Question on feasibility for a beginner

[Xefan]

Hi. First off, apologies for the beginner post, I know they're not the most exciting to deal with but I have read the stickied posts and such regarding beginner questions and feel my question is a fair one to ask.

I would like to have a go with robotics, but I have a specific goal in mind. I've got good knowledge of AI as implemented in software on the PC platform such as topics including genetic algorithms, neural networks, symbolic AI, swarm AI and so on as well as some slightly more robotic oriented concepts like subsumption architectures and so on. I also have a good understanding of software development from large scale application architecture down to low level problem solving due to being a software developer for an engineering firm by profession. From a computer science and maths point of view I've got a strong understanding of many areas of both pure maths and mechanics with a particular interest and understanding of theoretical concepts such as Turing machines, URMs and also real world implementations built off these underlying mathematical concepts such as x86 architecture and assembly. I mention my abilities not with the intention of sounding like I'm showing off in anyway but so that people will hopefully be able to answer my question with a focus on what I don't know rather than what I already do know!

What I don't know, is really anything about actually building a robot and I've never had an awful lot of electronics experience bar replacing a few components such as blown capacitors in old CRT monitors many years back when I was a junior IT technician and had to fix them.

What I'd like to achieve is to be able to build robots where I can try implementing some of my PC based understanding of AI and watch it go to work in the real world! Specifically I have a strong interest in starting out with swarm AI (once I've built some basic robots, I was considering this course to get me a quick, easy, hassle free start, opinions? - http://www.gameinstitute.com/Introduction_to_Robotics.html). How good is the kind of hobbyist hardware you can get for robots, is it fairly feasible to build robots that can communicate with each other and perhaps even a computer wirelessly (I realise I may have to build my own peripheral for robot to PC connectivity)? Is the kind of hobbyist hardware you can get to build robots capable of handling realtime vector calculations at a decent speed for example?

My dream is to build a few robots (3 or 4) that will work together to perform specific goals, such as perhaps finding something or for finding their way out of a simple maze and relaying instructions to the other robots to aid them in finding their way out and that sort of thing. Has anyone had any experience doing this sort of thing? Is it realistically feasible for an individual to do given enough dedication to the project? Are the components required for this sort of idea affordable without completely breaking the bank but with enough money to commit to it as your average hobby?

I will likely get started with robotics regardless of this but of course my intrigue is with AI and in particular the beauty of swarm intelligence in action.

Any advice or comments appreciated!

Thanks in advance.

[pomprocker]

If you want to focus on software then go with kits

if you want to learn electronics and the mechanics, then start reading

lots of books out there on getting started on electronics, and basic mechanical engineering.

I always say read everything you can on this site, and build the $50 robot whether you think it will benefit you or not, it will beleive me.

I started that robot last February when I first found this site and fell in love with robotics, and i'm still working on it, its so upgradeable, and it must be like a $300 robot for me now

[colorclocks]

I'll guess the simplest architecture that gets you quickly into the territory you want to play in would be a bunch of cheap, dumb slaves that can move and do stuff, but can't do much thinking, and a computer that does nothing itself, but tells the slaves what to do.

One advantage of the swarm notion is that you're already committed to two-way wireless communication.  You might as well take advantage of this to offload the thinking from the swarm bots to a computer with a compiler, lots of RAM, etc, so you can start work in an environment you already know how to use.  You can emulate the separation between swarm intelligences on a single computer, get algorithms that you like, and then, when you know what computing resources are required, re-implement with distributed intelligence.

The alternative--starting with smart swarm bots--might be what you want in the end, but I think you don't want to start there, because you'd have to do all your trial development on a computationally crippled platform, and you'd have to choose the platform before you really know what capabilities are required.

Since you're not strong on the building-robots side, I'd start with mass-market, utterly stupid "bots", like cheap remote controlled toy cars, and get your computer to emulate the kid driving them (i.e., interface your computer to the hand-held controller).  This won't get you any data from swarm bot to computer, but it's an inexpensive way to start moving things.  Eventually, you'll want a microcontroller on each bot, to run sensors and send data to the computer.  Maybe the BASIC stamp-based kit you linked would be good for this, but you certainly don't want to do AI development on a stamp.  You'd spend all your time fighting limitations, and none on algorithm development.  I'll guess things like the Basic stamp are better for simple, one-off robots, and not as cost effective for a swarm--even a swarm of three or four--as a microcontroller to which you download object code, such as an AVR.