Initial project direction is to create a mice Vs human gaming system. The speed of one mouse running inside a hamster wheel will be dynamically mapped onto the up and down movement of a plane avatar. The mapping will be semi-arbitrary meaning that shifting of directions is completely random but the speed will be directly proportional to the hamster wheel. Another mouse will be controlling the firing rates of the plane through a similar hamster wheel system. The human will then control another plane avatar trying to shoot down the mice controlled NPC.
The setup creates an illusion of two mice talking to each other through a computing platform. Each mouse rotates a hamster wheel and the generator connecting to it sends a digital on/off signal to the arduino microcontroller. The program will then randomly play out words from a pool of about 250 vocabularies. There is a simple algorithm to the order of words playing out:
… 1 verb, noun or adjective –> 1 connector (e.g. verb to be, pronoun, preposition) –> 1 verb, noun or adjective…
There is no overlapping from words playing out by individual mouse. That is, a word will only play after a prior word triggered by the same mouse finished playing. However, since the system is supposed to create an illusion of a conversation, the two sequences of words generated by the two mice will overlap each other.
I changed from the original proposal because the initial project idea is not creating a strong enough ‘personality’ even though it does incorporate a ‘creature’ in the system. The mapping between the npc and the mice is too abstract and random which would be difficult to form a compelling end result. The current project is very effective in terms of creating new and unique personalities across living bodies on top of already existing behaviors and individualities. Even though the mapping still does have a layer of abstractness, it creates an effective delusion of free-will. It looks as if the mice are splitting out words at their own will without the intelligence of forming grammatically correct and/or meaningful sentences.
If the project is to be put one step further, I would manipulate the algorithm of words output. A more sophisticated algorithm that closely resembles the English language should yield a more interesting output. In addition to that, I would like to play with other abstract mapping like the faster the hamster wheel spins, the faster the system talks. Or even map the speed of the hamster wheel to the pitch of the voice generated by the computing platform.
Great concept for the talking mice, and compelling demonstration. Simple details in the graphical quality of your screen output (the talk bubbles), combined with the live mice, made for a lot of charm and personality. It would have been nice to see at least a little more of a connection between the mice’s actions and the sentence output (as you mention for further directions); the current output ends up being either “off” or “on” instead of with a wider-ranging paramenter. But overall very convincing.