[Dprglist] Pat's robot

[David P. Anderson]

Howdy DPRG,

At the meeting last night we talked about ways to use testing to measure 
a robot's performance, and I pointed Pat to a webpage on odometery 
calibration.  But I don't want to get the cart before the horse here.

Pat, you mentioned that there might be some mechanical issues and those 
should of course be addressed first.

I've found it useful to strip down to the simplest exercises. You might 
consider finding a nice flat long place, turn off odometery/navigation, 
and just set the robot's speed controllers to run both motors at the 
same speed.  Then point it down the hallway or whatever and let it go.  
Do this a couple of times. Does it tend to drift right or left?  Can you 
ascertain that the chassis is straight and the wheels are parallel?  Is 
there slippage in the drive train?   The wheel-to-shaft coupling?  The 
tire to wheel attachment?  And so on.

I would nail down the mechanical issues first before looking at 
odometery calibration and so on.

cheers!

dpa