GPRAM 2013

Dear viewers,

Thank you for interesting in this work. In this demo, we show what we have accomplished in 2012 in the aspect of hardware demo of GPRAM prototype. This page is divided into three sections, (a) Motivation. We aim to answer questions like why we focus on this demo? What is the role of demo in GPRAM theory? Etc. (b) Detail of demo and (c) What we will do the next step.

(a) Motivation: As you all are aware of that the computer has become so small, so powerful, and so popular, it will put eyes, legs and arms in your smart phones, more importantly it will allow millions of people to try out their own idea to fuse and process those data. One day, one or many will open up new ways which lead us to intelligent machines, nothing like those proclaimed by experts in the past and in today’s scientific fields. If GPRAM theory is correct, then these ways will not be mathematically provable. The key is demo, building up real things that may not be perfectly following logic and may be very humble in its beginning, but it works and undeniable. Our GPRAM prototype is one of trials in this direction.

(b) Detail of Demo:?In this platform, we mount a wireless camera on a radio controlled toy car. The wireless camera will send images to a computer and then computer will analysis images to determine what to do (i.e., moving forward/backward, turning left/right). Finally the computer sends command to car and car moves.

In this 3 minutes demo, we showed that when we pointed a torch (my iphone) toward the camera, the car will follow the direction of torchlight. If we shine it from the left hand side, it will move forward and turn left. On the top right of video screen, you can find computer screen, which displays the detected torchlight and its direction. If there is no torchlight, then the car moves randomly.

In demo 1, (demo1 (78 M bytes)), we use perfect logic design.

In demo 2, (demo2 (63 M bytes)), we replace perfect logic by GPRAM.

*Note, password protected, if you are interested in our work, please contact us at bwndai at gmail dot com to get the password.

(c) Future works:

Step 1: Improve the learning capability and code expansion

Step 2: Multi-tasks and FPGA implementation