CABot1

I've currently got a EPSRC sponsored grant running to build a games agent, which processes language, does vision, and acts, all based on fLIF neurons.
This started in August 2006
I had scheduled three versions of code with the first due on July 31st 2007; the full agent wasn't going to be put together until the final version.
One of the grant reviewers said it could not be done.
So, I felt I had to prove the reviewer wrong and spent most of that year building CABot1.
It works.
The parser is stack based.
Vision is relatively faithful to retinal and primary visual cortex.
It recognizes pyramids and stalactites from the (crystal space) games environment.
Planning is derived from the Maes net with goals set from the user input.
The most sophisticated thing it does is "Go to the pyramid."
There are symbolic hooks to get the user commands and send the CS stub commands.
It is far from perfect, but we clearly can do it.

The CA roBot version 1 (CABot1)