Older blog entries for roschler (starting at number 5)

I have built a phpBB2 forum for use by Robots.net members

only. I talked to steve about letting the members know about it and he liked the idea. You can find it here:


I am keeping the forum private to keep out spammers. The recent attack on our diary system seems to confirm this as a good decision.

You need to register first so you can be granted access to the forum; all registrations will be hand approved. Use the following as your username/handle or your registration will not be valid:

Use your Robots.net handle, followed by the underscore character ("_"), followed by the handle of the first person that certified you as shown on your bio page.. For example, in my case that would be "roschler_steve". You username will be fixed (i.e., appended name removed) once you are approved.

If one of the master level members would like to turn this into an article I would appreciate that. This will make sure that everybody sees this message. Due to the diary attack, </b>messages are starting to disappear</b> from the new diary entries stream. Please delete this final paragraph if you do so. Thanks in advance!

That's weird, my Feb 17 post on the Tomy robot just got duplicated under Wendy's zxcvb123 name. Is there a problem with the diary database? Thanks.

Joining this site has stirred up some old memories. I had almost completely forgotten about another robotics project with which I was involved.

In the 1980's there was a robot called the Tomy Robot. It had an extendable gripping arm, a head that swiveled, and tank tracks that let it zip around a house.

My friends ripped the thing's head off and placed a camera on it. We had one of the first Picture-In-Picture TV's back then, and on the TV we could show the image from the robot's camera. Since the camera was mounted on the swivel axis, it could be rotated.

Then they tore apart the remote control that controlled the robot. They wired each of the buttons on the remote control to one of 16 opto-electronic switches. I was able to drive the on-off position of each switch on the board of switches, by sending a 16-bit word out to the parallel port connected to the board of switches. Each bit turned one switch on or off, depending on its value of 1 or 0.

Next I programmed a user interface that allowed a handicapped person to operate the robot's movement, swivel, and gripper capabilities through a graphical interface. I wrote the driver's for an infrared sensor connected to the PC, that could detect head movement (with the aid of a small piece of reflective material adhering to the user's forehead). This allowed a handicapped person to control the X, Y movement of the mouse cursor with their head. An eye blink sensor placed near the person's eye allowed them to trigger a mouse click by blinking their eye. The overall package gave the handicapped user, even if they were a quadraplegic, the ability to operate the graphical interface to the mutated Tomy robot.

In summary, a handicapped person could drive the Tomy robot around the house, see what it was seeing and even look around the room, and grab things with the gripper hand.

I remember that project with great fondness.

This is a subject that I just can't leave alone. I have had running discussions with several people who feel that the race to control robots with brain interface connections, will take a very long time.

Their argument is based on the premise that, to interface with a brain you have to map its functions in detail, to get any real performance.

I believe this is patently false.

Neurons have an innate ability to build connections and interact with an amazing variety of signals.

The brain has a wonderful ability to build connections to a plethora of electrical stimuli; as long as that stimuli provides a reasonable pattern of repetition; even if there is a fair amount of chaos and aperiodicity in the signal.

Anyone who has read up on the development of the brain in human babies, cannot help but walk away astonished at the adaptability of the human (and primate) brain.

There have been recent experiments with monkeys and implantable electrode interfaces that show how quickly a brain can create new connections and adapt to a new signal environment.

Things are going to happen a lot faster than I think most people even dream about. Not because of advances in brain function mapping, but due to the incredible pattern analysis capabilities of the brain, and its ability to adapt to an amazing variety of signal patterns.


I know this was covered a few years ago on robots.net, but The MIT Leg Library now has an MPEG video (QuickTime) of Troody on their web page:


To recapitulate, Troody is a robot built to study and emulate the biomechanics of Dinosaurs. Hey, it's a robot that looks like a dinosaur!

I figured I would add some more detail to text on my bio page about the six jointed robotic hand project. Here's the architecture that drove it.

I built the hand in the middle 1980's. I used two computer systems to drive it; a Windows 3.1 system and a pure DOS machine. The DOS machine ran the Robix driver programs which communicated with the robot hand joint motors over an RS-232 port. I had to use two machines because the Robix drivers needed real time access to the robot hand motors. This meant you could only run low level programs that used it's robot programming language and not Window 3.1, on the same machine. The genetic algorithm code was much easier to write and test under Windows 3.1. The Windows 3.1 machine communicated to the DOS machine over a custom written RS-232 comm bridge module.

Share this page