[Butia-devel-list] Articulo sobre robot similar al Butia, controlado por red neural que imita la de un nematodo, "C. Elegans", neurona por neurona

[Daniel Larrosa]

http://www.newscientist.com/article/mg22429972.300-first-digital-animal-will-be-perfect-copy-of-real-worm.html?full=true


Algunos extractos:

"THE Lego robot trundles forward, encounters a wall, stops, reverses. You
might think there is nothing clever about that, except that this one has
not been programmed to tell it when to stop and when to turn. Instead, it
has an artificial brain precisely modelled on that of a nematode worm.
WormBot is part of an effort to build the world's first digital animal."
...
"The bot's artificial brain has the same number of cells as a real nematode
brain, and they are connected up in exactly the same way. But instead of a
fluid tubular body animated by 95 muscles, WormBot has a plastic body and
two wheels."
...
"Its network of 302 neurons connect via 6393 synapses – its connectome –
and link to the worm's 95 muscles at 1410 junctions."
...
"There's a long list of reasons for creating a virtual animal. "The mere
act of trying to put a working model together causes us to realise what we
know and what we don't know," says John Long, a roboticist and
neuroscientist at Vassar College in New York State. Digital animals could
complement and in some cases dispense with animal experiments. C. elegans
is such a boon because it is so simple, yet 80 per cent of its genes are
the same as ours."
...
"Instead of muscles, WormBot has two wheels controlled by a matrix of 95
cells, representing the 95 muscles of C. elegans. Busbice hooked up the
worm's chemosensory neurons, which a real worm uses to detect smells and
tastes, to a microphone that is triggered beyond a certain decibel
threshold. He also connected the worm's "nose-touch" neurons to a sonar
that would send a message upstream to the brain if WormBot gets within 20
centimetres of an obstacle.

When all this was assembled, Busbice flipped the switch and began
whistling, softly at first then louder and louder. Once his whistles got
loud enough, the wheels were set in motion and WormBot moved towards the
sound much as a real worm would be lured by the smells wafting from food.

When the bot came across a chair leg, its sonar picked up the obstacle, and
without being programmed to do so, it stopped and reversed. "It's only the
connectome that makes it happen," says Busbice."

---

Podria ser un experimento muy interesante para realizar con los elementos
de hardware y software del Butia, quizas en colaboracion con gente del
Inst. Clemente Estable y/o de otras facultades (ej. Fac. de Ciencias,
Medicina, Veterinaria).

Se podria pensar alguna variante, ej. comenzar con la version con ruedas
como la del articulo y pasar luego a una version que utilice 95 servos,
para replicar uno a uno los 95 musculos del nematodo y obtener un
"comportamiento" mas realista, que admita otro tipo de movimientos.

Se podria pensar en crear el modelo fisico de nematodo, que pudiera servir
para las pruebas de distintos modelos de redes neurales y/o algoritmos y
circuitos de control, de modo que distintos estudiantes pudiesen trabajar
con el para sus proyectos de investigacion, tesis, etc, apuntando cada
grupo a un aspecto especifico de la simulacion o a modelos completos de
comportamiento y que lo utilicen por turnos para probar sus modelos o
incluso que distintos grupos colaboren para crear un modelo de mayor
alcance entre todos, etc.

El tema plantea aspectos interesantes desde el punto de vista del hardware
tambien, por ej. como simplificar el manejo de los 95 servos en forma
simultanea, para minimizar el hardware requerido, sin perder flexibilidad
para el manejo individual de cada servo ni velocidad de respuesta del
conjunto, para emular adecuadamente los comportamientos, etc.

Saludos,
     Daniel.
------------ próxima parte ------------
Se ha borrado un adjunto en formato HTML...
URL: <https://www.fing.edu.uy/pipermail/butia-devel-l/attachments/20141127/7bad30c9/attachment.html>