Date: Fri, 2 Jun 2006 06:40:58 -0500 (CDT)
Subject: new remote control
X-UID: 175
Content-Type: IMAGE/JPEG; name="img2905.jpg"

The old HP-48 SX calculators have a RS-232 serial port and general purpose 
programmability with FORTH. I purchased mine many years ago while an 
undergraduate from the campus bookstore. At the time, the serial cable kit 
was a separate product and very expensive. I never purchased one. But 
before I left home, my brother gave me his cable.

Brother, thank you!

One of the Maxstream 900 MHz radio modules is connected through a RS-232 
to 5 volt TTL level converter and a DB-9 crossover connector to the 
calculator. A separate NiMH battery pack powers the radio. It's messy with 
electrical tape, rubber bands, dangling cables and twist ties. However, 
I've learned too many lessons about over-engineering prototypes. That is a 
waste of effort. Why polish something you don't even know works? Assembly 
of this remote control took about half an hour.

The remote control software on the calculator is five commands.

CLOSE - close serial port, save power
OPEN  - open serial port, initialize radio
SPEED - set upper and lower bounds on drive motor duty cycle (speed)
PARK  - stop, set drive in fowards gear, center wheels 
DRIVE - go for autonomous, repeats command at one second intervals

If the remote control should break or radio contact be lost, then the 
robot will stop after a second or two. The command string from DRIVE will 
be missed after a time window. Otherwise, the robot continues to receive 
commands to DRIVE and continues.

I expect this to allow much more interesting and aggressive testing as it 
allows me to be much more mobile. I can carry everything in one trip and 
run alongside the robot ready to stop it. Before, the rugged case and 
laptop base station arrangement was too heavy and fragile to move. As the 
robot is not reliable, I could never drive it farther than a few feet away 
lest I lose control and be too distant to recover or stop it. Typically, 
I'd set up the laptop base station right next to my car due to weight and 
bulk.

So the previous control arrangement was very limiting. It meant that the 
robot had an effective radius of action of maybe 10 to 20 yards from where 
my car was parked. This greatly limited the available terrain for testing. 
That's why all testing so far has been in open parking lots.

A PDA like an older model HP iPaq running Linux would be even better. 
These have transreflective resistive touchscreen displays and much more 
capability than my calculator. However, the calculator is enough. Also, I 
don't want to introduce yet another system to learn and integrate.

The US Marine Corps had rugged iPaq PDA prototypes a few years ago. These 
looked like a normal commercial model except with an integrated Otter box 
shell and a VHF spread spectrum radio. When you start trying to address 
the need for computing and communications in something rugged and powerful 
yet small and light enough to be portable, the PDA is a pretty good 
solution.