Prototyping the Arduino NAO Skin today

Ho and his mother in law made this... Its a mitten that goes over NAOs hand.

and I will test it today...

The principle of reading this sensor is here...

http://www.plugandwear.com/default.asp?mod=cpages&page_id=34

But we've got a 5v signal going to 7 possible sensors which vary their resistance when they are touched. No idea if this is going to work. Lets try it out!

OK (not bad, 1 hours work) wired it up on a very bloody messy mini-breadboard, (sushi board) and it works. [couldnt fit two of the sensor wires in].

The signals received for various kinds of tactile stimulation of the glove are shown below.

Problem is shown below! The responses to pressure in ANY part of the glove are much too correlated such that it is impossible to tell WHERE on the glove there has been a touch. It seems that this fundamental circuit idea does not work with this circuit.

Using

http://boolscott.wordpress.com/2010/02/04/arduino-processing-analogue-bar-graph-2/

Push on left side of glove.

Dhhhh, dimwit Fernando. Ho solved the problem by simply giving an extra pull down through a 1kOhm resistor for EACH analog input signal. If I understood electronics this would make sense. This gives a perfectly nice and specific reading, see below...

So the glove works.

Next step.

1. Improve the design of the glove Mark I, to make it more attractive, and flexible, e.g. with velcro movable sensor pads so that some reconfigurability can be achieved.

2. Get the Xbee wireless communication working so that we can begin to put the base onto the robot back.

Nanda is going to take the Xbee stuff home and get it working by Friday at 4pm. Friday is going to be great fun because Chris Jack and I are going to make an eye on a stick in a box, controlled by MEX which will run a Darwinian neurodynamic algorithm, and be submitted to Living Machines on 22nd March.