Hi,
I am interested.
In my life I spent some time learning artificial intelligence.
And in last couple of years I developed decision making software which acts like a robot's mind.
The software is in production. It was done for high availability of IBM and Juniper security products.
Regarding the task, I am thinking about something truly self motivated (maybe needing no markup at all).
I mean "world" exploring strategy when the robot uses its sensors to detect obstacles and its own position/bearing and uses that information to build a map of the world.
Most difficult part here is to make him capable of knowing (or detecting) its position and orientation. Software part looks easy to me.
An idea is simple here -- robot will start its journey from the socket and will rover around building the map. But it will always know where was the starting point (socket), therefore going back to that place will be straightforward task.
Another option would be signal transmitting from the socket. So robot could radar it. This would solve the task of multiple sockets. But obstacles still must be seen for easy moving around. And radio issues, debugging difficulties and so on.
Laser censoring could be interesting. But it will not be suitable for home robots.
Those were my thoughts.
Will look forward to hear from you!
Best regards,
Gleb