Detect Edge of Table

[fuper]

Hello,

For my first robot I am planning on making a small little guy that can scoot around the table, detect an edge, back up turn around etc.

I'm thinking of using a regular infrared emitter/detector that I picked up from radio shack.

From what I understand so far, I think I use my PIC/555 to set the frequency to 38k. If I pick up a response from the detector, I infer that the signal was bounced and therefore there is something close by. If I don't detect a signal, in the case of the IR pair being over an edge then I should respond to that state accordingly.

Is this the correct approach?

 

[ikalogic]

mhh.. you said you'r firt robot right?.. well you'r quite ambitious! just kidding (but not too much! it is still very likely that you will discover that you under estimated that project)

For the 38Khz, well, as you are going to use a micro controller, you can save space/money by letting your micro controller generate that frequency. recent 8051, AVR and PIC micro controllers can do that.

good luck! and keep us posted though!

 

[fuper]

Hah! Ya I even focused on trimming down my goals. I think I can pull it off. Radio Shack has a 38k IR receiver which I will give a shot. I will just use the 555 or more likely my pic to generate the signal. I'll just test the sensor first then move on from there. Should be a fun project.

For the body I am seriously thinking of using an old 1U fan heat sink I have, and dremel out the middle fins. Gives me a metal body and a reason to dremel something since it's still in the box. Will build the sucker on a small breadboard that should fit snugly inside. I have a couple of 100 ma motors, again from Radio shack. Only thing now I'm debating is if I should build my own h-bridge (might do that just to get some experience) or go buy a low current IC driver chip.

I'm tempted to not even use a microcontroller for my first robot and just use a comparator to make the little guy turn when one of the sensors goes off. Will see.

Fuper

 

[ikalogic]

not using a micro controller will just make it 10 times harder, and you will lose a looooooooooot of time, and afterall, you'r gonna learn uC anyway..! so what not now!

 

[Krumlink]

You can take a microswitch with a roller at the end and flip it upside down and have it depressed in so when it goes off the table, you have NO or NC to deal with

 

[fuper]

I like it, thanks!

 

[billericayboy]

Roomba

Have you looked at how iRobot have kitted out their Roomba with 4 downward facing IR tranmitters and receivers to detect the floor. If the flood disapears the Robot stops and reverses.

Just what you were looking at.

 

[Krumlink]

With a DIP switch, you have no worries about noncontact issues. Say if you were riding over a very IR absorbing material (black carpet how wierd it is) it would send a false reading.

Direct contact is usually the most reliant.

 

[Norlin]

unless you have a tiled table! maybe it would lose contact in the grooves
Heh, one of my early little robot is still one of my favorites from the book "Junkbots, Bugbots & Bots on Wheels" called a "Herbie Photovore" for which you can probably find circuits on the net. Has two light sensing "eyes" and drives around heading towards the strongest light source. It has a "feeler" on the front, so if it bumps into something, it backs up and then drives on again.

 

[tank1357]

Ambitious using infrared to detect the edge on your first robot! In my personal experience, infrared can be a bit difficult to work with. Not only can it pick up interferences, sometimes it'll tell you theres notrhing there when there is, and vice versa. Until you learn how to program/tweak around this, I'd stick to touch sensors, allow something to run along the ground in front of you, and as soon as it drops slightly, it depresses the sensor. In addition, this same sensor and bumper can be used to detect if you run into anything. That way, you'll be able to use the same program and equipment for avoiding falls and walls.

 

[tank1357]

along those lines, LEGO makes some great sensors, touch, infrared, temperature, rotation, you name it. Always reliable, durable, and cheap! Pretty hard to ruin durring testing if you accidentally overload it. I'd sugest testing with these, then switch to a more sensitive instrument.

 

[Krumlink]

This thread is old, besides the original poster only was on for this. He decided to go with (I think) the micro switch, as it is very easy to interface. Problem with legos, is then you have to get a bunch of stuff to use it, ex bricks rods etc. Lego's are also not very integratable nor versatile. the touch switch is not great either.

 

[tank1357]

yeah, but one great thing about threads like these is that they can give other people ideas for awhile after the OP is done with them. Even though i just joined, I've been visiting this site for a while now. Towards the lego bit, I'm talking about completly canibalizing them. Cutting the wires comming out to hook up to your application, hotglueing the sensor, even removing the housing. And the touch sensors are highly reliable, even if they do need to be fully depressed in order to send a signal.

 

[Krumlink]

yeah, but one great thing about threads like these is that they can give other people ideas for awhile after the OP is done with them. Even though i just joined, I've been visiting this site for a while now. Towards the lego bit, I'm talking about completly canibalizing them. Cutting the wires comming out to hook up to your application, hotglueing the sensor, even removing the housing. And the touch sensors are highly reliable, even if they do need to be fully depressed in order to send a signal.

It's just easier to buy a dedicated microswitch, you can even get them with rollers.