[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]

Agrego link a un nuevo articulo sobre el tema, que profundiza mas en los
aspectos tecnicos de la simulacion de la red neuronal e incluye un video
del robot en accion:

http://hackaday.com/2014/12/15/gift-your-next-robot-with-the-brain-of-a-roundworm

Saludos,
     Daniel.

---


On Thu, Nov 27, 2014 at 2:30 AM, Daniel Larrosa <dflc en cs.com.uy> wrote:

>
>
> 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/20141215/ac415caa/attachment.html>