Here is a better summary of where I'm at with my project. The source is an email to the TCRG list -
> ps -- Rob, how is Robocities coming along? If you need some help, the > group is ready to step up and provide whatever you need. Right group :?) >
Robocities is making me cry. Major engineering hurdles, no time. Maybe the group could take my wife and kids for a few weeks? :-)
Seriously folks -
Bruce I'm looking at constructing not so much a city but a ziggurat, kind of like a square tower of Babel out of pink foamular(tm) insulation foam (how strange is that). Unfortunately planning out an entire city would be just too hard for a protocity, so if this structure is too weird for the Science Museum, hey I understand.
What I'll need is a 11' x 11' space, about 11' high. I'll need another 10' x 10' space to put the pre-build material. These dimensions can be scaled down as needed.
What we're looking at, with as many corners cut as I could manage, is 4 robots; 1 to place bricks, 1 to deliver bricks, 1 to deliver roof/floor tiles, 1 to deliver ramps. The walls will be 16" high making each level 17" high with the tiles. The roof tiles are approximately 32" x 32", the individual bricks are 8" x 4" x 8". 7 levels total with a 32" x 32" spot on the top of the building for the robots to sit when the structure is complete. The walls will actually be laid as "hallways" or "tunnels" that run through the structure; I currently have no plan for building a closing wall around these tunnels so you'll be able to see through the ziggurat structure.
The whole system is based on the brick laying robot because that is the robot that will move the least. One challenge was to determine if this robot would give direction, or if a computer would control the show. Having the robot give direction to the other robots was really attractive because then I wouldn't have to program any other systems like a computer. But, because I have some calibration considerations and yet I don't want to lug a computer around, I've decided to program a PDA to do the decision making in an extremely centralized method. I expect to change this in future upgrades (famous last words).
I'm trying to create an initial system with the least complexity possible to insure completion by the deadline. This means I'll be doing a lot of experimenting in the future weeks and then adding hardware/software as needed. For example initially I want to see if dead-reckoning is enough to get the protocity built, then I'll add an accelerometer, compass, etc. What I don't want to do is add stuff that I think I'll need and increasing the complexity.
Mechanically, I finally built the arm design I like the most for the brick layer, wired up the base, and I'm adding sensors and getting ready to program. This is the first robot, but don't panic yet (let me do the panicking here), this is the hard one, the other three robots will be significantly easier to build and program after this one is done.
Another problem is preparing the bricks and the tiles, and how to configure them as pre-build material that's accessible to the robots. This will probably be the last thing I work on, which means this is probably the thing that's going to get skipped and I get to hand my robots material by hand (bummer). Oh and I doubt the robots will have enough battery power to complete the structure, so I'll need to swap out the batteries without having the system lose it's place. Charging the robots is currently not an option since this will increase the complexity.
Rapid-Prototyping Boards have been a really big help, but there are still some flaws that are being worked out like my motor drivers stack on the motor interface board poorly.
Lastly, I was just relating to Jeff Sampson earlier that I don't think I've ever completed a robotics project to this degree, so some funny stuff is coming up like what software component to program first.
To sum it up - overall I would say I'm about 6 months behind but progressing nicely.