Date: Fri, 17 Jun 2005 02:18:04 -0500 (CDT)
Subject: battery and electronics tray
X-UID: 62
Content-Type: IMAGE/JPEG; name="img1634.jpg"
Content-Type: IMAGE/JPEG; name="img1639.jpg"
Content-Type: IMAGE/JPEG; name="img1640.jpg"
Content-Type: IMAGE/JPEG; name="img1643.jpg"
Content-Type: IMAGE/JPEG; name="img1645.jpg"

The new addition is the tray on either side of the drive motors. It's made 
of quarter inch thick aluminum and carbon steel tubes. This is overkill 
and prompted me to worry about excessive weight. Ideally, the trays would 
be carbon fiber over polymer honeycomb with composite tubes. That would 
cost around $100 to $150 for the material.

Batteries, electronics, basically everything except for lights, lasers, 
cameras and IR rangefinders will stand on the trays (strapped down, of 
course). The robot will have a very low profile. This simplifies transport 
as it fits in the trunk of my car without any disassembly.

There isn't enough clearance between the trays and the front wheels. 
Depending on the height of the front suspension, the wheels either barely 
clear or contact the trays. I did not anticipate this. I'll have to cut 
away some of the aluminum.

I plan on painting most steel tubes white. The ones used for the wheel and 
motor mounts will be painted blue. Aluminum will be left shiny as will be 
threaded rod and nuts. I don't really know if this is an attractive color 
scheme. Another thought I had was industrial yellow and black. Home Depot 
was out of yellow. Everyone must have the same idea.

Again, I feel that I'm not making enough progress on the robot as a 
system. I see all of these DARPA Grand Challenge projects. You know the 
teams are working at a punishing pace.

I'm thinking that I want to start a Sourceforge project for the software. 
There is precedent for this. Five years ago, a small team of amateurs 
started a UAV (unmanned aerial vehicle) project in an autonomous 
helicopter. They developed it to a fairly advanced degree for an amateur 
project but stopped in 2003. I had thought that it was yet another project 
that died on the vine. Instead, what happened is that the amateur work 
transitioned to a commercial enterprise: http://www.rotomotion.com/ .

I was looking through the Sourceforge robot projects and haven't seen 
anything that is equivalent to the UAV project for a UGV (unmanned ground 
vehicle). There are some robot projects. But they don't have the hardware 
investment. They aren't directly addressing the problems of field 
robotics.

I've viewed this robot as a miniaturized grand challenge vehicle. It 
should address all of the same problems but in a smaller form factor (40 
pounds instead of 4000 pounds). I think that there is a need for truly 
modular and reusable autonomous ground vehicle software given the current 
domestic interest and activity in military oriented field robotics. 
Congress has legislated (whether it happens is another matter) that a 
third of all military ground vehicles will be robots by 2015. So while the 
Japanese have a goal of building an android with the capability of a five 
year old 30 years from now, the US is focused on autonomous aerial and 
ground vehicles for the military. If it were easier to do this, then more 
people would do it. It is just like Linux. It has made many kinds of 
technology solutions practical. If you had to start from scratch, it would 
just be crazy.