Darwinbots Forum
Code center => Suggestions => Topic started by: Ulciscor on June 11, 2005, 12:50:30 AM
-
Just had another thought which could be interesting, maybe some sort of camouflage ability?
It would maybe take a certain amount of energy per cycle to sustain depending on the degree of camouflage you would want.
The full effect I suppose would be to render you invisible to other bots but this would be at a high price. Maybe you are unable to shoot while camouflaged? Lesser degrees would maybe mess with other bot's vision, maybe make them think you are closer/further than you are, maybe a little further left or right. Or maybe you would have to be within more than one of their eye's fields of vision before they could see you.
Could make for interesting strategies with DNA I guess.
-
This idea has been around for a while. It needs working environment greed though... so, we are still waiting.
-
I know I'm being dumb but what does it need the e-grid for?
-
No, it's not dumb. It's just that whenever we introduce a new feature, it needs to be something open-ended. In this case instead of just "camouflaging" we should allow bots to detect the color of environment and change their own color. Which means that we need environement grid with different regions having different color.
People actually have not been too hot about this idea. It seems too powerful. Another argument against having comouflage is that color is a very convinient featue for us to see different species of bots. If you allow bots to change color, then this becomes useless.
-
I am for this, but only in the way Sharz suggested; in combination with the Egrid.
Btw Ulciscor dont count on it to be implemented all too soon, I suggested this two times now and little has been done with it untill this day... :lol:
-
Well, bots can always keep their white eye. That would be pretty cool. "Hey, where'd all my bots go? And why are there all these energy shots floating aroun-- Oh."
-
I don't get why they actually need to change colour, the effect could just be in respect to other bots. Maybe a bot camouflaging itself could fade slightly just so you could see what it was doing but a complete colour change is going to look a bit weird...
And yeah it would be a very potent weapon for DB but it would also be an expensive one to use or would have drawbacks, such as the inability to shoot when camouflaged.
-
The main problem I see is that currently, bots either see something or they don't. There's no gradation. A camouflage system works best in an environment where being offgreen in a green environment is worse than being perfectly green but better than being neon yellow.
-
Camoflage? I'm guessing you mean blending in with the background. Hopefully you don't mean invisibility, because that would detract from the biology part of it. I guess it would be a great combat strategy, but it's just not natural. If you mean blending in w/ the background, it would require DB to have a background color the robots could determine and defining colors the Dbots would have and see.
-
Well, as near as I can tell camouflage is invisibility be definition. Invisibility - can't be seen. If you're camouflaged well, you can't be seen.
-
One point is that camoflage only works when you don't move so it is used either as a wait and pounce stategy or simply to hide from predators.
It could be a very good addition to DB provided it is implemented properly.
Simply not seeing a stationary object isn't good enough or they would never be able to find fixed veggies.
-
It should be don't move and match the background color. We should probably provide a way for background color to be changed in some areas of the simulation by the user (or allow bots to see the color of the background pic! That could be really cool!).
Color change via muatations should be included. Every cycle you're changed from your regular color costs energy, and as you get closer to the background pic (like within a few values of red, blue or green) you begin to be seen less and less by other bots. Like if you're off color by a bit other bots can only see you 95% of the time. As you get closer in color, bots see you less and less often. Provided of course you're not moving.
And, before Carlo makes the point, this camouflage color and the color visible to the users now need to be seperate. On some level the color the bots are is just a handy visual representation, so we should have all this other camouflage color stuff be seperate, and allow the user to switch between view representations.
-
One point is that camoflage only works when you don't move so it is used either as a wait and pounce stategy or simply to hide from predators.
Oh, I see your point. I didn't see it that way. I was thinking invisibility cloak, and that means high tech stuff. I guess that does work with biology. Stupid me.
-
Yea , like the null from cosmic rift. The whole point of the advantage is that it moves while cloaked. The ristriction in cr is must uncloak when shooting. So we can make it also when shooting or tie feeding. I think multibots should have the ability to cloack too, unless they tie feed from each other (passing info. can be during cloack)
The best way to make it is to have it like a "cloak mode" option.
Idk but the more you guys talk about ordinary camoflage the more over-complicated it sounds... :pokey: I mean whats the chance a bot will evolve a color to match the e-grid and then by the time it realises if it stays still it wont get eatin the e-grid colors change again...
Better yet, the background picture is composed of tons of pixels, how do you suppose a robot that occupies at least 30 pixels is going to blend with witch specific pixel? Then it has to memorize were does it need a specific color, I thought you guys thought building a virtual map is too complex for the little things. Not to mention a virtual map as complex as an average photo.
-
It wouldn't need a virtual map. You could just make the robots transparent while camoflaged with just a little faint circle left behind.
VB is capable of multiple layers of background pics so you could even make th e bots work like a window to another layer. They could most likely also be set as filters to slightly change the background color just by anding the display.
They can also take snapshots of background (with afore mentioned filters) so that when they move, they keep the same segment of picture with them until they copy the background again. This could be done by bitblting small squares of foreground onto a background picture and moving them with each bot as it acts as a window to the background picture (with filters).
Maybe DB needs to move to directx graphics. :D
As for realism, seen any good squid or octopi lately? An octopus is virtually undetectable (when it wants to be) until it moves.
-
It definately shouldn't be possible to be camoflaged and moving. There just aren't any animals I'm aware of that can move and still be unnoticable (unless they're movement mimics plants moving in the breeze, or something like that).
-
Yes, our eyes are great at catching any directional movement even in very complicated environments. Maybe it's because we are monkeys...
The whole camouflage thing is actually quite complicated under the surface - it involves fooling pattern-recognition by brains, which is a pretty complicated thing....
-
PY, the actual programing is inportant yes , but first we have to figure out what we really want to do.
It definately shouldn't be possible to be camoflaged and moving. There just aren't any animals I'm aware of that can move and still be unnoticable (unless they're movement mimics plants moving in the breeze, or something like that).
Look , if you want to add acamouflage mode to basic Db that fine.
I think I am not a biologist so I dont know how camouflage is supposed to work, camouflage to the egrid I think is the best way...
How will robots look to the user while camoflashed to the e-grid I dont know , thats why I was thinking to paint the e-grid into realtime and update it every cylce.
I just really like Strykes idea and I want to add this:
Yea , like the null from cosmic rift. The whole point of the advantage is that it moves while cloaked. The ristriction in cr is must uncloak when shooting. So we can make it also when shooting or tie feeding. I think multibots should have the ability to cloack too, unless they tie feed from each other (passing info. can be during cloack)
The best way to make it is to have it like a "cloak mode" option.
As a "!!!!NEW SPECIAL MODE!!!!"
I mean , I feel pissed now , why can num and py play arround with silly modes ( like "shapes mode" or like "space mode") and we cant. I want actual cloack with movment , its does not have to effect ordinary DB in any way. ITS A "SPECIAL MODE USING A SPECIAL CHECKBOX IN OPTIONS"
If the mode is checked this works and mutates:
50 'number of cycles to cloak , dumps 50 energy energy into cloack
.setcloack
store
'if the robot shoots or tie feeds the mode is lost, but the energy (whats left of it) is returned. ' I know that shooting or tie feeding while cloacked is too cheap.
If the mode is not! checked! this how it looks like:
50
0
store
does! nothing!
-
I mean , I feel pissed now , why can num and py play around with silly modes (like "shapes mode" or "space mode") and we can't. I want actual cloak with movement , its doesn't have to effect ordinary DB in any way. ITS A "SPECIAL MODE USING A SPECIAL CHECKBOX IN OPTIONS"
All we can do is ask/suggest, or program it ourselves. There's no need to whine.
Henk <_<
-
I mean , I feel pissed now , why can num and py play arround with silly modes ( like "shapes mode" or like "space mode") and we cant. I want actual cloack with movment , its does not have to effect ordinary DB in any way. ITS A "SPECIAL MODE USING A SPECIAL CHECKBOX IN OPTIONS"
:blink:
All we are doing is having some fun with a few concepts to figure out what works here. Nobody has programmed the stuff yet and frankly I don't know if they ever will.
To have any shot at realism we will have to restrict camoflage to stationary objects on a colored background, egrid, back picture, whatever. Cloaking is just way too Star Trek.
-
I give up.
-
Warning: PY will not comprehand this post:
Nobody has programmed the stuff yet and frankly I don't know if they ever will.
To have any shot at realism we will have to restrict camoflage to stationary objects on a colored background, egrid, back picture, whatever.
No one seems to understand the consept of cloacking as an "extra mode". Shvartz can play arround with his ordinary biological Dbs all he wants when this extra mode is turned off.
Warning: Bots is one lazy bum bc he could of added some mutch of hes crap to Db by now its not funny. But no, even uploading takes too long for him.
-
No one seems to understand the consept of cloacking as an "extra mode". Shvartz can play arround with his ordinary biological Dbs all he wants when this extra mode is turned off.
More to the point.
Nobody likes the idea of cloaking. This isn't a space combat game. A limited form of camoflage while not moving is quite enough thanks.
-
Stryke posted: http://s9.invisionfree.com/DarwinBots_Foru...ndpost&p=718115 (http://s9.invisionfree.com/DarwinBots_Forum/index.php?showtopic=566&view=findpost&p=718115)
One point is that camoflage only works when you don't move so it is used either as a wait and pounce stategy or simply to hide from predators.
Oh, I see your point. I didn't see it that way. I was thinking invisibility cloak, and that means high tech stuff. I guess that does work with biology. Stupid me.
Yea another newbe just owned the Nobody
title ;
-
Talk about lack of comprehansion:
Camoflage? I'm guessing you mean blending in with the background. Hopefully you don't mean invisibility, because that would detract from the biology part of it. I guess it would be a great combat strategy, but it's just not natural. If you mean blending in w/ the background, it would require DB to have a background color the robots could determine and defining colors the Dbots would have and see.
:banghead: I started posting before I read that one twice, I just read it the second time now
Sry , py , all cool now