Author Topic: My Craziest idea ever  (Read 3382 times)

Offline jknilinux

  • Bot Destroyer
  • ***
  • Posts: 468
    • View Profile
My Craziest idea ever
« on: November 08, 2008, 05:59:59 PM »
Hi everyone! You've been warned.

So, I was thinking that it should be at least possible for an embedded processor to run Darwinbots code. Then, I realized- add two motors and 2 sensors, and you have a real-life darwinbot. AND, it should be even easier than that because there are robot kits that are commercially available that could do the trick inexpensively...

 Good choices are the e-puck robots from k-team.com or the cybug robot insect from jcminventures.com with the hbf-3 add-on board. The HBF-3 has a microcontroller that we could program to accept DBcode, just like how the Basic stamp microcontroller from parallax accepts PBASIC code and turns it into PIC assembly.

With the cybug, we can even have the robots feed off of each other's batteries-
(aka "eat" each other- see here: Ecosystem)
(detailed here: Upgrades)
relatively easily.

They're also inexpensive, but only have 2 IR sensors and 2 light sensors, and cannot form MBs. However, it still would be pretty neat to get even a simple darwinbot coded up in one of them. Seeing a darwinbot in "flesh and blood".... that would be cool.

And take a look at these links:

http://en.wikipedia.org/wiki/S-bot_mobile_robot

http://www.swarm-bots.org/

If we got an s-bot ( or similar) to run the DB code, then they could even form MBs! However, I don't think we could buy any just yet...


Anyway, uh... who'd like to try it?

Seriously, though, maybe it's not that hard...
What do you think?

EDIT: THE LINKS WORK NOW!
« Last Edit: November 10, 2008, 01:32:58 PM by jknilinux »

Offline Moonfisher

  • Bot Overlord
  • ****
  • Posts: 592
    • View Profile
My Craziest idea ever
« Reply #1 on: November 08, 2008, 08:55:31 PM »
If you're going to apply DB to RL robots then I would try to make them use the few sensors they have in an optimal way. Hope to see them evolve into actualy using the sensors to accomplish more than one would immagine was possible.
Basicaly what I find interesting about the idea is that robots often fail because the sensors are code by a human instead of letting the robot try to figure out a better way of using them.
Personaly when I first saw the robot that nasa peple had been working on for repairing space shutles I got a bad taste in my mouth and was wondering how people working for nasa could fail so miserably at such an easy task. The thing walked like it had a load in it's pants and used like 23 steps to get through a door, could barely keep balance, and generaly seemed like they got it all wrong from day one then kept moving in the wrong direction.
Latest I saw they where working on image reconition and such becayse they must have realised how embarasing it was to see utube clips from teens having done better at mototrics with almost no resources or fancy diplomas... anyway I could go on for hours about this, lets just say their results where so poor that I was shocked, and it seemed to me like the reason was that the team was full of physicist (Ok drunk+spelling physisist = error). It looked like they had tried to use absolute logic and physics to make the bot work, wich would probaly also require HUGE precision in making new bots because they would have to match the original exactly in order to work propperly. The ground is not always flat, and a robot won't be exactly the same each time, so trying to build it this way... well to me it just seemed ... stupid. I just expected a lot more from nasa that's all.
Anyway I was actualy playing with the idea of building small robots, so if I do then I may try to play around with this, but my goal would probably be mroe around making more of motorics and sensors through evolution or NN's, rather than utopian logic and physic scenarios that will only work in theory and never IRL.

Offline bacillus

  • Bot Overlord
  • ****
  • Posts: 907
    • View Profile
My Craziest idea ever
« Reply #2 on: November 09, 2008, 11:24:42 PM »
Obivous problems that would stop an otherwise great idea:
reproduction
body mass
defenses (not that serious, admittedly)
viruses (okay, an IR signal could solve this)
autotrophs (solar  cells, but you wouldn't exactly do this outside   )
"They laughed at Columbus, they laughed at Fulton, they laughed at the Wright brothers. But they also laughed at Bozo the Clown."
- Carl Sagan

Offline jknilinux

  • Bot Destroyer
  • ***
  • Posts: 468
    • View Profile
My Craziest idea ever
« Reply #3 on: November 10, 2008, 01:29:35 PM »
Quote from: bacillus
Obivous problems that would stop an otherwise great idea:
reproduction
body mass
defenses (not that serious, admittedly)
viruses (okay, an IR signal could solve this)
autotrophs (solar  cells, but you wouldn't exactly do this outside   )

Reproduction- all robots are connected wirelessly to a central computer- whenever a robot "reproduces", it sends the child's DNA to the central computer, which then overwrites the DNA of any least fit/dead robot with the new DNA. The child is "born".

Bodymass- You could make the max speed for the motors decrease as battery level goes up.

Defenses/viruses- The cybug robots have a 360-degree IR tranciever built-in on the HBF-3 processor board.

autotrophs- They also offer a solar cell/ wall outlet-powered stationary charger, called the "sunflower", which can be modified to feed the cybugs, is also available for like $20- see the ecosystem link.

Also, here's the coupling mechanism to trasnsfer energy that they reccommend, in ASCII art-

. . . . . . antenna on prey- connected to + terminal on huge capacitor
. . . . . ./
. . . . ./ . . . . . . \_____________ . .- antenna on predator- connected with diode to + terminal on huge capacitor
(+)--/ . . . . . . . . . . . . . . . . . . . . . .\---->|---- +
----- . . . . . . . . . . . . . . . . . . . . .-------------------
. .\ (-) . . . . . . . . . . . . . . .  . . . . . . .\(-) . . . . . . -Whiskers on bottom always connect - terminal on battery with tinfoil on floor
--------------------------------------------------------  -Tinfoil always connects negative terminals on bots

A better strategy:

 V- below are the prey coupling contacts on the left and the predator's on the right. The --->|--- things are diodes.

. . . . . . . .\ -Special feeding whiskers, this one is above bot
. . . . . . . . \
(+)-----O . .\---->|--(+) to (+) contact
 ----| . . . . . .(----------- PCB
(-)------O . ./----|<--(-) to (-) contact
. . . . . . . . /
. . . . . . . / - (-) whisker below bot


btw- THE LINKS WORK NOW.
« Last Edit: November 10, 2008, 02:15:33 PM by jknilinux »

Offline jknilinux

  • Bot Destroyer
  • ***
  • Posts: 468
    • View Profile
My Craziest idea ever
« Reply #4 on: November 13, 2008, 11:14:10 AM »
Is the energy sharing mechanism unclear? Sometimes I feel like I'm the only one who understands what my ascii art means- and that's probably my fault, too.

Offline bacillus

  • Bot Overlord
  • ****
  • Posts: 907
    • View Profile
My Craziest idea ever
« Reply #5 on: November 13, 2008, 06:46:17 PM »
I think that even with the basics, such a mechanism would be amply difficult. For example, feeding in DB requires a shot to impact a bot, regardless of where it hits, while with these bots, they have to pinpoint the plug of the prey, and compensate for motion etc.
"They laughed at Columbus, they laughed at Fulton, they laughed at the Wright brothers. But they also laughed at Bozo the Clown."
- Carl Sagan

Offline jknilinux

  • Bot Destroyer
  • ***
  • Posts: 468
    • View Profile
My Craziest idea ever
« Reply #6 on: November 13, 2008, 10:10:13 PM »
If you've looked at pictures of the cybugs, you'll notice they have two whiskers at the front which are connected directly to the batteries. With the original feeding strategy, they run around on aluminum foil such that the foil is always connected to their negative terminal. That way, all they need to do is touch their whiskers together to share energy. A diode on the predator can ensure that it's the one gaining energy. So, it's actually very simple. It's shown in the first picture.

With my idea, the robot is seen edge-on. It's basically the idea shown in the predator + prey picture I'm showing here (number 3). They don't need to couple the connectors exactly, because the connector spans the entire front and back of the robot.
« Last Edit: November 13, 2008, 10:13:20 PM by jknilinux »