General Misc > Misc

Solar Roach-Bot Project Updates

(1/3) > >>

Gopher:
Instead of posting updates scattered around in my old question threads (which have all received enough answers already) I decided to make an announcement thread.

Roachbot

Goal
To create a simple 6-legged, autonomous, solar/battery hybrid microbot capable of living out it's empty little cockroach life with minimal human involvement. Moreover, to create this system without the benefit of a microcontroller, as a pure hardware system.
 
Description
The Roachbot will be a novelty/toy, meaning it will serve no useful purpose. It's high-level behavior can be described as consisting of two states: "Hunger" and "Fear."  In the "Hunger" state, it will seek a bright spot to sit. Like a real roach, which spends most of it's time searching for food, the roachbot will spend most of it's time in this state, passively soaking up light to charge it's main battery. The bot will switch from hunger to fear in two ways. First, if it's battery is fully charged it will stay in "fear" mode, hiding in a dark spot; or when it hears a loud noise, it will switch to fear mode (unless it's battery is critically low). Switching back from fear to hunger will occur when the battery power drops below a minimum level, or when no noise has been  detected for a period of time. In addition to this basic behavior, it will have a pair of "antenna" which it uses to detect ledges. When a ledge is detected, it will back up and turn away to avoid falling. This allows Roachbot to be kept on a desk or table, making it an interesting conversation piece.
 
To add interest, I would like to add a set of dip switches allowing it to be switched between normal (autonomous) mode,  always-on (walks constantly, ignoring light but still avoiding ledges), recharge-only (no movement, but still charges battery from solar), and off.

Unresolved Issues and Potential Problems
Much of the design has been worked out, but a few major elements remain undefined
* Fear timer
Without a microcontroller and given the tiny scale planned, a real timer is impractical. I'll be playing around with ideas over the weekend (10/6/06).
* ledge-detection and avoidance
The thing that complicates this one is the "spastic" style of locomotion, which will not provide a very stable platform for mounting the sensors. I have a solution in mind, but it is unproven, and can't be tested until the base chassis is up and walking.

Status
I've bought most of my parts today, though a few key items are still missing, including voltage triggers, voltage comparators, photoresistors, and pins onto which the gears can be mounted. Tomorrow morning I will be going out in search of as many of these components as I can find. Any I am unable to track down will have to be ordered. Until I get at least some of these components, I am unable to start constructing the chassis or testing most of the circuitry, but regardless of what I find my principle task this weekend will be formalizing my circuit diagrams. This will allow me to calculate the actual resistance and load of the bot's circuits, which will tell me if the battery I'm using will be sufficient. I will also be able to produce a complete, detailed list of parts, which will allow me to calculate both the cost and the weight of the final bot.

Expect regular blow-by-blow progress updates.

Charlie1138:
This sounds incredible. I want to build one too!

What do you roughly estimate all the parts will cost? Just a general ball park figure.

Gopher:
Hard to say at this point; I'll have a better estimate after I put together my complete parts list. Yesterday I bought a lot more stuff than needed at Radio Shack, like 150 assorted resistors, 30 each of PNP and NPN transistors, etc. (When you can buy 3 for $2, or 15 for $6, it's kindof a no-brainer!) Also a new multimeter (I've got one somewhere, I think it's in the closet at my parents' house, but this one is a little nicer and it was still cheap) and all 4 of the ZZMT engine kits (I certainly won't need 12 motors for this bot, but it's nice to have extras and they were just so cheap!)

I think it's going to fall in the $40-$60 range. If you're not as clever at scavenging parts as me, might be more; if you're better at it than me (which isn't hard), you might build it for less. There aren't really any expensive parts, since it won't use a microcontroller.

I'm thinking about writing a couple of tutorials on the project if I get it all working; One on constructing the basic 6-legged walker chassis, one about the makeshift tactile sensors (I quickly figured out from my research that DIY seems to be the only way to go with tactile sensors for microbots, but it's a lot easier than you might think), and one about the brain that wires it all together.

Now I'm off to hit another radio shack, HobbyTown, Sarge's Army Surplus, and any other places I can think of that might have any of the parts I need.

Charlie1138:
Sounds great. I seriously can't wait. This sounds like it would be right up my alley! Especially with a tutorial. :D

Gopher:
Yeah, it should be an interesting and educational journey :)

Got a late start yesterday, and my shopping expedition was less than perfectly successful. I didn't have high hopes of finding voltage triggers or comparators, but the radio shack that was supposed to have photoresistors, didn't. Amusingly, they checked the computer and it claimed the location I went to the day before had them. Ended up going all over town, and didn't have time to do any work before meeting some friends.

:sigh:

Otherwise things went well; found the materials I needed to start building the chassis. Experimented this morning with ways of cutting and shaping the plastic and metal sheets, and now I'm going to start making parts. It's probably going to take some experimenting to get the gears mounted firmly and accurately enough to avoid twisting and slipping when powered, so this should keep me busy until I can get the missing electronic components.

Navigation

[0] Message Index

[#] Next page

Go to full version