Apart from that, the flash news is, yes I have ordered an xport and thinking to do a a project where network involves for the end of semester. I really want to get my hands dirty with physical networks.

I think the question we have to ask to ourselves is before starting any production, how I am going to approach to it. Is it going to be an extension for a current acoustic instrument,or is it going to be a total new interface? Personally I am kind of curious about two things currently. Using slit-scanning as a real time composition tool and using brainwaves to perform and compose.

I have found couple of good papers around those subjects:
Computer Music Research page of Plymouth University, UK
Interfacing the Brain Directly with Musical Systems: On developing systems for making music with brain signals
Tactile Composition Systems for Collaborative Free Sound
An Informal Catalogue of Slit-Scan Video Artworks by Golan Levin.

Those could be a good start for inspiration.-

Joseph Paradiso (he is the director of the Responsive Environments Group)has paper dated 1998 where he describes new electronic music interfaces.

So I started basically with building a bot which follows light to see and to experience its behaviours according to basic directions. Then I try to see this simulation on the computer which gave me the possibilities of multiplying the agent.

What I am showing in this example is agents are just merely following light with their basic neurons, so they are checking at that point their proximity to light and behaving according to it. While this is the most common approach today, I questioned if there could be a way to extend this behaviour which makes it to learn behave smarter in time.

So I added couple of new things to my bot, I multiplied its neurons to get more accurate results, I added capabilities which are called avoid systems meaning if it is too close to light or getting light from more in its tail, just run backwards.

The most challenging one was to make it understand certain patterns so that it can learn to behave according to these patterns. The main problem I was having is I had only one bot and no generations. The genetic algorithms work in this way.

url

Results and further questions
It is not easy to implement genetic algorithms into physical devices.
How could this approach be useful in our daily lives and user experiences with devices?
Think of a building where the floors should be painted could this approach be applied to this kind of scenario?

A few things to look at:

According to http://www.fairchildsemi.com/ds/2N/2N3904.pdf the pinout from closest to farthest in the picture on the web page is emitter – base – collector. The emitter in an NPN transistor like the 2N3904 corresponds to the side with the arrow in the schematic. It looks like the transistors are wired backwards and you need to reverse the connections to the two outer pins (the middle pin, the base, is fine). Or you could turn the transistor around.

Second, and equally important, there are too many diodes in series with your connection to the power supply. My guess is that you started adding diodes until you measured 3V with a voltmeter. The problem with this is that _sort of_ like a resistor, the more current is flowing the more voltage will drop across the diode (but in a diode the relationship isn’t nice and linear like with a resistor). So if you want to measure the voltage, you need to do it with the motor connected, drawing power. You could get pretty close without the transistor, by just connecting the motor straight from the last diode to ground. However a pretty good starting point in most cases is to estimate the voltage drop across each diode as about 0.7 volts. This means you would need 3 diodes in series to drop the 2 volts between supplies (also the transistor will drop a bit of voltage so you might find that two work better if the motor can handle a bit over 3V). With 7 diodes in series from a 5V supply only a tiny bit of current can flow (enough to feed your voltmeter, but not enough for your motor).

Also, I don’t see the protection diode on the schematic and that you mention in the caption. It should be across the two outer terminals of the transistor, with the band pointed toward the collector (the pin farthest from us in the picture).

If you make those changes, the circuit _should_ work. Another thing to consider though is that the 2N3904 is only rated to carry an average of 200mA of current, which may be too low for your motors, but I’m not sure how much they take. It all depends on the specific motor, but you can measure how much it draws with a meter set to DC Amps. If this is a problem, I don’t see any reason that the same circuit should work with a TIP120 (the pins are different though – in a TIP120 they go Base – Collector – Emitter from left to right as you are looking at the part number) which can carry plenty of current, but you will probably only want 2 diodes in series with your power at that point, because the TIP120 drops more voltage than the 2N3904. If the circuit works with the 3904 and the transistor doesn’t get hot enough to burn you within a few seconds to minutes you are probably OK though.

One more thing – if you have any doubt as to whether or not your PIC is the problem, you can always pull the wire from the PIC pin and connect it to the 5V supply which should turn the motor on.

Hope this helps.
-N.

it was a good experience for me to make it a bot work, and it is even reacting to light. So couple of next steps :

I should make the two motors speed equal or as equal as I can do, right now it is going a little bit left( not much) for this I am going to use HPWM as friends suggested, I got a code example of this for my servo somewhere, I have found one in Tom’s site as well related to LEDs.

I am going to replace transistors with an H-Bridge so that I can add backwards capability to my photovore. And with that I can add an avoidance capability as well(when it is so close to light, just step back from the source)

I am thinking to expand the photovore’s I/O relationship which is digital right now. It can be either positive proportional or a gaussian function.

And at last but not at least, I am thinking to add some algorithm to make the photovore not more aware of the environment but actually comparing the environment according to the data it keeps for its previous experiences.. Could it be possible? I am going to experiment and see…

that’s it for now.

There are some problems with wheels and I am sure they are going to give me headaches but still I am pretty happy with the progress I made, here is the code that takes the values from CdS and print on the hyperterm.

‘ 10-bit A/D conversion
‘ Connect analog input to channel-0 (RA0)

‘ Define ADCIN parameters
‘ Set number of bits in result
DEFINE ADC_BITS 10
‘ Set clock source (3=rc)
DEFINE ADC_CLOCK 3
‘ Set sampling time in microseconds
DEFINE ADC_SAMPLEUS 10

adcVar VAR WORD ‘ Create variable to store result

‘ Set PORTA to all input
TRISA = %11111111

‘ Set up ADCON1

ADCON1 = %10000010
v1 var byte ‘ variable v1 holds cdS #1 information
v2 var byte ‘ variable v2 holds cdS #2 information
Pause 500 ‘ Wait .5 second

main:
ADCIN 0, v1 ‘ Read channel 0
ADCIN 1, v2 ‘ Read channel 1
SEROUT2 PORTC.6, 16468, ["CdS 1 = ", DEC v1, 10, 13]
pause 5
SEROUT2 PORTC.6, 16468, ["CdS 2 = ", DEC v2, 10, 13]
Pause 100 ‘ Wait.01 second
GoTo main

- Image and Vision Computing
- Pattern Recognition

And some random article fishing :)
-Object segmentation using maximum neural networks for the gesture recognition system
-Hand gesture recognition using a real-time tracking method and hidden Markov models

I got the parts from Images SI url, I have two gear motors, 2 rubber wheels and one omnidirectional wheel as the front wheel.. I got the plate last week from canal street witht he u bracket and the rods so do I need anymore parts on a second thought… I guess I am quite allright, I just need the photocells, and I am going to buy it from NYU store so I am ready to start building the stuff.

I am goingto go and try to finish it tomorrow at least make it work in these two days, it looks easy and we will see.

So my goal is make it respond to the light sources around itself in a certain threshold, first goal.
By the way I have found a nice code about Braitelberg Devices and someone actually did this experiment in processing before and this guy is one of the guys that helped Casey Reas back in Whitney Exhibition Software Structures url He is William Ngan url.

So I am trying to improve the code, oh ye ah I am trying to understand it first to improve it :)
More to come.

BEAM is alternatively said to stand for: Building Evolution Anarchy Modularity Biotechnology Ethnology Analogy Morphology.
Most BEAM robots are unusually simple in design compared to traditional mobile robots, and trade off flexibility in purpose for robustness of performance.

BEAM style robots are coming from behavior-based robotics approach.
The school of behavior-based robots owes much to work undertaken in the 1980s at the Massachusetts Institute of Technology by Professor Rodney Brooks, who with students and colleagues built a series of wheeled and legged robots utilising the subsumption architecture.

Behavior-Based Robotics by Ronald C. Arkin from MIT press.