DPRG List  

[DPRG] G0t P1D? was Understanding PID

Subject: [DPRG] G0t P1D? was Understanding PID
From: Giuseppe Marullo giuseppe at marullo.it
Date: Fri Apr 28 07:18:58 CDT 2006

This works much better now:

http://www.xar.it/download/umbmark.mpg (30Mbyte)

The robot will execute two UMBMark runs ( very small 1m square path, 1m )
one at 10cm/s and the other at more increased speed.

The test is incomplete, it should run clockwise and counterclockwise several
times but you can get the idea ( erros included, it is not tuned yet).

Done with two HEDS 500CPR encoders on the motors(e192-25-12 from
micromotorssrl), two New Micros 5A H-Bridge and the fabulous Isopod!

Two PID loops, the outer will execute the speed profile (PID) and the inner
will keep orientation ( a different PI controller).
I set the loop to runs at 50Hz in complete floating point, I will verify the
exact speed.

Changing the encoders enabled me to increase the sampling time. Now I got
166x the original 300CPR resolution at wheel level.

It is not perfect but I like it now, I will try to tune both the PIDs and
the systematic errors(umbmark).

Giuseppe Marullo

More information about the DPRG mailing list