"Realistic" precision.

mlw · 2005-04-12T14:05:09+00:00

It occured to me over a few, um, discussions? where I think our argumentsare coming from. It is sort of interesting, but think about it.I come from a very strong engineering background, while I know the theorycold, I implement from a very practical perspective.Take for instance a discussion about a robot moving at 4 miles per hour, andhow 0.01 seconds represents 3/4 of an inch, or something like that. Yes,all these numbers are true, and they are 100% absolutely correct, but is ita practical and realistic expectation of precision?I don't know about you, but my sub $500 robot is not a precision instrument.Chances are, your robot isn't either. At Denning Mobile Robotics, we spent months trying different types of wheelsmade from rubber, urathane, and plastic, inflated and solid, differentdiameters, smooth or with tread, flat or conical, just to conclude thatthere is no reliable way to know the robots true position based on wheelmotion, even on a controlled surface.Once the navagation team realized that there was no reliable precision inthe position based on the wheels, they had to think differently. They coulduse the wheel position as an approximation that degrades over time. Withsome corrective process periodically. (beacons) The "realistic" precision is quite low. So, in the end, you can only get"so" close no matter how much precision you apply to the problem. It turnsout that you can get pretty darn close with a hell of a lot less. That ishow I have approached my robot. Take the dual wheel differential drive. Even if there was no variation inwheel diameter, if I could keep .001% variation in motor speeds, the robotwould still drift because of surface imperfections (dirt, rugs, dog, kid).How much better could it realistically get? Seriously, think about it.The motor control will work fine 99.999% of the time, every once in a whileit may miss a beat or so, but the code accounts for it, and willcompensate.

Apr 12, 2005 20 Replies

Rate this thread:

I'm building my first robot atm, but have also been thinking about precision. This will depend on the situation the robot is to be used in. The navigation people personally use varies. If we are outside in a field wanting to get to the other side we just wander in that general direction using obstacle avoidance for trees, cowpats etc. Bring the scale down and move into a house and we move slower, watch our footsteps, hands grab doorhandles etc. Our motor precision must increase. Then try walking a tightrope, most of us haven't developed the skills for this as it's not generally required. We use different senses to get around so a robot is also better off with as much relevant information about it's surroundings/position as possible. Maybe because it's my first bot that I can just casually toss the term 'relevant information' in, I'm sure I'll find out :)