Al-Bot / STANN Bot
[Background]
This project was my entry into the local Bay Area Science and Engineering Fair (BASEF). It was designed to be a development platform for artificial intelligence technology. It had a modular board setup where all boards plugged into a backplane, which allowed boards to be swapped around.
As well it had a colour camera in the form of a CMUCam that could be used to perform object tracking.
This did use to be called STANN Robot. The name changed when I was working on it though to Al-bot, thats is AL-bot and not AI-bot. The reason for this is my friends young cousins mis-pronounced AI-bot (for Artificial Intelligence bot) as 'Al-bot' in a game, and the name stuck.
[A Huge Thanks]
This project has really been intersting so far, and I have learned a lot from it. However there have been a few people I would like to thank greatly, they are:
Atmel - Provided me with STK500, STK501, JTAG ICE and some AtMega128 samples. These development tools have been a key part in the design of my robot, and make it possible to do much more in less time!
Eivind Sivertsen - Webmaster of AVRFreaks.net
Jacob Lunn Lassen - Field Application Engineer for Atmel AVR Products in Norway
Tim Kitagawa - Atmel Field Application Engineer in Canada
Joe Young - Regional Sales Manager for Atmel in Canada
Honeywell - Provided me with M22S rotation sensor
Yan Gaudet - Technical Application Specialist
[Results]
At BASEF from April 2nd to 5th 2003 I won the following awards:
Best in Fair - 2nd Place
Intel International Science and Engineering Fair Trip
Gold Merit Award
Procor Senior
PEO (Professional Engineers of Ontario) Hamilton Chapter - First Place
Mechanical Construction Association
McMaster University Tuition - Engineering or Science
IISEF - Yale Science & Engineering
Dofasco - Process Automation
Bell Computer & Communications - 2nd
See http://basef.mcmaster.ca/2003/results/gallery/display_project.cgi?project_id=MA11 for more detail on my project there (has picture).
At the Intel International Science and Engineering Fair in Cleveland (May 2003) I won a third place award for the catagory of engineering and I attened this fair as a member of Team Canada 2003.
[Download]
The download file includes much C code, VHDL, schematics, and documention. May be useful if you want to see:
- colour object tracking using CMUCam
- my own communications protocol called the StannBus using a backplane design
- motor speed controller
- FRAM access
- using SONAR
- how neural networks work as it includes a large documentation package with introduction on them
Alternately you can grab the files from a local mirror: STANN_bot.zip [1.8 MB]
[Photos]
Here are two photos that were on by backboard of the display, the show photos of the completed robots as well as parts:
The Vacuum Flourescent Display on the bot
Another picture of the processor board. Note that I reversed one of the connectors, hence the adapter board glued on that corrects that error...
UPLOADED FEB 26, 2003
An updated image of the robot base. You can see a few parts mounted. As well the top of it is Lexan, but the protective paper is still on it.
An updated picture of the finished backplane. Here you can see everything mounted on it, as well as a blue wire that corrects an error in the PCB...
My processor board hooked up to my JTAG ICE. The square chip is an AtMega128. The large 28-DIP is my FRAM chip. At the bottom you can see a space for two connectors that plug into the backplane. But I forgot to order them from DigiKey so time for a new shipment...
The JTAG ICE is really helpful here, I could quickly test that my FRAM was working without having to get my serial port working and finding an RS232 line... plus the fact I had to use most of the port-pins in the AtMega128 here anyway made it hard to attach a LCD display to spare lines or whatnot.
The 'sensor board'. This thing connects to the camera, the stepper motor, and the sonar. Beside the AtMega8 (the 28-pin DIP) there is a SMD op-amp mounted dead bug style.
The 'not 3.3 volt complient' stickers are there because the 'Stann Bus' (the bus that is on the backplane) should be able to work at 3.3 OR 5 volts. This card doesn't sense that though, and would rip apart a 3.3 volt bus by applying 5 volts on it. Actually none of the cards really work at 3.3 volts, but they DO recognize a 3.3 volt bus and just shut down to prevent them from breaking other cards... I guess since I DID make the protocol I could just take the 3.3 volt part out but oh well...
UPLOADED JAN 20, 2003
This is the base the robot will be built on, at least one more level will be added.
My STK500 + STK501 testing the FRAM interface that will hold the neural network data
The SONAR module that will allow the robot to detect obsticals
The backplane to which the other boards will plug into. U5 is not yet mounted on it as you can see...
The CMUcam + motor + M22S sensor that track on object. The camera can now track a ball and is fairly intelligent, for example if the ball goes out of range the camera will continue to move in the last known direction of ball movement to have a better chance of "catching up".