In researching the current state of personal (hobby?) robots I looked at an unbelievable number of sites. One link would lead to another, which would lead to another almost ad infinitum.... The number of people experimenting, and doing some great work, with robots in their home workshops, or on the kitchen table is really impressive. It reminds me a lot of the late 1970's right after the introduction of the MITS Altair 8800 personal computer kit.
After looking at quite a few projects I started to realize that a large number of the robots took on a circular shape, and most of those used two primary drive wheels with a third, and sometimes fourth passive caster type wheel. In many of the designs the wheels seemed to be positioned right at the robot chaiss perimeter. The Rouge Blue Robot kit, that I plan on using as an initial platform, uses this approach as well.
I naturally assumed that there must be some good, logical reasoning behind the design and made a mental note to look into it later. I wanted to figure why everyone seemed headed down the same path. Then, this morning, I happened across the Dallas Personal Robotics Group site. I found a reference to their site on the Seattle Robotics Society site, and as usual, I clicked on the link and started exploring. Both sites are chock full of helpful articles written by robot experimenters, and contain a weallth of good information. The writers are extremely knowledgeable, direct, and often humourous. They tend to report both their success and their failures - even the big disasters. I really chuckled reading about how one experimenter's robot did some major damage to the cabinets in their kitchen because it moved too fast for the contact sensors to stop it in time. The flavor of their writing is a welcome change from the PR put out by 'real companies' that always gives the impression that they walk on water and never make a mistake.
And there, buried in an experimenter's article was the simple answer to why all these robots are round with the wheels on the perimeter. It makes the software design much simpler when you're trying to build routines to maneuver the robot through a maze, or to attack a competitor during a sumo match. If the robot can turn within it's own footprint, and has round edges, then you don't need to worry about getting hung up on corners. With other designs when you come to an obstacle, the typical response is to back up a short distance, turn a bit, then try again. But, if the robot is round and can turn without having to backup, and doesn't change its center position (location) in the process, then the whole problem is a couple orders of magnitude simplier.
Really cool... Learn from other people's experiences - success and failure. Incorporate what is useful, store the rest for later use. Move on to the next step. Assume that a solution for every problem exists - some place, some where. Then your challenge is to find it. Assume that some one on the planet has already solved the problem, or something close enough to it to be useful to you. Go and find out what they did and harness it to solve your particular problem. Then build on it to advance to the next step. And, pay them back - document what you did - add your learning to the general pool so that others can benefit from your experience.
"The design which has emerged is a small dual-differential drive platform with the geometry of an 11" circle. The drive wheels and tail caster sit on the perimeter of this circle, and thus it can rotate in it's own space. This greatly simplifies software for maneuvering and collision avoidance."