3D lattice of RGB LEDs driven by ANN with environmental sensors as inputs
- Programmable cube ($390) with Mic and Acc: http://cubetube.org/
- http://www.instructables.com/id/8X8X8-RGB-LED-Cube/
- NeoPixel (https://learn.adafruit.com/adafruit-neopixel-uberguide/overview) single wire, RGB
- Audio Spectrum Analyzer with Pi
- Something acquires a sample each from many sensors
- These samples are passed to a Neural Network which does some nonlinear processing on the vector of input time series.
- The outputs of the NN are passed to a an output processor, which takes the outputs and writes them to the LEDs.
- The NeoPixel or DotStar style of Arduino / Raspberry Pi compatible LEDs are a single strip of addressable RGB LEDs.
- there are 3rd party products also like HKBAYI
- The FadeCandy board takes a USB input and can drive 8 strips having 64 LEDs each. That's a total of 8x64 = 512 LEDs.
- could do a cube with 4 sides + 1 top. 10x10 LEDs per side = 500 total.
- The LED strips can be mounted on some clear plastic rods so as to make the shape into something like a cube.
- use a 3D printer to make some wild shapes to mount it on: trees, spheres, Japanese lantern, Klein bottle
- maybe hang them from a frame like Hanging Gardens or the living trees in Avatar
- needs ~60 mA per LED for full power. Should use a 5V, 10A AC/DC adapter and a power bus to spread power to each strip.
- GitHub Markdown: https://guides.github.com/features/mastering-markdown/