algoritm.

How gun works?

  • get data
  • create a Pattern based on data
  • find similar pattern(s)
  • shot as PM, create a virual bullet containig the Pattern and his weigth
  • if bullet hit modify actual gun weigth with virtual bullet weigth

Which data save?

(difference are calculate every tick, if one or more scan fail gun can try to estimate lost data)

  • enemy distance travelled bin from 0 to 8 (int)
  • enemy bearing change bin from 0 to 46 (int)
  • i've shot bin from 0 to 3 (int)
  • enemy has shot bin from 0 to 3 (int)
  • distance from me bin from 0 to 10 (int)
  • enemy life bin from 0 to 10 (int)

How find PM match?

  • Look for exactly same data in an Arraylist<String> (index is tick) who contains bins indexin this form: (index bin 1).(index bin 2).ecc..

if exactly match hasn't been found:

  • get all same data with same bin index
  • find the most similar results for every bin
  • look how much enemy has moved from the most similar result until predicted hit

How find similar result?

  • sum the distance of data bins index and sum it
  • find the lowest sum
  • if there are many low sum do the same but using the distance between absolute value and not bins index

I hope you have understend^^ ==lestofante 13:11, 26 February 2009 (UTC)