that is what I was suggesting, although the workstation would probably ben sending not only a positive/negative target acquired message but some relative position information as well.
um, this is a ton of work. you may want to do a bit of reading and research before heading down this path. not to discourage of course but just to make sure you are aware of the amount of effort involved to go where it sounds like you might be wanting to go.
I think I will try that later. especcially considering this is my first robot, just wanted to get a goal to acheive through incrementalism.
Alternatively…
You could put something on your cat that would aid your robot in finding it.
Stick an emitter of some variety (Light, IR, sound, etc) on your cat and have the reciever on the bot.
Then program the bot to “seek” a stronger signal.
My cat walks around with a bell on it’s neck. would it work? then I just make a sound sensor on 4 sides of the hexapod or three whatever, and program to follow whichever sensor has the highest value?
anybody ever thought about using maybe 2.5 or 4 inch tubing for the ends of the hex as opposed to the 3 inch lengths in the tutorial? just wonder what would happen.
I’m planning on having my bot distinguish between cat/dog/human by using a TPA81 thermopile and sonar sensors. They all operate with I2C so using the ATOM shouldn’t be a problem in terms of code length and I/O pins.
By using trigonometry with:
-the height of the bot,
-angle between the ground and the warmest part of the body (often the head or torso) and
-distance from the sonar rangefinder
I should be able to tell the difference between different animals around the house and take appropriate action, hopefully using a nerf missile launcher - only for humans, of course, not animals!
If the set-up is accurate enough it might be able to tell the difference in people’s heights and react differently depending on who it detects.
As far as using longer bits of tubing, any IK program you use will have to be changed accordingly. The turning moments exerted on each servo will also be greater - this might not be a problem if you use high torque servos though. I’m currently designing custom servos around the PGHM/GHMs from Lynxmotion, using Openservo control hardware and software. This way, I should be able to build something big enough to interact with people properly, although movement will be slower.
Well, ideally, you’ll want a constant, or at least frequent, sound.
If your cat hasn’t figured out how to walk around without jingling constantly, then it should work.
My old cat would flick her head and catch the bell in her mouth.
Then she’d quietly prance around the house.
I think she found it hard to pounce on things that heard her jingling a mile away.
Anyhow…
Here’s the thread that Mike created when he was building his audio sensing circuit.
lynxmotion.net/viewtopic.php?t=198&start=0
He used two microphones and had them compare which microphone’s loudness was greater.
You might want to PM him and ask him whether or not he’s got all of his sensor board’s reserved.
Last time I checked, he was getting a bunch of them, but needed people to pool together.
With that board, you’d have all the necessary circuitry to interface two microphones, a range sensor, and a speaker.
Ideally, it was meant for a biped, but there’s no reason that a hexapod couldn’t use it.
I expect the board to come in this week and then I need to assemble it and test it out.
The sound sensing portion of the circuit, can react to sounds coming from either the left or right mic element. I suppose someone with strong programming skills can make this device find the angle of a sound source mathematically by taking the two values from each mic, and running an algorithm based off those values. Also for tracking and identifying a sound source, I think this would prove to be very difficult.
What can be done is have the bot change directions towards the sound source, and walk in that direction, or perform a movement routine of some sort. It can even be used to trigger a programmed behavior as another example. Through code, I suppose a lot can be done for different application uses. I’m not very good with advanced programming. I’m an “IF THEN ELSE” kinda guy.
hey dwalker! is there any more details to that? links? how did you get your nerf missile launcher to work? does it reload itself? I really like your idea but can you post some code or somehting so I can get a better I dea in my head? I bet if you tried hard enough you could get one of those nerf guns on a pan/tilt assembly on top of the bot. 
and yet another bot worthy of “I’ll be back” is born."
hi, I don’t have a missile launcher working yet, it’s just something I planned for later. I’d actually prefer to use something spring/elastic band based because it would be easier to fire than compressed air. I’m not aiming to hurt people, just annoy flatmates
I’m trying to get a walking base working before I add lots of sensors to it, but I know I definately want to use the TPA81:
robot-electronics.co.uk/htm/tpa81tech.htm
It outputs a 8x1 pixel image that shows how much hotter than the surroundings an object is. If you scan it around using a pan and tilt mechanism you can build up a useful image, like the one shown in the picture about halfway down that page^.
When I do finally get to the stage of fitting a missile launcher of some sort I’d like to try and attach it and my sensors to the end of a Lynx 5 arm so it can scan all around it and build up a better picture of its environment.
I guess the only problem with that would be what you’re trying to see.
Even if the sensor is good enough to detect a change in radiant temperature between a 98.6 degree Farenheit body and a 70 degree Farenheit room (which it may or may not be able to discern), you’ll probably run into trouble on hot days.
If the day is 90 degrees, then the temperature difference is only 10 degrees.
So, if you’re bot is going to be in a climate-controlled area, this would work better.
I believe that this sensor gets a lot of use with the fire-fighter bots, since the difference between a couple hundred degrees and room temperature is rather significant.
I hadn’t considered very hot days when a body may have a similar temperature to the environment. The technical data from Devantech shows that it can pick out reasonably small differences in temperature, but on hot days it may be “dazzled” by the ambient temperature. I don’t know whether moving the bot and sensor into the shade/a cooler place would help - it may be that bodies will stand out from the lower ambient temperature, but the surrounding hot air could also register.
I’ll be buying one of these sensors and some sonar rangefinders in about a month, so I’ll put them on a pan and tilt and see what works.
There is a thermal sensor available that would allow your bot to search out warm blooded entities such as people and animals. I plan to use on on my bot to make it find and follow people. 
The sensor is kind of expensive at around $100, but it would sure make an interesting addition along with an IR and untrasonic sensor, and perhaps a sound sensor.
That wouldn’t happen to be the Devantech, would it?
If not, can you link us?
