So I suddenly have the itch to build a telepresence robot.
I have never built anything resembling a robot before, but I dream in machine code, so my design is likely to emphasize custom software solutions over custom hardware approaches.
Here's the idea. Feel free to tell me how insane I am, or give tips as to how to proceed, etc.
The goal of the robot is to be a limited version of me, 2000 miles away from me.
It would have some mobility, cameras, mikes, speaker, and a screen at eye level to render my face.
I would not be able to do any maintenance on the robot for months at a time, although I could rely on some folks' benevolence to handle minor tasks if needed.
The robot would have almost no autonomous behavior, and would be remote controlled through a wifi network available in the entire space.
It would navigate in an office environment (flat ground and elevators at least. possibly doors and stairs later)
It would need to be able to recharge its batteries easily (probably by reaching a specific location, then it's roomba-style contacts or maybe some nifty induction thingy.)
So far, I'm leaning toward a setup with 4 wheels, a platform holding the batteries and the embedded computer, a pole to bring the I/O stuff to head-level, and the cam/mike/speaker/screen gears attached to the top of the pole.
That doesn't account for pushing an elevator button, but the space will be shared with many people, so waiting/nagging would eventually get the desired result.
(although if there's an easy way to get a little arm in there, I'd definitely go for it.)
I'm leaning toward an ARM setup running linux with various off the shelf components for the I/O stuff.
The skeleton, wheels, battery and charging systems are still very undefined.
That's all I've got right now. I'm planning to take my time building it, in part to make up for the lack of domain familiarity.
All advice is welcome, although I'm hoping I won't get too many "stick to your compiler, code monkey" replies.