Targeting, Prediction

[ufo_1_98]

Ok, My first time posting here (I just signed up) and I wanted to see
how you guys felt about my approach to targeting && predicting movement.

Ok so here is my theory:

Scan once to find velocity, heading, and distance of the enemy. The
distance will be used in the Law of Cosines (google it if you dont
know what it is). I will scan again to monitor changes in velocity,
for acceleration and deceleration, and changes in heading, for
turning. The second distance will also be used in the first law of
cosines equation.

Ok so now I have whether or not it is turning and/or slowing down or
speeding up. Lets assume the robot is moving a constant speed at a
constant direction (which is easy to predict but for the sake of
theory please bear with me).

I can use the Law of Cosines to determine at what point on the field I
will have to shoot at because I can form a second, predicted,
triangle, using all the given information. Law of cosines will help
determine the angle at which I need to shoot and how far it is. Of
course this angle will have to be modified if he is changing speed
and/or course, but that is the 2nd phase of this implementation.

My question now to you is this; Is this a "hard" approach? Has anyone
done this or something similar?

Thanks, Michael.




 
Yahoo! Groups Links

<*> To visit your group on the web, go to:
    http://groups.yahoo.com/group/Robocode/

<*> Your email settings:
    Individual Email | Traditional

<*> To change settings online go to:
    http://groups.yahoo.com/group/Robocode/join
    (Yahoo! ID required)

<*> To change settings via email:
    mailto:Robocode-digest@...
    mailto:Robocode-fullfeatured@...

<*> To unsubscribe from this group, send an email to:
    Robocode-unsubscribe@...

<*> Your use of Yahoo! Groups is subject to:
    http://docs.yahoo.com/info/terms/
 

[Thomas Dalton]

> My question now to you is this; Is this a "hard" approach? Has anyone
> done this or something similar?

It sounds like what's commonly called "Linear Targeting" - you assume
they're going to keep going in the same direction at the same speed
and fire at where they'll be once the bullet reaches them. It's the
2nd easiest method (the easiest is Head-on targeting, which is
shooting at where they are now). One step harder is Circular
Targeting, which is much the same as Linear targeting, except you take
into account their change of heading and assume they're moving in a
circle (which most robots do, at least to some extent, as they try and
move perpendicular to their enemy). After that, you start getting into
targeting systems that actually learn about the enemy's movement and
work out where they're likely to be based on what they've done in the
past.

See http://robowiki.net for more info.


 
Yahoo! Groups Links

<*> To visit your group on the web, go to:
    http://groups.yahoo.com/group/Robocode/

<*> Your email settings:
    Individual Email | Traditional

<*> To change settings online go to:
    http://groups.yahoo.com/group/Robocode/join
    (Yahoo! ID required)

<*> To change settings via email:
    mailto:Robocode-digest@...
    mailto:Robocode-fullfeatured@...

<*> To unsubscribe from this group, send an email to:
    Robocode-unsubscribe@...

<*> Your use of Yahoo! Groups is subject to:
    http://docs.yahoo.com/info/terms/
 

[ufo_1_98]

--- In Robocode@..., "Thomas Dalton" <thomas.dalton@...>
wrote:

>
> > My question now to you is this; Is this a "hard" approach? Has anyone
> > done this or something similar?
>
> It sounds like what's commonly called "Linear Targeting" - you assume
> they're going to keep going in the same direction at the same speed
> and fire at where they'll be once the bullet reaches them. It's the
> 2nd easiest method (the easiest is Head-on targeting, which is
> shooting at where they are now). One step harder is Circular
> Targeting, which is much the same as Linear targeting, except you take
> into account their change of heading and assume they're moving in a
> circle (which most robots do, at least to some extent, as they try and
> move perpendicular to their enemy). After that, you start getting into
> targeting systems that actually learn about the enemy's movement and
> work out where they're likely to be based on what they've done in the
> past.
>
> See http://robowiki.net for more info.
>

Yes, I know about the wiki. From what you explained it seems as mine
is a combination of Circular and Linear, because when its done it (in
theory) should target both.