DPRG
DPRG List  



[DPRG] DPRG ORG 08 Photos

Subject: [DPRG] DPRG ORG 08 Photos
From: David dpa at io.isem.smu.edu
Date: Sun May 18 10:09:45 CDT 2008

Howdy,

Randy M. Dumse wrote:
> One of my professional short comings is I was never
able to getinto the 68332 TPU.
> ...
> So I'm curious about some of the technical details,

jBot's microcontroller is a Mini Robominds 68332 controller
with 512K battery backed ram, 512K of flash and an SCI
periferal 7 channel A/D and 20x2 LCD display.

The 68332 is really four processors integrated onto one
chip: the CPU32 core, the TPU (16 channel timer processor
unit), a serial I/O processor, and a system control processor.
These can all generate and pass interrupts to each other.

The TPU on jBot uses 4 channels for two four quadrant decoders,
(two wheel encoders) 6 channels for three RS232 serial ports
(IMU, GPS, and yet to be used CMUcam), in addition to the
system processor's RS232 terminal port, and the remaining pins
are for four Polaroid sonar rangers timing,  which also uses a
couple of unused SCI signal pins.  Mark also provide chip
selects for driving the onboard LCD and A/D in hardware, which
are handled by the integrated system processor.

The sample rate for the wheel encoders is 20Hz and the TPU
interrupt reset switches between 4 and 2 quadrant decode when
the velocity is greater than/less than 2000 counts per sample.

Analysis of feedback from the participants of the DPRG ORC 
reveals the following hardware choices from the winning
robot entries:

Erik Petrich: home brew Luminary Cortex ARM
John Abshier: Parallax Propeller
Robert Scheer: Atmega128 + fitPC/linux
Ron Grant: home brew AVR Mega-Donkey
Dean Hall: ATmega103 with Larray Barellos's AvrX microkernel

Another interesting note is that both !!Dean and Randy observe
that they spent so much time on PID control tuning that it left
little time for developing good steering control, while Robert's
winning Tarzan used no motor control at all and just set the
PWM to a fixed value.  John uses a simple P-only velocity control
which seemed to work quite well on his winning robot.

As Jeff pointed it, it's great to see so many different approaches
to the same common problem set, and bodes well for the future of
this fleet of outdoor hobby robots.

best regards,
dpa

More information about the DPRG mailing list

Copyright © 1984 - 2006 Dallas Personal Robotics Group. All rights reserved.
Website Design by NCC

For the latest robot news visit robots.net