Let's Make Robots!

Cheap and small swarmbots

So. I'm finally building swarmbots. My goal is to have bot that is small (should be 25mm x 25mm, will be ~30mmx30mm) and cheap (less than 50$ per one. even when building only one).

 

 

It has to be cheap because - i have no job or funding and i would like to make at least 5 of them. And small because small components are cheaper (eq. small motors) and cheap distance sensors(self built) work in very small range. I have found acceptable method of making PCBs (with photoresist. Local art schools have UV lamps..) and i think i can now solder well enough to solder SMD components.

 

The Robot:

Mechanics:

For now, it has only 2 motors. 2 x 3.7g micro servos. Servos have been modified for continuous rotation and smaller size. For two front wheels (with motors) I use gears. In the back, for third wheel/caster, will be loop of wire, connected with Ground for charging. On the motors, there is PCB and on that will be 400mAh lipo battery. On the top there will be antenna like thing for charging voltage.

Problems:

  • I have no idea if SMD components and PCB and everything is going to work. So far, like every component is untested and i have no practice with them..
  • I want to add an arm or gripper or something like this but i have no good ideas. Any suggestions? Maybe something like the Antbots have.

Electronics/Brain:

I plan to use Atmega88 in some freaky small SMD package. There will be low-dropout 3.0V regulator for electronics and MAX1555 to charge lipo. And some leds for debugging and light show. And thats about it..

Problems:

  • I fear lipo charging.



Sensors:

Four infrared leds with phototransistors that work like proximity sensors, i also hope they will detect the absense of land (aka a cliff) in front of them, my tests have proven that it is possible. Also robots can communicate with each other, using them. And, i have idea of measuring the background IR radiation from all sensors, with that i can say where is the lightest spot (for IR. like a window, a lightblub or a beacon) that will be useful for measuring direction or finding things. There will be one sensor on the rear and three on front.

3mm leds for sensors

Also i will have one 38khz modulated ir sensor on the robot. I will block its range of vision to about 5 degrees so it can be used to find direction of far (1-4m) modulated IR beacons. Useful for finding home/charger.

Problems Ideas what to add:

  • Microphone or speaker. maybe..
  • Current sensing for motors. Only 1 resistor per motor and so much(?) data.
  • Odometer - finally something useful.
  • More ir sensors - you never know when there is enough.
  • Some weird magnet sensors
  • Some weird 'lets-test-if-the-ground-or-something-in-front-of-me-is-an-electrical-conductor' sensor. like the ones antbots had. for finding food and special places.


So, ideas, thoughts, encouragement are welcome.

 

3. Jan.

I put up pictures. First one is my schematic, seems to be fine. On second picture there are 2 cont. rotation micro servos with gears as wheels. Third pic is 3mm leds and phototransistors for sensors.

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

Your links are not linking what we wanted them to link. Please check.

 

Sorry about that, seems that I pasted something else I had in my clipboard instead of the proper links. 

Corrected.

Long ago I have started a swarm project but did not finish it, because I got sick of making the first prototype with tiny wires on a proto board instead of a PCB. I will get back to them sometime when I'll be in the mood of designing a PCB. $12 for 10 boards at iTeadStudio is well worth the waiting time.

Here is the link on my blog for the inspiration: http://seriousrobotics.wordpress.com/2010/05/14/new-project-swarm-robots/

and here for the first Drone prototype: http://seriousrobotics.wordpress.com/2010/05/15/first-drone-prototype/

BTW. I named them Drones, but in Beekeeping, the drone is the male that does nothing for the hive, just lives to mate with the queen. The workers are doing everything. I will be renaming them properly and design the PCB to look like an insect. I was also thinking to design the "queen" as a Wall-E that will stack cubes gathered by the workers from the field.

What if you gave each robot a specific binary address? You could connect all of the IR LED's up to a transistor (The transistor is driving all of the led's) and made each robot turn the IR on and off, such as a signal on a carrier wave. Each signal would be unique for each robot, causing robots to be able to "identify" the Infrared light around them from the other bots. Then, you could in software look for that unique signal.

I have done a few experiments using Mr. Generals IR sensors for both comunications and object detection. It can be done but I had two problems.

1. If a nearby robot is detecting an object then it will interfer with another robots communications

2. my IR LEDs and photo transistors only work over a small angle (about 15 degrees) so they must be pointing almost directly at each other for communications to work

My solution to communications being interfered with was to program the robots to only check for objects once every second allowing enough time between checks for communications to occur

The best solution for the angle problem was to use more LEDs and more phototransistors to cover a wider angle. 

 

Good start! This looks like a very promising start!