Getting Started...?

[PSvilans]

Hey all!

I'd like to start out just by introducing me a bit...I've been interested in robots and AI most of my life, and with that I'm really into programming (mainly in Actionscript 3.0 (yeah...whatever...) and Java, but I've also mumbo jumboed in C++, specifically with robot interaction with the FIRST robotics challenge (through highschool). To me, the FIRST robotics challenge was highly unsatisfying, since it's dumbed down, the challenges aren't too original, and frankly, it was more hardware than software.
In terms of what I've done with AI, I've only been working with mainly pathfinding and some logic, and a bit of GA's(Genetic Algorithms)...

Anywho, what I wanted to know is the best point for starting out in building some simple robots. I've checked out Arduino, and it looks like something solid to start with, though I'd prefer to start out with something in C++ as soon as possible, just so I can start getting used to working with it...but I guess a lot of simple stuff doesn't travel through c++ too often.

Where did you guys start?

Also, links to awesome articles getting into specifics about robots and hardware interaction, or anything related to AI, feel free to shoot my way!

(Sorry if this thread is too...amateur? Not sure on how professional of a level this forum is...)
Thanks!

P.

[bens]

Hello.

If you're more interested in the programming aspect, you could start out with something like the 3pi robot.  The electronics and mechanical side are taken care of for you, so you can jump right into the programming and do some rather fun things with the built-in sensors, such as work out an algorithm to find the shortest path through an arbitrary looped maze.  If you want to make things more complex, you can always expand the platform by adding your own electronics (see the videos at the bottom of the product page).  The 3pi will be available for $50 during the Pololu Black Friday Sale to the first 50 customers ($75 after that for the duration of the sale).

- Ben

[knossos]

I would have to say SoR is what inspired me to build robots.  I decided to start out with the Axon II since it would give me a good starting platform that was also powerful enough that I wouldn't outgrow it any time soon.  You will find all levels of experience on SoR.  There are also some excellent tutorials that you should really look through if you haven't yet.  You can use C or C++ with the Arduino and other related boards.  I use C++ with the Axon, and Webbot has been talking about changing his Webbotlib over to C++.

[azy]

if you search this webby there is a huge collection of excellent hardware howto's...everthing from modding serovs too giro drift

cant help you out on the hardware side as im total noob in that respect, but when it comes to strong AI , thats my thang

& theres the 50$ robot kit, if u just want to wade in and get yer hands dirty.  Theres just as much an art to hardware as there is to genetic algorithms, but when u can marry the two skillsets you can get something going on

I built my first linefollower back in 83, the programming equivilent of "hello world" , so i understand your desire to step up from that level

http://www.societyofrobots.com/robotforum/index.php?topic=12597.0

[PSvilans]

if you search this webby there is a huge collection of excellent hardware howto's...everthing from modding serovs too giro drift

My problem exactly! I just thought I'd ask the pro's where it's really best to start!

Thank you all for giving all the excellent links and infos! I did a fair amount of research over the past day. I'm still not sure exactly which one (still leaning to Arduino, as it's extensible and doesn't seem TOO hard to setup for even just a simple robot that blinks different kinds of lights as a starter...), but you guys are definitely helping me narrow the choice!

The 3pi demo's looked really really cool! Loved the pathfinding one!
The Axon II also looks good, though if I were to get that, I think I might as well get Arduino, and also thanks for informing me on the C++/Arduino thing...I guess I missed that amidst my amazement at looking through demo's/info on the web.

AI is definitely the main thing I'd like to develop through all of this, since AI in game's doesn't really become useful in reality...which is why I finally want to apply everything to something real!

Thanks again! My research continues. Again, any articles on robotics/AI, I am more than interested!

P.

[azy]

im new to the hardware stuff myself........too much choice & too much infromation...job to know where to start

what are you aiming for ?

do you want the brain to be smart code embeded just in the boards

or do you want the brain to be running on the computer/netbook & the boards just acting as input output sensors , arm wheel controlers

these are the two main ways of doing things as far as i can see

It is also possible to have a mobile bot feeding video & data wirelessly back to you basestation pc. A larger bot could carry round a netbook and use that as the main brain

[PSvilans]

Pfff...as an ultimate goal, I have no idea. I mean, I have ideas, just I'm way too far off to be deciding that; I have my whole life ahead of me to either plan my own project, or get a job that's related to this exact thing. I'm just thinking that I should actually start doing something in the field as early as possible, and at least start experimenting somewhat. Plus if I do come up with something nice, it would look good on a University application.

As a starter, I'm going the simplest of the simplest ways, just program an Arduino(or other) board, so it handles logic and input/output. I feel insecure about doing the whole program to hardware communication, and would eventually (years ahead) like to work at an even lower level than Arduino. My main interests currently in AI are neural networks and adaptive/learning behaviours. I plan to start a game project soon, where AI would eventually adapt to the players style (player likes to run and gun, AI eventually starts hiding and waiting more, player likes to sneak around, AI starts to do the running and gunning, something not TOO complex, but something where evolution is obvious. Plus if the game is even marginally successful, it could fund my robotics hobby.

Since you have a background specifically on AI, it would be cool to consult/collaborate/just share ideas with you in the future! I saw your project on actually building a pretty big robot and giving it a brain and what not...

(And another thing that may interest some people, one of my biggest inspirations is Boston Dynamic's Big Dog...looks like an animal made of metal.)